diff --git a/.agents/skills/debug-cudf-pandas/SKILL.md b/.agents/skills/debug-cudf-pandas/SKILL.md index e19fe214a4ae..fbb7ea23b5f6 100644 --- a/.agents/skills/debug-cudf-pandas/SKILL.md +++ b/.agents/skills/debug-cudf-pandas/SKILL.md @@ -71,7 +71,7 @@ The file `python/cudf/cudf/pandas/scripts/pandas-testing-plugin.py` contains thr - **`NODEIDS_TO_SKIP`** — tests marked `skip` (not run at all). Keys are alphabetically sorted. - **`NODEIDS_PATHS_TO_SKIP`** — prefix-based path skips covering entire modules. -The pandas-tests harness runs with `xfail_strict = false` (set in the vendored `pandas-tests/pyproject.toml` to tolerate flaky XPASSes — [rapidsai/cudf#22681](https://github.com/rapidsai/cudf/issues/22681)). A test listed in `NODEIDS_THAT_FAIL` that now *passes* is therefore reported as `XPASS` **without failing the run**, so a stale entry will not flag itself. You must change the false to true yourself before testing your fix; otherwise the test reports `XPASS` instead of a genuine `PASSED` and the dead marker lingers silently. Do not commit this change in any commit. +The pandas-tests harness runs with `xfail_strict = false` (set in the vendored `pandas-tests/pyproject.toml` to tolerate flaky XPASSes — [NVIDIA/cudf#22681](https://github.com/NVIDIA/cudf/issues/22681)). A test listed in `NODEIDS_THAT_FAIL` that now *passes* is therefore reported as `XPASS` **without failing the run**, so a stale entry will not flag itself. You must change the false to true yourself before testing your fix; otherwise the test reports `XPASS` instead of a genuine `PASSED` and the dead marker lingers silently. Do not commit this change in any commit. Search for the node ID: diff --git a/.agents/skills/perf-compare-cudf/SKILL.md b/.agents/skills/perf-compare-cudf/SKILL.md index 30bdd2423486..0d82c5a9cff2 100644 --- a/.agents/skills/perf-compare-cudf/SKILL.md +++ b/.agents/skills/perf-compare-cudf/SKILL.md @@ -4,14 +4,14 @@ description: Benchmark a cuDF branch, WIP changes, or a PR against the `main` br --- Use this skill when the user asks to compare libcudf benchmark performance for: -- **the current branch or WIP changes** against `rapidsai/cudf` `main`. -- **a cudf PR link or number** against `rapidsai/cudf` `main`. +- **the current branch or WIP changes** against `NVIDIA/cudf` `main`. +- **a cudf PR link or number** against `NVIDIA/cudf` `main`. # Goal -Run the same selected libcudf NVBench benchmarks on the target (current WIP or cudf PR) and then on `rapidsai/cudf` `main`, then report meaningful differences. +Run the same selected libcudf NVBench benchmarks on the target (current WIP or cudf PR) and then on `NVIDIA/cudf` `main`, then report meaningful differences. -`` is the git remote for `https://github.com/rapidsai/cudf` (often `upstream`). Detect it with `git remote -v`. +`` is the git remote for `https://github.com/NVIDIA/cudf` (often `upstream`). Detect it with `git remote -v`. ## Prerequisites @@ -37,7 +37,7 @@ Run the same selected libcudf NVBench benchmarks on the target (current WIP or c - For current-branch or WIP targets: keep target changes applied for the target run. - For PR targets: Stash any unrelated local changes, record the stash name, and check out the PR: ```bash - gh pr checkout --repo rapidsai/cudf + gh pr checkout --repo NVIDIA/cudf ``` - For PR targets: After switching, check if the PR branch is behind `/main` and add a merge commit. DO **NOT** push anything. If there are merge conflicts, stop and guide the user to fix them. diff --git a/.agents/skills/reproduce-ci/SKILL.md b/.agents/skills/reproduce-ci/SKILL.md index 8e833d3df503..7dfe92921114 100644 --- a/.agents/skills/reproduce-ci/SKILL.md +++ b/.agents/skills/reproduce-ci/SKILL.md @@ -39,7 +39,7 @@ You need three arguments for `run.sh`: **container image**, **CI script**, and * ### Option A: Discover from a GitHub Actions Job URL Use this option when the user provides a URL like: -`https://github.com/rapidsai/cudf/actions/runs//job/?pr=` +`https://github.com/NVIDIA/cudf/actions/runs//job/?pr=` **Parse the Job URL:** @@ -66,7 +66,7 @@ If the script is unavailable, extract manually: Use the extracted IDs to get job details: ```bash -gh api repos/rapidsai/cudf/actions/runs/$RUN_ID/jobs \ +gh api repos/NVIDIA/cudf/actions/runs/$RUN_ID/jobs \ --jq ".jobs[] | select(.id == $JOB_ID)" ``` @@ -77,7 +77,7 @@ From the job JSON, note: Download the full job log: ```bash -gh run view "$RUN_ID" --repo rapidsai/cudf --job "$JOB_ID" --log > /tmp/ci_job_log.txt +gh run view "$RUN_ID" --repo NVIDIA/cudf --job "$JOB_ID" --log > /tmp/ci_job_log.txt ``` Read through the log to identify: @@ -139,7 +139,7 @@ The script automatically detects `RAPIDS_SHA` (the PR's head commit) using `gh p This is required by CI helper scripts inside the container to locate build artifacts. If auto-detection fails (e.g., `gh` is not authenticated), set it manually: ```bash -export RAPIDS_SHA=$(gh pr view --repo rapidsai/cudf --json commits --jq '.commits[-1].oid') +export RAPIDS_SHA=$(gh pr view --repo NVIDIA/cudf --json commits --jq '.commits[-1].oid') ``` The script launches a detached container, runs the CI script, and leaves the container running for inspection. After `--timeout` minutes of idle (default: 30), the container is automatically removed. @@ -173,7 +173,7 @@ After `run.sh` completes, analyze the local output against the CI outcome: | Problem | Fix | |---------|-----| -| `GIT_DESCRIBE_NUMBER is undefined` | `git fetch https://github.com/rapidsai/cudf.git --tags` | +| `GIT_DESCRIBE_NUMBER is undefined` | `git fetch https://github.com/NVIDIA/cudf.git --tags` | | Interactive GitHub auth prompt inside container | Ensure `GH_TOKEN` is set — `run.sh` passes it through automatically via `gh auth token` | | GPU driver mismatch causing test differences | Note driver version from CI log; compare with local `nvidia-smi` | | Log download returns empty or 403 | Verify `gh auth status` has `repo` scope; re-auth with `gh auth login` if needed | diff --git a/.agents/skills/reproduce-ci/parse-job-url.py b/.agents/skills/reproduce-ci/parse-job-url.py index 8413068b8dfb..50c07349a5ca 100755 --- a/.agents/skills/reproduce-ci/parse-job-url.py +++ b/.agents/skills/reproduce-ci/parse-job-url.py @@ -5,7 +5,7 @@ # Parse a GitHub Actions job URL and emit shell-friendly variable assignments. # # Usage: -# python parse-job-url.py "https://github.com/rapidsai/cudf/actions/runs//job/?pr=" +# python parse-job-url.py "https://github.com/NVIDIA/cudf/actions/runs//job/?pr=" # # Output (eval-able): # RUN_ID=XXXXXXXXXX diff --git a/.agents/skills/reproduce-ci/run.sh b/.agents/skills/reproduce-ci/run.sh index ca66d721261b..3a6da4ffbfeb 100755 --- a/.agents/skills/reproduce-ci/run.sh +++ b/.agents/skills/reproduce-ci/run.sh @@ -90,7 +90,7 @@ fi # Determine RAPIDS_SHA from the PR's head commit RAPIDS_SHA="${RAPIDS_SHA:-}" if [[ -z "$RAPIDS_SHA" ]]; then - RAPIDS_SHA=$(gh pr view "$PR_NUMBER" --repo rapidsai/cudf --json commits --jq '.commits[-1].oid' 2>/dev/null || true) + RAPIDS_SHA=$(gh pr view "$PR_NUMBER" --repo NVIDIA/cudf --json commits --jq '.commits[-1].oid' 2>/dev/null || true) if [[ -z "$RAPIDS_SHA" ]]; then echo "Warning: Could not determine RAPIDS_SHA for PR #${PR_NUMBER}." echo "Artifact downloads inside the container may fail." @@ -106,7 +106,7 @@ DOCKER_ARGS=( --volume "$PWD:/repo" --workdir /repo --env "RAPIDS_BUILD_TYPE=pull-request" - --env "RAPIDS_REPOSITORY=rapidsai/cudf" + --env "RAPIDS_REPOSITORY=NVIDIA/cudf" --env "RAPIDS_REF_NAME=pull-request/${PR_NUMBER}" --name "$CONTAINER_NAME" -d diff --git a/.agents/skills/review-cudf/SKILL.md b/.agents/skills/review-cudf/SKILL.md index e70af8135378..647924faeaa6 100644 --- a/.agents/skills/review-cudf/SKILL.md +++ b/.agents/skills/review-cudf/SKILL.md @@ -8,15 +8,15 @@ Use this skill when the user invokes `/review-cudf` with: - currently checked out cudf PR - specified cudf code changes or a diff -cudf GitHub repository is located at: https://github.com/rapidsai/cudf +cudf GitHub repository is located at: https://github.com/NVIDIA/cudf # Review cuDF Pull Request 1. **Fetch PR metadata and diff** ```bash -gh pr view --repo rapidsai/cudf --json title,body,files,additions,deletions,baseRefName,headRefName -gh pr diff --repo rapidsai/cudf +gh pr view --repo NVIDIA/cudf --json title,body,files,additions,deletions,baseRefName,headRefName +gh pr diff --repo NVIDIA/cudf ``` Hint: Check if `GH_TOKEN` (or GitHub CLI auth) is already configured in the environment (for example via your secret manager) so `gh` can authenticate and bypass rate limits; do not run `gh auth token` from within the agent. If `gh` auth is unavailable, fall back to GitHub's raw diff/patch URLs, `git fetch` of the PR ref, unauthenticated GitHub REST API with `curl`, or any other available methods. diff --git a/.devcontainer/cuda12.9-conda/devcontainer.json b/.devcontainer/cuda12.9-conda/devcontainer.json index 4187c5a2b1d0..ef44f4f11b31 100644 --- a/.devcontainer/cuda12.9-conda/devcontainer.json +++ b/.devcontainer/cuda12.9-conda/devcontainer.json @@ -8,7 +8,7 @@ "BASE": "rapidsai/devcontainers:26.10-cpp-mambaforge" }, "cacheFrom": [ - "ghcr.io/rapidsai/cudf/devcontainer:26.10-cuda12.9-conda" + "ghcr.io/nvidia/cudf/devcontainer:26.10-cuda12.9-conda" ] }, "runArgs": [ diff --git a/.devcontainer/cuda12.9-pip/devcontainer.json b/.devcontainer/cuda12.9-pip/devcontainer.json index 51b1b4d22d5c..2e65922ae418 100644 --- a/.devcontainer/cuda12.9-pip/devcontainer.json +++ b/.devcontainer/cuda12.9-pip/devcontainer.json @@ -8,7 +8,7 @@ "BASE": "rapidsai/devcontainers:26.10-cpp-cuda12.9-ucx1.19.0-openmpi5.0.10" }, "cacheFrom": [ - "ghcr.io/rapidsai/cudf/devcontainer:26.10-cuda12.9-pip" + "ghcr.io/nvidia/cudf/devcontainer:26.10-cuda12.9-pip" ] }, "runArgs": [ diff --git a/.devcontainer/cuda13.3-conda/devcontainer.json b/.devcontainer/cuda13.3-conda/devcontainer.json index 7ceefa75516b..6af32c08eab2 100644 --- a/.devcontainer/cuda13.3-conda/devcontainer.json +++ b/.devcontainer/cuda13.3-conda/devcontainer.json @@ -8,7 +8,7 @@ "BASE": "rapidsai/devcontainers:26.10-cpp-mambaforge" }, "cacheFrom": [ - "ghcr.io/rapidsai/cudf/devcontainer:26.10-cuda13.3-conda" + "ghcr.io/nvidia/cudf/devcontainer:26.10-cuda13.3-conda" ] }, "runArgs": [ diff --git a/.devcontainer/cuda13.3-pip/devcontainer.json b/.devcontainer/cuda13.3-pip/devcontainer.json index a6194402e962..c871de37f323 100644 --- a/.devcontainer/cuda13.3-pip/devcontainer.json +++ b/.devcontainer/cuda13.3-pip/devcontainer.json @@ -8,7 +8,7 @@ "BASE": "rapidsai/devcontainers:26.10-cpp-cuda13.3-ucx1.19.0-openmpi5.0.10" }, "cacheFrom": [ - "ghcr.io/rapidsai/cudf/devcontainer:26.10-cuda13.3-pip" + "ghcr.io/nvidia/cudf/devcontainer:26.10-cuda13.3-pip" ] }, "runArgs": [ diff --git a/.github/ISSUE_TEMPLATE/pandas_function_request.md b/.github/ISSUE_TEMPLATE/pandas_function_request.md index 19f1377dfe73..f102acb83c30 100644 --- a/.github/ISSUE_TEMPLATE/pandas_function_request.md +++ b/.github/ISSUE_TEMPLATE/pandas_function_request.md @@ -7,7 +7,7 @@ assignees: '' --- -This issue template is intended to be used primarily for requests related to pandas accelerator mode. If you'd like to file a general cuDF feature request, please [click here](https://github.com/rapidsai/cudf/issues/new?assignees=&labels=%3F+-+Needs+Triage%2C+feature+request&projects=&template=feature_request.md&title=%5BFEA%5D). +This issue template is intended to be used primarily for requests related to pandas accelerator mode. If you'd like to file a general cuDF feature request, please [click here](https://github.com/NVIDIA/cudf/issues/new?assignees=&labels=%3F+-+Needs+Triage%2C+feature+request&projects=&template=feature_request.md&title=%5BFEA%5D). **Missing Pandas Feature Request** diff --git a/.github/PULL_REQUEST_TEMPLATE.md b/.github/PULL_REQUEST_TEMPLATE.md index 301037c8660b..eb19a89afea0 100644 --- a/.github/PULL_REQUEST_TEMPLATE.md +++ b/.github/PULL_REQUEST_TEMPLATE.md @@ -4,6 +4,6 @@ ## Checklist -- [ ] I am familiar with the [Contributing Guidelines](https://github.com/rapidsai/cudf/blob/HEAD/CONTRIBUTING.md). +- [ ] I am familiar with the [Contributing Guidelines](https://github.com/NVIDIA/cudf/blob/HEAD/CONTRIBUTING.md). - [ ] New or existing tests cover these changes. - [ ] The documentation is up to date with these changes. diff --git a/.github/workflows/pr.yaml b/.github/workflows/pr.yaml index 500f5a71636b..a67ef3971bed 100644 --- a/.github/workflows/pr.yaml +++ b/.github/workflows/pr.yaml @@ -441,7 +441,7 @@ jobs: with: build_type: pull-request script: ci/test_cpp.sh - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) conda-python-build: needs: [build-details, conda-cpp-build] @@ -488,7 +488,7 @@ jobs: with: build_type: pull-request script: "ci/test_python_cudf.sh" - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) conda-python-other-tests: # Tests for dask_cudf, cudf_polars, custreamz, cudf_kafka are separated for CI parallelism @@ -504,10 +504,10 @@ jobs: if: fromJSON(needs.changed-files.outputs.changed_file_groups).test_python_conda with: build_type: pull-request - # https://github.com/rapidsai/cudf/pull/22381/changes#r3196736965 + # https://github.com/NVIDIA/cudf/pull/22381/changes#r3196736965 container-options: "--cap-add CAP_SYS_PTRACE --shm-size=8g --ulimit=nofile=1000000:1000000" script: "ci/test_python_other.sh" - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) conda-java-tests: needs: [conda-cpp-build, changed-files] @@ -651,7 +651,7 @@ jobs: with: build_type: pull-request script: ci/test_wheel_cudf_streaming.sh - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) wheel-build-pylibcudf: needs: [build-details, checks, wheel-build-libcudf] @@ -705,7 +705,7 @@ jobs: with: build_type: pull-request script: ci/test_wheel_cudf.sh - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) wheel-build-cudf-polars: needs: [build-details, wheel-build-pylibcudf] @@ -833,7 +833,7 @@ jobs: if: (fromJSON(needs.changed-files.outputs.changed_file_groups).test_python_wheels || fromJSON(needs.changed-files.outputs.changed_file_groups).test_cudf_pandas) && fromJSON(needs.changed-files.outputs.changed_file_groups).neither_cudf_polars_nor_dask_cudf with: # This selects the latest supported Python + CUDA minor versions for each ARCH/CUDA major version combo - # Filter out GB300 due to https://github.com/rapidsai/cudf/issues/23498 + # Filter out GB300 due to https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: group_by([(.ARCH), (.CUDA_VER|split(".")|map(tonumber)|.[0])]) | map(max_by([(.PY_VER|split(".")|map(tonumber)), (.CUDA_VER|split(".")|map(tonumber))])) | map(select(.GPU != "gb300" and .GPU != "gh200")) build_type: pull-request script: ci/cudf_pandas_scripts/run_tests.sh diff --git a/.github/workflows/test.yaml b/.github/workflows/test.yaml index b15763c2654d..9ef2847a6dc2 100644 --- a/.github/workflows/test.yaml +++ b/.github/workflows/test.yaml @@ -54,7 +54,7 @@ jobs: date: ${{ inputs.date }} script: ci/test_cpp.sh sha: ${{ inputs.sha }} - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) conda-cpp-benchmark-tests: permissions: @@ -123,7 +123,7 @@ jobs: date: ${{ inputs.date }} sha: ${{ inputs.sha }} script: "ci/test_python_cudf.sh" - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) conda-python-other-tests: # Tests for dask_cudf, custreamz, cudf_kafka are separated for CI parallelism @@ -141,7 +141,7 @@ jobs: date: ${{ inputs.date }} sha: ${{ inputs.sha }} script: "ci/test_python_other.sh" - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) conda-java-tests: permissions: @@ -224,7 +224,7 @@ jobs: date: ${{ inputs.date }} sha: ${{ inputs.sha }} script: ci/cudf_pandas_scripts/run_tests.sh - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) third-party-integration-tests-cudf-pandas: permissions: @@ -279,7 +279,7 @@ jobs: date: ${{ inputs.date }} sha: ${{ inputs.sha }} script: "ci/test_wheel_cudf_polars.sh" - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) cudf-polars-polars-tests: permissions: @@ -296,7 +296,7 @@ jobs: date: ${{ inputs.date }} sha: ${{ inputs.sha }} script: "ci/test_cudf_polars_polars_tests.sh" - # https://github.com/rapidsai/cudf/issues/23498 + # https://github.com/NVIDIA/cudf/issues/23498 matrix_filter: map(select(.GPU != "gb300" and .GPU != "gh200")) narwhals-tests: permissions: diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index 7e6924f1f271..c46a8f93d4f3 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -45,7 +45,7 @@ repos: "polars>=1.35,<1.43", "numpy>=1.26", "pyarrow-stubs>=19.0", - "pyarrow>=19.0.0,<24.0.0", # https://github.com/rapidsai/cudf/issues/22229 + "pyarrow>=19.0.0,<24.0.0", # https://github.com/NVIDIA/cudf/issues/22229 "pytest>=8.0.0,<9.1.0", "types-cachetools>=5.5.0", "rmm-cu12==26.10.*,>=0.0.0a0; sys_platform=='linux'", diff --git a/CHANGELOG.md b/CHANGELOG.md index 4e27c0c89826..d39ff2c4ebf1 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -1,9677 +1,9677 @@ # cudf 26.08.00 (5 Aug 2026) ### 🚨 Breaking Changes -* Enforce pandas 3.0 deprecations and subsequent related cuDF deprecations part 1 by @mroeschke in https://github.com/rapidsai/cudf/pull/20823 -* Enforce Series.interpolate deprecations by @mroeschke in https://github.com/rapidsai/cudf/pull/20905 -* Remove FutureWarning checks from series/test_pct_change by @mroeschke in https://github.com/rapidsai/cudf/pull/20904 -* Enforce Series.replace argument deprecations by @mroeschke in https://github.com/rapidsai/cudf/pull/20903 -* Enforce pandas 3 DataFrame/Series.first/last removal by @mroeschke in https://github.com/rapidsai/cudf/pull/20909 -* Enforce pandas 3 fillna deprecation by @mroeschke in https://github.com/rapidsai/cudf/pull/20912 -* Enforce pandas 3 timedelta frequency string deprecation by @mroeschke in https://github.com/rapidsai/cudf/pull/20915 -* Enforce pandas 3 concat with empties deprecation by @mroeschke in https://github.com/rapidsai/cudf/pull/20916 -* Enforce pandas 3 `to_numeric(errors='ignore')` deperecation by @mroeschke in https://github.com/rapidsai/cudf/pull/20920 -* Enforce pandas 3 date_range(periods=) deprecation by @mroeschke in https://github.com/rapidsai/cudf/pull/20921 -* Enforce pandas 3 resample deprecations by @mroeschke in https://github.com/rapidsai/cudf/pull/20923 -* Enforce pandas 3 TimedeltaIndex/DatetimeIndex deprecations by @mroeschke in https://github.com/rapidsai/cudf/pull/20924 -* Enforce pandas 3 future_stack argument switch by @mroeschke in https://github.com/rapidsai/cudf/pull/20926 -* Remove pandas 3.0 Series.__getitem__ int positional deprecation by @mroeschke in https://github.com/rapidsai/cudf/pull/20930 -* Test pandas 3.0.0rc1 by @mroeschke in https://github.com/rapidsai/cudf/pull/20941 -* Change `to_datetime` to match pandas-3 by @galipremsagar in https://github.com/rapidsai/cudf/pull/20913 -* More misc cuDF classic test fixes for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21049 -* Switch default string dtype by @galipremsagar in https://github.com/rapidsai/cudf/pull/20890 -* Change CUDF_STRING_DTYPE to pandas.StringDtype for pandas 3.0 by @mroeschke in https://github.com/rapidsai/cudf/pull/21062 -* Adjust cuDF grouby unit tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21081 -* Fix StringDtype dask serialization in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21103 -* Avoid object dtype in ColumnAccessor.to_pandas_index for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21110 -* Move CategoricalDtype higher in cudf.dtype elif chain for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21129 -* Match groupby idxmin/max skipna=True pandas 3.0 behavior with all null groups by @mroeschke in https://github.com/rapidsai/cudf/pull/21189 -* Map numpy dtypes to new default string dtype in np_dtypes_to_pandas_dtypes for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21201 -* Fix test_string_char_types for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21210 -* Adjust more misc tests for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21228 -* Preserve StringDtype correctly in StringColumn.to_pandas for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21206 -* Update pandas pinnings in conda-forge recipes for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21302 -* Update cudf_pandas proxies for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21606 -* Change `date_range` time unit inferencing logic by @galipremsagar in https://github.com/rapidsai/cudf/pull/21694 -* Make copy-on-write the default for pandas3 by @vyasr in https://github.com/rapidsai/cudf/pull/21801 -* Fix multiple repr issues in `cudf` with `pandas-3.0` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21778 -* Fix `replace` API & pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21938 -* Fix `cudf.pandas` pytest failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/22137 -* Fix setitem dtype validation, NaT column inference, and boolean-dtype join by @galipremsagar in https://github.com/rapidsai/cudf/pull/22255 -* Reject pd.NA string-to-object casts in pandas-compatible mode by @galipremsagar in https://github.com/rapidsai/cudf/pull/22295 -* Align groupby apply/rolling defaults with pandas 3 by @galipremsagar in https://github.com/rapidsai/cudf/pull/22448 -* Align cudf find_common_type with pandas for empty/bool/datetime+timedelta inputs by @galipremsagar in https://github.com/rapidsai/cudf/pull/22466 -* Add validations in various code-paths by @galipremsagar in https://github.com/rapidsai/cudf/pull/22336 -* Merge `pandas3` feature branch into `main` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22518 -* Forward-merge release/26.06 into main by @madsbk in https://github.com/rapidsai/cudf/pull/22585 -* Remove deprecated `.values_host` property in cudf by @mroeschke in https://github.com/rapidsai/cudf/pull/22649 -* Remove deprecated nvtext::edit_distance_matrix by @davidwendt in https://github.com/rapidsai/cudf/pull/22644 -* Fix capture group handling in cudf::strings::findall by @davidwendt in https://github.com/rapidsai/cudf/pull/22408 -* Removed deprecated multi-pattern overload of replace_re by @davidwendt in https://github.com/rapidsai/cudf/pull/22639 -* Skip sum updates once SUM_WITH_OVERFLOW flag is set by @PointKernel in https://github.com/rapidsai/cudf/pull/22696 -* Remove deprecated strings::like string_scalar API by @davidwendt in https://github.com/rapidsai/cudf/pull/22678 -* Fix Categorical astype for nulls in pandas-compatible mode by @galipremsagar in https://github.com/rapidsai/cudf/pull/22652 -* [cudf_streaming] Using rapidsmpf spill unspill partition by @nirandaperera in https://github.com/rapidsai/cudf/pull/22810 -* Change behavior for regex class pattern [a-m-z] by @davidwendt in https://github.com/rapidsai/cudf/pull/22601 -* Fix groupby `any`/`all` on null-containing string columns by @galipremsagar in https://github.com/rapidsai/cudf/pull/22926 -* Align cudf indexing/setitem validation and alignment with pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/22912 -* refactor(streaming): flatten namespaces and rename to snake_case by @vyasr in https://github.com/rapidsai/cudf/pull/22909 -* [cudf_streaming] rapidsmpf backrefmixin API changes by @nirandaperera in https://github.com/rapidsai/cudf/pull/22911 -* Refactor packed metadata to use an explicit table header by @madsbk in https://github.com/rapidsai/cudf/pull/22951 -* Support `skipna` in groupby reductions (first/last, sum/prod/mean/median/min/max, idxmin/idxmax) by @galipremsagar in https://github.com/rapidsai/cudf/pull/22925 -* Make `pylibcudf.Table.columns()` return a tuple by @madsbk in https://github.com/rapidsai/cudf/pull/23040 -* Align DataFrame.reindex dtype and validation behavior with pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/22914 -* Match pandas suffix handling in merge (None/non-string suffixes, invalid-container TypeError) by @galipremsagar in https://github.com/rapidsai/cudf/pull/23059 -* Rapidsmpf logger API changes by @nirandaperera in https://github.com/rapidsai/cudf/pull/22972 -* Refactor ``OrderScheme`` to track multiple ``Ordering`` descriptions by @rjzamora in https://github.com/rapidsai/cudf/pull/22896 -* Return null for non-participating regex capture groups in strings::extract by @galipremsagar in https://github.com/rapidsai/cudf/pull/23123 -* wheels: build CUDA 13 wheels with latest CTK (13.3.0) by @jameslamb in https://github.com/rapidsai/cudf/pull/23174 -* Support explicit row counts for zero-column tables by @madsbk in https://github.com/rapidsai/cudf/pull/22765 -* Fix MultiIndex indexing/.loc to match pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/22966 -* Match pandas merge/join dtype, index and error semantics by @galipremsagar in https://github.com/rapidsai/cudf/pull/23060 -* Rapidsmpf backref API changes for Host and Pinned MR by @nirandaperera in https://github.com/rapidsai/cudf/pull/23128 -* Rewrite mixed inner/left/full join with post-filtering by @PointKernel in https://github.com/rapidsai/cudf/pull/23012 -* Fix DataFrame.stack level resolution, ordering, and dtype preservation by @galipremsagar in https://github.com/rapidsai/cudf/pull/23370 -* Rework approach to cudf-streaming bloom filter sizing by @wence- in https://github.com/rapidsai/cudf/pull/23067 -* Match pandas ordering, dtypes, and validation in unstack/_pivot by @galipremsagar in https://github.com/rapidsai/cudf/pull/23368 -* JNI bridge for the retention filter support for deletion vectors by @jihoonson in https://github.com/rapidsai/cudf/pull/23423 +* Enforce pandas 3.0 deprecations and subsequent related cuDF deprecations part 1 by @mroeschke in https://github.com/NVIDIA/cudf/pull/20823 +* Enforce Series.interpolate deprecations by @mroeschke in https://github.com/NVIDIA/cudf/pull/20905 +* Remove FutureWarning checks from series/test_pct_change by @mroeschke in https://github.com/NVIDIA/cudf/pull/20904 +* Enforce Series.replace argument deprecations by @mroeschke in https://github.com/NVIDIA/cudf/pull/20903 +* Enforce pandas 3 DataFrame/Series.first/last removal by @mroeschke in https://github.com/NVIDIA/cudf/pull/20909 +* Enforce pandas 3 fillna deprecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/20912 +* Enforce pandas 3 timedelta frequency string deprecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/20915 +* Enforce pandas 3 concat with empties deprecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/20916 +* Enforce pandas 3 `to_numeric(errors='ignore')` deperecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/20920 +* Enforce pandas 3 date_range(periods=) deprecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/20921 +* Enforce pandas 3 resample deprecations by @mroeschke in https://github.com/NVIDIA/cudf/pull/20923 +* Enforce pandas 3 TimedeltaIndex/DatetimeIndex deprecations by @mroeschke in https://github.com/NVIDIA/cudf/pull/20924 +* Enforce pandas 3 future_stack argument switch by @mroeschke in https://github.com/NVIDIA/cudf/pull/20926 +* Remove pandas 3.0 Series.__getitem__ int positional deprecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/20930 +* Test pandas 3.0.0rc1 by @mroeschke in https://github.com/NVIDIA/cudf/pull/20941 +* Change `to_datetime` to match pandas-3 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20913 +* More misc cuDF classic test fixes for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21049 +* Switch default string dtype by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20890 +* Change CUDF_STRING_DTYPE to pandas.StringDtype for pandas 3.0 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21062 +* Adjust cuDF grouby unit tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21081 +* Fix StringDtype dask serialization in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21103 +* Avoid object dtype in ColumnAccessor.to_pandas_index for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21110 +* Move CategoricalDtype higher in cudf.dtype elif chain for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21129 +* Match groupby idxmin/max skipna=True pandas 3.0 behavior with all null groups by @mroeschke in https://github.com/NVIDIA/cudf/pull/21189 +* Map numpy dtypes to new default string dtype in np_dtypes_to_pandas_dtypes for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21201 +* Fix test_string_char_types for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21210 +* Adjust more misc tests for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21228 +* Preserve StringDtype correctly in StringColumn.to_pandas for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21206 +* Update pandas pinnings in conda-forge recipes for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21302 +* Update cudf_pandas proxies for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21606 +* Change `date_range` time unit inferencing logic by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21694 +* Make copy-on-write the default for pandas3 by @vyasr in https://github.com/NVIDIA/cudf/pull/21801 +* Fix multiple repr issues in `cudf` with `pandas-3.0` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21778 +* Fix `replace` API & pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21938 +* Fix `cudf.pandas` pytest failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22137 +* Fix setitem dtype validation, NaT column inference, and boolean-dtype join by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22255 +* Reject pd.NA string-to-object casts in pandas-compatible mode by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22295 +* Align groupby apply/rolling defaults with pandas 3 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22448 +* Align cudf find_common_type with pandas for empty/bool/datetime+timedelta inputs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22466 +* Add validations in various code-paths by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22336 +* Merge `pandas3` feature branch into `main` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22518 +* Forward-merge release/26.06 into main by @madsbk in https://github.com/NVIDIA/cudf/pull/22585 +* Remove deprecated `.values_host` property in cudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/22649 +* Remove deprecated nvtext::edit_distance_matrix by @davidwendt in https://github.com/NVIDIA/cudf/pull/22644 +* Fix capture group handling in cudf::strings::findall by @davidwendt in https://github.com/NVIDIA/cudf/pull/22408 +* Removed deprecated multi-pattern overload of replace_re by @davidwendt in https://github.com/NVIDIA/cudf/pull/22639 +* Skip sum updates once SUM_WITH_OVERFLOW flag is set by @PointKernel in https://github.com/NVIDIA/cudf/pull/22696 +* Remove deprecated strings::like string_scalar API by @davidwendt in https://github.com/NVIDIA/cudf/pull/22678 +* Fix Categorical astype for nulls in pandas-compatible mode by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22652 +* [cudf_streaming] Using rapidsmpf spill unspill partition by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22810 +* Change behavior for regex class pattern [a-m-z] by @davidwendt in https://github.com/NVIDIA/cudf/pull/22601 +* Fix groupby `any`/`all` on null-containing string columns by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22926 +* Align cudf indexing/setitem validation and alignment with pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22912 +* refactor(streaming): flatten namespaces and rename to snake_case by @vyasr in https://github.com/NVIDIA/cudf/pull/22909 +* [cudf_streaming] rapidsmpf backrefmixin API changes by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22911 +* Refactor packed metadata to use an explicit table header by @madsbk in https://github.com/NVIDIA/cudf/pull/22951 +* Support `skipna` in groupby reductions (first/last, sum/prod/mean/median/min/max, idxmin/idxmax) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22925 +* Make `pylibcudf.Table.columns()` return a tuple by @madsbk in https://github.com/NVIDIA/cudf/pull/23040 +* Align DataFrame.reindex dtype and validation behavior with pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22914 +* Match pandas suffix handling in merge (None/non-string suffixes, invalid-container TypeError) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23059 +* Rapidsmpf logger API changes by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22972 +* Refactor ``OrderScheme`` to track multiple ``Ordering`` descriptions by @rjzamora in https://github.com/NVIDIA/cudf/pull/22896 +* Return null for non-participating regex capture groups in strings::extract by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23123 +* wheels: build CUDA 13 wheels with latest CTK (13.3.0) by @jameslamb in https://github.com/NVIDIA/cudf/pull/23174 +* Support explicit row counts for zero-column tables by @madsbk in https://github.com/NVIDIA/cudf/pull/22765 +* Fix MultiIndex indexing/.loc to match pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22966 +* Match pandas merge/join dtype, index and error semantics by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23060 +* Rapidsmpf backref API changes for Host and Pinned MR by @nirandaperera in https://github.com/NVIDIA/cudf/pull/23128 +* Rewrite mixed inner/left/full join with post-filtering by @PointKernel in https://github.com/NVIDIA/cudf/pull/23012 +* Fix DataFrame.stack level resolution, ordering, and dtype preservation by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23370 +* Rework approach to cudf-streaming bloom filter sizing by @wence- in https://github.com/NVIDIA/cudf/pull/23067 +* Match pandas ordering, dtypes, and validation in unstack/_pivot by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23368 +* JNI bridge for the retention filter support for deletion vectors by @jihoonson in https://github.com/NVIDIA/cudf/pull/23423 ### 🐛 Bug Fixes -* Add back Index.fillna by @mroeschke in https://github.com/rapidsai/cudf/pull/21002 -* Account of pandas.NamedAgg subclass change in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21023 -* Account for non-fixed `pandas.offsets.Day` in `date_range` for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21036 -* Fix `columns` validation logic of DataFrame.from_records by @mroeschke in https://github.com/rapidsai/cudf/pull/21034 -* Unify `_is_null_host_scalar` into `is_na_like` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21079 -* Fix datetime binops issues in `pandas-3.0` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21172 -* Fix convert_dtypes with new StringDtype in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21200 -* Preserve StringDtype variants from pandas inputs for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21208 -* Re-enable `dtype=category` in pandas-3.0 branch by @galipremsagar in https://github.com/rapidsai/cudf/pull/21604 -* Fix all string type bugs by @galipremsagar in https://github.com/rapidsai/cudf/pull/21609 -* Fix astype `CategoricalDtype` cases. by @galipremsagar in https://github.com/rapidsai/cudf/pull/21627 -* Fix column type for empty columns by @galipremsagar in https://github.com/rapidsai/cudf/pull/21638 -* Fix interval_range type inference bugs by @galipremsagar in https://github.com/rapidsai/cudf/pull/21686 -* Fix `Index.intersection` for categorical types by @galipremsagar in https://github.com/rapidsai/cudf/pull/21723 -* Fix string udf failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21727 -* Fix join bug that shows up in `Index.difference` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21733 -* Fix all binops failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21738 -* Use reduction_axis for dataframes by @vyasr in https://github.com/rapidsai/cudf/pull/21743 -* Fix `astype(category)` for empty column scenarios by @galipremsagar in https://github.com/rapidsai/cudf/pull/21741 -* Fix all explode failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21766 -* Fix orc pytest failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21847 -* Fix failures in `rolling` pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21846 -* Fix all `Series.astype` bugs by @galipremsagar in https://github.com/rapidsai/cudf/pull/21860 -* Fix numpy `ufunc` return type by @galipremsagar in https://github.com/rapidsai/cudf/pull/21876 -* Fix dtype issues in `test_between.py` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21887 -* Fix `isin` test failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21886 -* Fix `string.to_pandas` for `nullable` & `arrow_type` cases by @galipremsagar in https://github.com/rapidsai/cudf/pull/21890 -* Fix `concat` & `union` bugs by @galipremsagar in https://github.com/rapidsai/cudf/pull/21889 -* Fix all `test_iloc.py` failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21872 -* Fix all failures in `test_constructors.py` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21905 -* Fix all string related failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21939 -* Fix Parquet pytest failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21967 -* Switch default time resolutions for `DatetimeIndex` & `TimedeltaIndex` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21997 -* Fix `add_prefix` & `add_suffix` pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/22000 -* Fix all json test failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21998 -* Fix runtime warnings in `apply` pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/22017 -* Fix string column `to_arrow` to generate large strings when needed by @galipremsagar in https://github.com/rapidsai/cudf/pull/21891 -* Fix type inconsistencies in pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/22016 -* Fix `replace` pytest by @galipremsagar in https://github.com/rapidsai/cudf/pull/22019 -* Fix parquet metadata failures in pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/22020 -* Fix `copy-on-write` failures related to `astype` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22022 -* Fix `str` reductions to match pandas3 by @galipremsagar in https://github.com/rapidsai/cudf/pull/22026 -* Fix `right-join` column ordering by @galipremsagar in https://github.com/rapidsai/cudf/pull/22002 -* [pandas3] Fix test_categorical_series_with_nan_repr for pandas 3 nan_as_null change by @galipremsagar in https://github.com/rapidsai/cudf/pull/22066 -* [pandas3] Fix test_diff_many_dtypes for boolean Series diff null representation by @galipremsagar in https://github.com/rapidsai/cudf/pull/22065 -* [pandas3] Fix test_timedelta_astype_unicode_dtype_pandas_compat for pandas 3 StringDtype by @galipremsagar in https://github.com/rapidsai/cudf/pull/22064 -* [pandas3] Fix test_categorical_empty: update categories dtype assertion by @galipremsagar in https://github.com/rapidsai/cudf/pull/22062 -* [pandas3] Fix test_rename_shallow_copy: use np.shares_memory instead of identity check by @galipremsagar in https://github.com/rapidsai/cudf/pull/22060 -* [pandas3] Fix test_groupby_pct_change_empty_columns for pandas 3 dtype inference by @galipremsagar in https://github.com/rapidsai/cudf/pull/22059 -* [pandas3] Fix two regressions in Series.describe() by @galipremsagar in https://github.com/rapidsai/cudf/pull/22058 -* [pandas3] Fix test_alltypes_plain_avro timestamp conversion for pandas 3 by @galipremsagar in https://github.com/rapidsai/cudf/pull/22061 -* [pandas3] Add _CachedAccessor to replace removed pandas.core.accessor.CachedAccessor by @galipremsagar in https://github.com/rapidsai/cudf/pull/22056 -* Update memory usage test by @galipremsagar in https://github.com/rapidsai/cudf/pull/22080 -* [pandas3] Fix float-category NaN handling in CategoricalColumn.to_pandas() by @galipremsagar in https://github.com/rapidsai/cudf/pull/22057 -* Skip pandas failure cases in `ewm` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22082 -* Fix `sort_values` when `return_indexer=True` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22098 -* Fix pandas 3 compatibility bugs in index, groupby, and dask_cudf tests by @galipremsagar in https://github.com/rapidsai/cudf/pull/22101 -* Make `cudf.pandas` pandas tests able to run by @galipremsagar in https://github.com/rapidsai/cudf/pull/22140 -* Fix `12k` pandas test suite failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/22158 -* Allow SeriesApplyKernel to support pd.StringDtype in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22177 -* Fix `datetime` & `timedelta` failures in pandas test suite by @galipremsagar in https://github.com/rapidsai/cudf/pull/22174 -* Fix return types for `rank` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22180 -* Fix regression in `eval` frame fetching by @galipremsagar in https://github.com/rapidsai/cudf/pull/22221 -* Propagate writes to internal or private properties to slow objects only by @galipremsagar in https://github.com/rapidsai/cudf/pull/22231 -* Fix cudf.pandas view semantics across fast/slow conversions and add __array__ to datetimelike ExtensionArray proxies by @galipremsagar in https://github.com/rapidsai/cudf/pull/22259 -* Fix cudf.pandas pickle round-trips for proxy classes, DatetimeTZ columns, and NaT by @galipremsagar in https://github.com/rapidsai/cudf/pull/22297 -* Add proxy types for missing offsets by @galipremsagar in https://github.com/rapidsai/cudf/pull/22314 -* Fix failures in masked array pandas unit tests under cudf.pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/22296 -* Fix `StataReader` proxy type by @galipremsagar in https://github.com/rapidsai/cudf/pull/22322 -* Fix `MultiIndex` getitem by @galipremsagar in https://github.com/rapidsai/cudf/pull/22326 -* Define attributes that are not discoverable by `dir(cls)` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22330 -* Drop invalidated frequency in `DatetimeIndex` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22328 -* Fix `to_period` parameter by @galipremsagar in https://github.com/rapidsai/cudf/pull/22329 -* Expose __from_arrow__ on masked extension dtype proxies by @galipremsagar in https://github.com/rapidsai/cudf/pull/22373 -* Fix datetime resolution for empty data inputs by @galipremsagar in https://github.com/rapidsai/cudf/pull/22363 -* Expose additional proxy attributes for IntervalArray, Styler, and _MethodProxy by @galipremsagar in https://github.com/rapidsai/cudf/pull/22374 -* Fix `tz` aware binops by @galipremsagar in https://github.com/rapidsai/cudf/pull/22243 -* Implement min_count for groupby reductions by @galipremsagar in https://github.com/rapidsai/cudf/pull/22372 -* Fix datetime tz + categorical creation by @galipremsagar in https://github.com/rapidsai/cudf/pull/22278 -* Expose DateOffset instance attributes on cudf.pandas proxy by @galipremsagar in https://github.com/rapidsai/cudf/pull/22447 -* Fix np.ufunc(Index, Series) dispatch and np.matmul scalar return type by @galipremsagar in https://github.com/rapidsai/cudf/pull/22444 -* Fix `fillna` in `binop` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22442 -* Fix `is_numeric_dtype` for `Index` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22438 -* Preserve StringDtype storage and na_value in get_dtype_of_same_kind by @galipremsagar in https://github.com/rapidsai/cudf/pull/22289 -* Raise TypeError for numeric column arithmetic with pd.NaT by @galipremsagar in https://github.com/rapidsai/cudf/pull/22451 -* Make StringArray proxy inherit from NumpyExtensionArray by @galipremsagar in https://github.com/rapidsai/cudf/pull/22457 -* Fix return types of `quantile` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22433 -* Preserve extension dtypes in groupby reductions and exclude grouping-key columns by identity by @galipremsagar in https://github.com/rapidsai/cudf/pull/22369 -* Fix several to_datetime edge cases for pandas 3 compatibility by @galipremsagar in https://github.com/rapidsai/cudf/pull/22446 -* Fix arrow types related failures in `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22194 -* Implement groupby all/any via bool-coercion + min/max by @galipremsagar in https://github.com/rapidsai/cudf/pull/22371 -* Fix `AssertionError: DataFrame.columns are different` failures in cudf.pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/22351 -* Fix factorize for pandas 3 compatibility by @galipremsagar in https://github.com/rapidsai/cudf/pull/22445 -* Fix empty series dtype by @galipremsagar in https://github.com/rapidsai/cudf/pull/22443 -* Implement groupby sum on StringDtype columns as per-group concatenation by @galipremsagar in https://github.com/rapidsai/cudf/pull/22370 -* Preserve pandas-nullable dtypes through ufuncs and empty value_counts by @galipremsagar in https://github.com/rapidsai/cudf/pull/22461 -* Fix all `datetime` type inconsistencies with extension types by @galipremsagar in https://github.com/rapidsai/cudf/pull/22409 -* Serialize engine config in new pdsh benchmark CLI by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22365 -* Fix unsafe, missing close() of HostColumnVectors by @mythrocks in https://github.com/rapidsai/cudf/pull/22432 -* Skip flaky test_make_spill_function on pinned-memory allocation failure by @Matt711 in https://github.com/rapidsai/cudf/pull/22551 -* Fix JSON token count estimate to account for the delimiter in recovery mode by @shrshi in https://github.com/rapidsai/cudf/pull/22559 -* Fix AST expression evaluator handling of intermediate values for fixed-point types by @davidwendt in https://github.com/rapidsai/cudf/pull/22512 -* Remove __arrow_array__ from NumPy ndarray cudf.pandas proxy by @galipremsagar in https://github.com/rapidsai/cudf/pull/22617 -* Expose per-column had_schema_mismatch diagnostic from JSON reader by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22450 -* Fix Arrow device string view variadic buffers by @0ax1 in https://github.com/rapidsai/cudf/pull/22620 -* [BUG] Parquet reader: treat LIST-annotated REPEATED group as the list, not a stub by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22597 -* Fix invalid use of data_type constructor in row_ir logic by @davidwendt in https://github.com/rapidsai/cudf/pull/22676 -* Preserve single-field unannotated repeated groups in parquet schema sanitization by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22567 -* Expose `__array_ufunc__` on NumpyExtensionArray cudf.pandas proxy by @galipremsagar in https://github.com/rapidsai/cudf/pull/22607 -* Expose __from_arrow__ on DatetimeTZDtype cudf.pandas proxy by @galipremsagar in https://github.com/rapidsai/cudf/pull/22608 -* Fix DataFrame mask Series condition alignment by @aryansri05 in https://github.com/rapidsai/cudf/pull/22701 -* fix(cudf.pandas): fix Excel/openpyxl test failures and add xlsxwriter dependency by @vyasr in https://github.com/rapidsai/cudf/pull/22722 -* Fix to_numpy object null handling by @aryansri05 in https://github.com/rapidsai/cudf/pull/22621 -* Match pandas dtype semantics in groupby quantile by @galipremsagar in https://github.com/rapidsai/cudf/pull/22750 -* Make `IntervalDtype` construction pandas-compatible by @galipremsagar in https://github.com/rapidsai/cudf/pull/22731 -* Remove cudf-polars CI timeouts by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22768 -* Fix scalar as column view for null scalars by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22773 -* Increase timeouts for slower cudf-polars tests by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22780 -* Enable chunked row masks in hybrid scan reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22716 -* Fix Parquet page header decoder to allow empty dict pages by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22777 -* declare nvidia-nvjitlink dependency in 'libcudf' wheels by @jameslamb in https://github.com/rapidsai/cudf/pull/22785 -* Fix replace_with_backrefs copy remainder logic for empty pattern match by @davidwendt in https://github.com/rapidsai/cudf/pull/22711 -* Fix ArrowDtype integer arithmetic to match pandas pyarrow semantics by @galipremsagar in https://github.com/rapidsai/cudf/pull/22633 -* Fix zero-count regex quantifier expansion by @fallintoplace in https://github.com/rapidsai/cudf/pull/22800 -* Build cudf_streaming wheel against this CI run's pylibcudf by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22819 -* Fix groupby.apply columns axis name for row-like UDF results by @galipremsagar in https://github.com/rapidsai/cudf/pull/22809 -* Use (un)spill_partitions from RapidsMPF by @pentschev in https://github.com/rapidsai/cudf/pull/22837 -* Fix cudf.pandas DateOffset equality and empty-frame column dtype preservation by @galipremsagar in https://github.com/rapidsai/cudf/pull/22806 -* Fix `MultiIndex.join` to join on all common levels by @galipremsagar in https://github.com/rapidsai/cudf/pull/22827 -* Let NVRTC choose automatic PCH directory by @bdice in https://github.com/rapidsai/cudf/pull/22812 -* [cudf-polars] Sort ORDER BY keys by column, not alias by @madsbk in https://github.com/rapidsai/cudf/pull/22781 -* Fix cudf_polars benchmarking summarizing mean times by @mroeschke in https://github.com/rapidsai/cudf/pull/22845 -* Fix masked dtype equality in cudf.pandas so real_dtype == proxy_dtype holds by @galipremsagar in https://github.com/rapidsai/cudf/pull/22842 -* Reduce size of input dataframe in test_parallel_scan by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22835 -* Fix findall logic handling for empty match regex patterns by @davidwendt in https://github.com/rapidsai/cudf/pull/22714 -* Reduce setup runtime of multi_orderby_range rolling benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/22869 -* Skip polars tests relying on newer sqlite features by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22870 -* Fix ARROW_FLAG_NULLABLE flag usage in cudf interop by @davidwendt in https://github.com/rapidsai/cudf/pull/22828 -* Pin pytest-cases version to fix nightly build errors by @davidwendt in https://github.com/rapidsai/cudf/pull/22888 -* Drop dynamic predicates from logical plan by @Matt711 in https://github.com/rapidsai/cudf/pull/22820 -* Fix group-key columns in `SeriesGroupBy` list-`agg` and `any`/`all` with `as_index=False` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22813 -* Preserve object dtype for string columns when constructing cudf DataFrame from pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/22831 -* Reduce concurrency of cudf-polars tests by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22894 -* Fix groupby and single-column `to_numpy` pandas-compatibility bugs by @galipremsagar in https://github.com/rapidsai/cudf/pull/22783 -* Preserve categorical column index in `DataFrame.describe` and handle pure iterators in `cudf.pandas` fallback by @galipremsagar in https://github.com/rapidsai/cudf/pull/22821 -* [BUG] Fix conditional linking of RTCX embed dependencies & Allow null fragment names by @lamarrr in https://github.com/rapidsai/cudf/pull/22898 -* Allow IR fallback to handle an ErrorNode by @mroeschke in https://github.com/rapidsai/cudf/pull/22855 -* Fix type annotations in pylibucdf hybrid scan by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22118 -* Destroy RTCX cache before teardown by @fallintoplace in https://github.com/rapidsai/cudf/pull/22843 -* `create_string_column` bug fix by @abigalekim in https://github.com/rapidsai/cudf/pull/22942 -* Fix categorical replace no-op remapping by @anthonyguerrera in https://github.com/rapidsai/cudf/pull/22436 -* Enable cudf.pandas `test_numpy_ufuncs_basic` via tolerant Index comparison by @galipremsagar in https://github.com/rapidsai/cudf/pull/22924 -* Make `to_string` match pandas (drop repr-style truncation/footer) by @galipremsagar in https://github.com/rapidsai/cudf/pull/22929 -* Fix cudf.to_numeric to match pandas semantics by @galipremsagar in https://github.com/rapidsai/cudf/pull/22934 -* Protect against integer overflow in ``groupby_actor`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22950 -* Skip code coverage where dynamic predicates are dropped by @Matt711 in https://github.com/rapidsai/cudf/pull/22973 -* Work around pola-rs/polars#23214 in streaming dataframe scan by @wence- in https://github.com/rapidsai/cudf/pull/23007 -* Fix `cudf.pandas --line-profile` clobbering `__file__` by @galipremsagar in https://github.com/rapidsai/cudf/pull/23017 -* Pin pandas to exclude 3.0.4 (segfault constructing `pd.Timedelta` on Python 3.14) by @galipremsagar in https://github.com/rapidsai/cudf/pull/23022 -* Fix stream ordering bug in join of expression-based keys by @wence- in https://github.com/rapidsai/cudf/pull/23024 -* Fix `DataFrame` and `GroupBy.value_counts` to match pandas ordering, `dropna`, and categorical handling by @galipremsagar in https://github.com/rapidsai/cudf/pull/22903 -* Fix ``Distinct`` expression lowering by @rjzamora in https://github.com/rapidsai/cudf/pull/22970 -* Add bounds checking for AST column_reference out-of-range indices by @utkarshparekh in https://github.com/rapidsai/cudf/pull/23021 -* Implement `DataFrame.count(axis=1)` on the GPU by @galipremsagar in https://github.com/rapidsai/cudf/pull/23016 -* Fix ptr null check in packed_data_from_cudf_packed_columns by @Matt711 in https://github.com/rapidsai/cudf/pull/23038 -* Fix cudf-java native_jstring to emit standard UTF-8 for supplementary characters and embedded NULs by @igorpeshansky in https://github.com/rapidsai/cudf/pull/23018 -* Fix cudf.pandas Excel reader failures (empty-column dtype, string offset width, proxy monkeypatch) by @galipremsagar in https://github.com/rapidsai/cudf/pull/22927 -* Keep Nanoarrow ArrowArrayStream chunks alive during cudaMemcpyBatchAsync by @mroeschke in https://github.com/rapidsai/cudf/pull/23009 -* Support host buffers (`BytesIO`) in `cudf.read_text` by @nethum529 in https://github.com/rapidsai/cudf/pull/23032 -* Fully initialize PageInfo before sort_pages copies it by @pramodsatya in https://github.com/rapidsai/cudf/pull/22957 -* Declare return type on device lambdas queried in host code by @PointKernel in https://github.com/rapidsai/cudf/pull/23046 -* Add exception for unsupported decimal128 AST operations by @davidwendt in https://github.com/rapidsai/cudf/pull/22982 -* Fix `explain()` failing on `UnaryFunction` with multiple children by @Matt711 in https://github.com/rapidsai/cudf/pull/23050 -* Raise parquet partitions_info validation errors by @fallintoplace in https://github.com/rapidsai/cudf/pull/22801 -* Re-derive cudf.pandas intermediate proxies when their parents change by @galipremsagar in https://github.com/rapidsai/cudf/pull/22998 -* Add Numba pin to cuDF wheel build requirements by @bdice in https://github.com/rapidsai/cudf/pull/23084 -* Return np.intp from argsort to match numpy/pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/23061 -* Fix pinned-host spill reservations for packed table alignment by @pentschev in https://github.com/rapidsai/cudf/pull/23088 -* Enable cudf_streaming distributed (MPI/UCXX) tests in CI by @nirandaperera in https://github.com/rapidsai/cudf/pull/22954 -* Fix DataFrame.rename for an unnamed MultiIndex level by @galipremsagar in https://github.com/rapidsai/cudf/pull/23062 -* Fix rtcx embed target-level dependencies under the Makefiles generator by @galipremsagar in https://github.com/rapidsai/cudf/pull/23106 -* fix(pylibcudf): accept inf/-inf as valid FLOAT32 scalar values by @nethum529 in https://github.com/rapidsai/cudf/pull/23099 -* Match pandas semantics in groupby.rolling and groupby.apply result construction by @galipremsagar in https://github.com/rapidsai/cudf/pull/23122 -* Hybrid scan reports correct number of input RGs by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/23052 -* Fix uninitialized read for empty segments in segmented top-k by @ttnghia in https://github.com/rapidsai/cudf/pull/23109 -* Aggregate category values instead of codes in window operations by @galipremsagar in https://github.com/rapidsai/cudf/pull/23120 -* Sum interleave benchmark byte counts across all columns by @fallintoplace in https://github.com/rapidsai/cudf/pull/23103 -* Fix type annotation for CachedParquetInfo.size by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23142 -* Raise TypeError for NaT in where() on masked dtypes by @galipremsagar in https://github.com/rapidsai/cudf/pull/23107 -* Fix ``Scan`` projection with pandas index by @rjzamora in https://github.com/rapidsai/cudf/pull/23135 -* Match pandas semantics in DataFrame.select_dtypes by @galipremsagar in https://github.com/rapidsai/cudf/pull/23116 -* Fix to_datetime nanosecond inference and DatetimeIndex tz wall-time interpretation by @galipremsagar in https://github.com/rapidsai/cudf/pull/23141 -* Return value_counts groups in order of first appearance by @galipremsagar in https://github.com/rapidsai/cudf/pull/23140 -* Implement factorize for MultiIndex by @galipremsagar in https://github.com/rapidsai/cudf/pull/23139 -* Fix cudf.pandas datetimelike plotting failures (canonical freq offsets + Period/offset pickling) by @galipremsagar in https://github.com/rapidsai/cudf/pull/23054 -* Use `rapids_test_add()` again by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/23181 -* Add error checking to format inference by @vyasr in https://github.com/rapidsai/cudf/pull/23172 -* Set `RUN_SERIAL` property on streaming tests by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/23188 -* Handle Arrow bool masks in DataFrame getitem by @dttutty in https://github.com/rapidsai/cudf/pull/23195 -* Supply stream and mr to column synthesizers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/23209 -* Fix out-of-bounds access in segmented top-k with partial offsets by @ttnghia in https://github.com/rapidsai/cudf/pull/23105 -* Remove duplicate return in ndsh apply_inner_join by @fallintoplace in https://github.com/rapidsai/cudf/pull/23104 -* Fix cuco::extent type for comparator_helper key_set in distinct_count by @davidwendt in https://github.com/rapidsai/cudf/pull/23236 -* Fix string reductions and isin across string dtype flavors by @galipremsagar in https://github.com/rapidsai/cudf/pull/23214 -* fix(pre-commit): normalize SPDX handling by @bdice in https://github.com/rapidsai/cudf/pull/23228 -* Fix cudf.pandas arithmetic test failures (pow null identities, RangeIndex range-preserving binops, Index proxy array priority) by @galipremsagar in https://github.com/rapidsai/cudf/pull/23176 -* Fix benchmark generate_input create_ascii_string_column function signature by @davidwendt in https://github.com/rapidsai/cudf/pull/23239 -* Raise OutOfBoundsDatetime for datetime conversions that exceed the target unit's range by @galipremsagar in https://github.com/rapidsai/cudf/pull/23216 -* Fix cudf.pandas series accessor failures across numpy/nullable/arrow dtypes by @galipremsagar in https://github.com/rapidsai/cudf/pull/23056 -* Fix Series.isin dtype and value semantics for masked (nullable) dtypes by @galipremsagar in https://github.com/rapidsai/cudf/pull/23057 -* Replace thrust::reduce_by_key with cub::DeviceReduce::ReduceByKey in tdigest generation by @PointKernel in https://github.com/rapidsai/cudf/pull/23204 -* Change cuco::extent type usage from size_type to size_t by @davidwendt in https://github.com/rapidsai/cudf/pull/23259 -* Fix DataFrame.quantile and Series.quantile to match pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/23058 -* Fix data_type ctor calls in the row_ir get_return_type function by @davidwendt in https://github.com/rapidsai/cudf/pull/23237 -* Fix groupby size dtype for masked series and apply group order for sort=False by @galipremsagar in https://github.com/rapidsai/cudf/pull/23260 -* Rewrite GroupBy.nth as a pandas-compatible positional row filter by @galipremsagar in https://github.com/rapidsai/cudf/pull/23257 -* Keep the dtype family in GroupBy.rank results by @galipremsagar in https://github.com/rapidsai/cudf/pull/23258 -* Fix stale num_rows argument breaking mypy in cudf-polars duplicated-output path by @galipremsagar in https://github.com/rapidsai/cudf/pull/23303 -* Pin pyarrow<24 in the cudf and pylibcudf conda recipes by @galipremsagar in https://github.com/rapidsai/cudf/pull/23319 -* Fix cudf.pandas isinstance checks for unproxied pandas subclasses and wrap groupby resamplers by @galipremsagar in https://github.com/rapidsai/cudf/pull/23273 -* Fix GroupBy.apply result assembly, UDF closure side effects, and empty-frame dtypes by @galipremsagar in https://github.com/rapidsai/cudf/pull/23272 -* Avoid leaving multi-rank execution in a bad state in cudf-polars tests by @wence- in https://github.com/rapidsai/cudf/pull/23340 -* Reconcile Index.union dtypes before the empty-operand short-circuits by @galipremsagar in https://github.com/rapidsai/cudf/pull/23318 -* Support the limit parameter in GroupBy.ffill and bfill by @galipremsagar in https://github.com/rapidsai/cudf/pull/23302 -* Stop forward-filling in GroupBy.pct_change by @galipremsagar in https://github.com/rapidsai/cudf/pull/23301 -* Return index labels from GroupBy idxmin/idxmax in agg and transform by @galipremsagar in https://github.com/rapidsai/cudf/pull/23298 -* Match pandas semantics for null reduction results and mode ordering by @galipremsagar in https://github.com/rapidsai/cudf/pull/23328 -* Fix incorrect DECIMAL128 groupby sums on Blackwell with CUDA 13.0 by @PointKernel in https://github.com/rapidsai/cudf/pull/23229 -* Fix null_precedence access in preprocessed_table::create by @davidwendt in https://github.com/rapidsai/cudf/pull/23238 -* Route GroupBy.transform size/cumcount/ngroup to their group-level implementations by @galipremsagar in https://github.com/rapidsai/cudf/pull/23297 -* Recognize named-aggregation lambdas as scans in GroupBy by @galipremsagar in https://github.com/rapidsai/cudf/pull/23300 -* Add timeout stacktrace utility for cudf-polars tests, remove pytest-timeout by @wence- in https://github.com/rapidsai/cudf/pull/23332 -* Stop mutating the source column dtype in equivalent-type numerical casts by @galipremsagar in https://github.com/rapidsai/cudf/pull/23364 -* Mirror cudf.pandas class-level monkeypatches onto the real type by @galipremsagar in https://github.com/rapidsai/cudf/pull/23001 -* Restore the original index in GroupBy transform broadcasts and ngroup by @galipremsagar in https://github.com/rapidsai/cudf/pull/23296 -* Honor level and sort_remaining in sort_index(axis=1) by @galipremsagar in https://github.com/rapidsai/cudf/pull/23367 -* Fix lazy quantifier priority handling in Glushkov regex engine by @davidwendt in https://github.com/rapidsai/cudf/pull/23381 -* Keep MultiIndex columns in GroupBy.agg for MultiIndex-column sources by @galipremsagar in https://github.com/rapidsai/cudf/pull/23366 -* Preserve masked integer dtypes in GroupBy cumsum/cumprod by @galipremsagar in https://github.com/rapidsai/cudf/pull/23299 -* Use new rapidsmpf utilities to handle cancellation in bloom filter awaitables by @wence- in https://github.com/rapidsai/cudf/pull/23293 -* RMM Library Linking Error Bug Fix by @abigalekim in https://github.com/rapidsai/cudf/pull/23341 -* Fix Parquet stats pinned host copy lifetime by @aryansri05 in https://github.com/rapidsai/cudf/pull/23219 -* Add cuFile wheel dependencies for Python tests by @bdice in https://github.com/rapidsai/cudf/pull/23392 -* Include sources.jar and javadoc.jar in the Java Maven repo output by @paul-aiyedun in https://github.com/rapidsai/cudf/pull/23389 -* Preserve MultiIndex column fidelity through ColumnAccessor round trips by @galipremsagar in https://github.com/rapidsai/cudf/pull/23365 -* [BUG] Fix ORC chunked writer root row statistics by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/23118 -* Fix flaky cudf-polars Quent integration test by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23482 +* Add back Index.fillna by @mroeschke in https://github.com/NVIDIA/cudf/pull/21002 +* Account of pandas.NamedAgg subclass change in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21023 +* Account for non-fixed `pandas.offsets.Day` in `date_range` for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21036 +* Fix `columns` validation logic of DataFrame.from_records by @mroeschke in https://github.com/NVIDIA/cudf/pull/21034 +* Unify `_is_null_host_scalar` into `is_na_like` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21079 +* Fix datetime binops issues in `pandas-3.0` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21172 +* Fix convert_dtypes with new StringDtype in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21200 +* Preserve StringDtype variants from pandas inputs for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21208 +* Re-enable `dtype=category` in pandas-3.0 branch by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21604 +* Fix all string type bugs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21609 +* Fix astype `CategoricalDtype` cases. by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21627 +* Fix column type for empty columns by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21638 +* Fix interval_range type inference bugs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21686 +* Fix `Index.intersection` for categorical types by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21723 +* Fix string udf failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21727 +* Fix join bug that shows up in `Index.difference` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21733 +* Fix all binops failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21738 +* Use reduction_axis for dataframes by @vyasr in https://github.com/NVIDIA/cudf/pull/21743 +* Fix `astype(category)` for empty column scenarios by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21741 +* Fix all explode failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21766 +* Fix orc pytest failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21847 +* Fix failures in `rolling` pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21846 +* Fix all `Series.astype` bugs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21860 +* Fix numpy `ufunc` return type by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21876 +* Fix dtype issues in `test_between.py` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21887 +* Fix `isin` test failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21886 +* Fix `string.to_pandas` for `nullable` & `arrow_type` cases by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21890 +* Fix `concat` & `union` bugs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21889 +* Fix all `test_iloc.py` failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21872 +* Fix all failures in `test_constructors.py` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21905 +* Fix all string related failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21939 +* Fix Parquet pytest failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21967 +* Switch default time resolutions for `DatetimeIndex` & `TimedeltaIndex` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21997 +* Fix `add_prefix` & `add_suffix` pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22000 +* Fix all json test failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21998 +* Fix runtime warnings in `apply` pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22017 +* Fix string column `to_arrow` to generate large strings when needed by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21891 +* Fix type inconsistencies in pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22016 +* Fix `replace` pytest by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22019 +* Fix parquet metadata failures in pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22020 +* Fix `copy-on-write` failures related to `astype` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22022 +* Fix `str` reductions to match pandas3 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22026 +* Fix `right-join` column ordering by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22002 +* [pandas3] Fix test_categorical_series_with_nan_repr for pandas 3 nan_as_null change by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22066 +* [pandas3] Fix test_diff_many_dtypes for boolean Series diff null representation by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22065 +* [pandas3] Fix test_timedelta_astype_unicode_dtype_pandas_compat for pandas 3 StringDtype by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22064 +* [pandas3] Fix test_categorical_empty: update categories dtype assertion by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22062 +* [pandas3] Fix test_rename_shallow_copy: use np.shares_memory instead of identity check by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22060 +* [pandas3] Fix test_groupby_pct_change_empty_columns for pandas 3 dtype inference by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22059 +* [pandas3] Fix two regressions in Series.describe() by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22058 +* [pandas3] Fix test_alltypes_plain_avro timestamp conversion for pandas 3 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22061 +* [pandas3] Add _CachedAccessor to replace removed pandas.core.accessor.CachedAccessor by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22056 +* Update memory usage test by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22080 +* [pandas3] Fix float-category NaN handling in CategoricalColumn.to_pandas() by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22057 +* Skip pandas failure cases in `ewm` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22082 +* Fix `sort_values` when `return_indexer=True` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22098 +* Fix pandas 3 compatibility bugs in index, groupby, and dask_cudf tests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22101 +* Make `cudf.pandas` pandas tests able to run by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22140 +* Fix `12k` pandas test suite failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22158 +* Allow SeriesApplyKernel to support pd.StringDtype in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22177 +* Fix `datetime` & `timedelta` failures in pandas test suite by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22174 +* Fix return types for `rank` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22180 +* Fix regression in `eval` frame fetching by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22221 +* Propagate writes to internal or private properties to slow objects only by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22231 +* Fix cudf.pandas view semantics across fast/slow conversions and add __array__ to datetimelike ExtensionArray proxies by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22259 +* Fix cudf.pandas pickle round-trips for proxy classes, DatetimeTZ columns, and NaT by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22297 +* Add proxy types for missing offsets by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22314 +* Fix failures in masked array pandas unit tests under cudf.pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22296 +* Fix `StataReader` proxy type by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22322 +* Fix `MultiIndex` getitem by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22326 +* Define attributes that are not discoverable by `dir(cls)` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22330 +* Drop invalidated frequency in `DatetimeIndex` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22328 +* Fix `to_period` parameter by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22329 +* Expose __from_arrow__ on masked extension dtype proxies by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22373 +* Fix datetime resolution for empty data inputs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22363 +* Expose additional proxy attributes for IntervalArray, Styler, and _MethodProxy by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22374 +* Fix `tz` aware binops by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22243 +* Implement min_count for groupby reductions by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22372 +* Fix datetime tz + categorical creation by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22278 +* Expose DateOffset instance attributes on cudf.pandas proxy by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22447 +* Fix np.ufunc(Index, Series) dispatch and np.matmul scalar return type by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22444 +* Fix `fillna` in `binop` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22442 +* Fix `is_numeric_dtype` for `Index` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22438 +* Preserve StringDtype storage and na_value in get_dtype_of_same_kind by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22289 +* Raise TypeError for numeric column arithmetic with pd.NaT by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22451 +* Make StringArray proxy inherit from NumpyExtensionArray by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22457 +* Fix return types of `quantile` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22433 +* Preserve extension dtypes in groupby reductions and exclude grouping-key columns by identity by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22369 +* Fix several to_datetime edge cases for pandas 3 compatibility by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22446 +* Fix arrow types related failures in `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22194 +* Implement groupby all/any via bool-coercion + min/max by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22371 +* Fix `AssertionError: DataFrame.columns are different` failures in cudf.pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22351 +* Fix factorize for pandas 3 compatibility by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22445 +* Fix empty series dtype by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22443 +* Implement groupby sum on StringDtype columns as per-group concatenation by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22370 +* Preserve pandas-nullable dtypes through ufuncs and empty value_counts by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22461 +* Fix all `datetime` type inconsistencies with extension types by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22409 +* Serialize engine config in new pdsh benchmark CLI by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22365 +* Fix unsafe, missing close() of HostColumnVectors by @mythrocks in https://github.com/NVIDIA/cudf/pull/22432 +* Skip flaky test_make_spill_function on pinned-memory allocation failure by @Matt711 in https://github.com/NVIDIA/cudf/pull/22551 +* Fix JSON token count estimate to account for the delimiter in recovery mode by @shrshi in https://github.com/NVIDIA/cudf/pull/22559 +* Fix AST expression evaluator handling of intermediate values for fixed-point types by @davidwendt in https://github.com/NVIDIA/cudf/pull/22512 +* Remove __arrow_array__ from NumPy ndarray cudf.pandas proxy by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22617 +* Expose per-column had_schema_mismatch diagnostic from JSON reader by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22450 +* Fix Arrow device string view variadic buffers by @0ax1 in https://github.com/NVIDIA/cudf/pull/22620 +* [BUG] Parquet reader: treat LIST-annotated REPEATED group as the list, not a stub by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22597 +* Fix invalid use of data_type constructor in row_ir logic by @davidwendt in https://github.com/NVIDIA/cudf/pull/22676 +* Preserve single-field unannotated repeated groups in parquet schema sanitization by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22567 +* Expose `__array_ufunc__` on NumpyExtensionArray cudf.pandas proxy by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22607 +* Expose __from_arrow__ on DatetimeTZDtype cudf.pandas proxy by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22608 +* Fix DataFrame mask Series condition alignment by @aryansri05 in https://github.com/NVIDIA/cudf/pull/22701 +* fix(cudf.pandas): fix Excel/openpyxl test failures and add xlsxwriter dependency by @vyasr in https://github.com/NVIDIA/cudf/pull/22722 +* Fix to_numpy object null handling by @aryansri05 in https://github.com/NVIDIA/cudf/pull/22621 +* Match pandas dtype semantics in groupby quantile by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22750 +* Make `IntervalDtype` construction pandas-compatible by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22731 +* Remove cudf-polars CI timeouts by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22768 +* Fix scalar as column view for null scalars by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22773 +* Increase timeouts for slower cudf-polars tests by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22780 +* Enable chunked row masks in hybrid scan reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22716 +* Fix Parquet page header decoder to allow empty dict pages by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22777 +* declare nvidia-nvjitlink dependency in 'libcudf' wheels by @jameslamb in https://github.com/NVIDIA/cudf/pull/22785 +* Fix replace_with_backrefs copy remainder logic for empty pattern match by @davidwendt in https://github.com/NVIDIA/cudf/pull/22711 +* Fix ArrowDtype integer arithmetic to match pandas pyarrow semantics by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22633 +* Fix zero-count regex quantifier expansion by @fallintoplace in https://github.com/NVIDIA/cudf/pull/22800 +* Build cudf_streaming wheel against this CI run's pylibcudf by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22819 +* Fix groupby.apply columns axis name for row-like UDF results by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22809 +* Use (un)spill_partitions from RapidsMPF by @pentschev in https://github.com/NVIDIA/cudf/pull/22837 +* Fix cudf.pandas DateOffset equality and empty-frame column dtype preservation by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22806 +* Fix `MultiIndex.join` to join on all common levels by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22827 +* Let NVRTC choose automatic PCH directory by @bdice in https://github.com/NVIDIA/cudf/pull/22812 +* [cudf-polars] Sort ORDER BY keys by column, not alias by @madsbk in https://github.com/NVIDIA/cudf/pull/22781 +* Fix cudf_polars benchmarking summarizing mean times by @mroeschke in https://github.com/NVIDIA/cudf/pull/22845 +* Fix masked dtype equality in cudf.pandas so real_dtype == proxy_dtype holds by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22842 +* Reduce size of input dataframe in test_parallel_scan by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22835 +* Fix findall logic handling for empty match regex patterns by @davidwendt in https://github.com/NVIDIA/cudf/pull/22714 +* Reduce setup runtime of multi_orderby_range rolling benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/22869 +* Skip polars tests relying on newer sqlite features by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22870 +* Fix ARROW_FLAG_NULLABLE flag usage in cudf interop by @davidwendt in https://github.com/NVIDIA/cudf/pull/22828 +* Pin pytest-cases version to fix nightly build errors by @davidwendt in https://github.com/NVIDIA/cudf/pull/22888 +* Drop dynamic predicates from logical plan by @Matt711 in https://github.com/NVIDIA/cudf/pull/22820 +* Fix group-key columns in `SeriesGroupBy` list-`agg` and `any`/`all` with `as_index=False` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22813 +* Preserve object dtype for string columns when constructing cudf DataFrame from pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22831 +* Reduce concurrency of cudf-polars tests by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22894 +* Fix groupby and single-column `to_numpy` pandas-compatibility bugs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22783 +* Preserve categorical column index in `DataFrame.describe` and handle pure iterators in `cudf.pandas` fallback by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22821 +* [BUG] Fix conditional linking of RTCX embed dependencies & Allow null fragment names by @lamarrr in https://github.com/NVIDIA/cudf/pull/22898 +* Allow IR fallback to handle an ErrorNode by @mroeschke in https://github.com/NVIDIA/cudf/pull/22855 +* Fix type annotations in pylibucdf hybrid scan by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22118 +* Destroy RTCX cache before teardown by @fallintoplace in https://github.com/NVIDIA/cudf/pull/22843 +* `create_string_column` bug fix by @abigalekim in https://github.com/NVIDIA/cudf/pull/22942 +* Fix categorical replace no-op remapping by @anthonyguerrera in https://github.com/NVIDIA/cudf/pull/22436 +* Enable cudf.pandas `test_numpy_ufuncs_basic` via tolerant Index comparison by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22924 +* Make `to_string` match pandas (drop repr-style truncation/footer) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22929 +* Fix cudf.to_numeric to match pandas semantics by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22934 +* Protect against integer overflow in ``groupby_actor`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22950 +* Skip code coverage where dynamic predicates are dropped by @Matt711 in https://github.com/NVIDIA/cudf/pull/22973 +* Work around pola-rs/polars#23214 in streaming dataframe scan by @wence- in https://github.com/NVIDIA/cudf/pull/23007 +* Fix `cudf.pandas --line-profile` clobbering `__file__` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23017 +* Pin pandas to exclude 3.0.4 (segfault constructing `pd.Timedelta` on Python 3.14) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23022 +* Fix stream ordering bug in join of expression-based keys by @wence- in https://github.com/NVIDIA/cudf/pull/23024 +* Fix `DataFrame` and `GroupBy.value_counts` to match pandas ordering, `dropna`, and categorical handling by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22903 +* Fix ``Distinct`` expression lowering by @rjzamora in https://github.com/NVIDIA/cudf/pull/22970 +* Add bounds checking for AST column_reference out-of-range indices by @utkarshparekh in https://github.com/NVIDIA/cudf/pull/23021 +* Implement `DataFrame.count(axis=1)` on the GPU by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23016 +* Fix ptr null check in packed_data_from_cudf_packed_columns by @Matt711 in https://github.com/NVIDIA/cudf/pull/23038 +* Fix cudf-java native_jstring to emit standard UTF-8 for supplementary characters and embedded NULs by @igorpeshansky in https://github.com/NVIDIA/cudf/pull/23018 +* Fix cudf.pandas Excel reader failures (empty-column dtype, string offset width, proxy monkeypatch) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22927 +* Keep Nanoarrow ArrowArrayStream chunks alive during cudaMemcpyBatchAsync by @mroeschke in https://github.com/NVIDIA/cudf/pull/23009 +* Support host buffers (`BytesIO`) in `cudf.read_text` by @nethum529 in https://github.com/NVIDIA/cudf/pull/23032 +* Fully initialize PageInfo before sort_pages copies it by @pramodsatya in https://github.com/NVIDIA/cudf/pull/22957 +* Declare return type on device lambdas queried in host code by @PointKernel in https://github.com/NVIDIA/cudf/pull/23046 +* Add exception for unsupported decimal128 AST operations by @davidwendt in https://github.com/NVIDIA/cudf/pull/22982 +* Fix `explain()` failing on `UnaryFunction` with multiple children by @Matt711 in https://github.com/NVIDIA/cudf/pull/23050 +* Raise parquet partitions_info validation errors by @fallintoplace in https://github.com/NVIDIA/cudf/pull/22801 +* Re-derive cudf.pandas intermediate proxies when their parents change by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22998 +* Add Numba pin to cuDF wheel build requirements by @bdice in https://github.com/NVIDIA/cudf/pull/23084 +* Return np.intp from argsort to match numpy/pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23061 +* Fix pinned-host spill reservations for packed table alignment by @pentschev in https://github.com/NVIDIA/cudf/pull/23088 +* Enable cudf_streaming distributed (MPI/UCXX) tests in CI by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22954 +* Fix DataFrame.rename for an unnamed MultiIndex level by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23062 +* Fix rtcx embed target-level dependencies under the Makefiles generator by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23106 +* fix(pylibcudf): accept inf/-inf as valid FLOAT32 scalar values by @nethum529 in https://github.com/NVIDIA/cudf/pull/23099 +* Match pandas semantics in groupby.rolling and groupby.apply result construction by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23122 +* Hybrid scan reports correct number of input RGs by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/23052 +* Fix uninitialized read for empty segments in segmented top-k by @ttnghia in https://github.com/NVIDIA/cudf/pull/23109 +* Aggregate category values instead of codes in window operations by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23120 +* Sum interleave benchmark byte counts across all columns by @fallintoplace in https://github.com/NVIDIA/cudf/pull/23103 +* Fix type annotation for CachedParquetInfo.size by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23142 +* Raise TypeError for NaT in where() on masked dtypes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23107 +* Fix ``Scan`` projection with pandas index by @rjzamora in https://github.com/NVIDIA/cudf/pull/23135 +* Match pandas semantics in DataFrame.select_dtypes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23116 +* Fix to_datetime nanosecond inference and DatetimeIndex tz wall-time interpretation by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23141 +* Return value_counts groups in order of first appearance by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23140 +* Implement factorize for MultiIndex by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23139 +* Fix cudf.pandas datetimelike plotting failures (canonical freq offsets + Period/offset pickling) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23054 +* Use `rapids_test_add()` again by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/23181 +* Add error checking to format inference by @vyasr in https://github.com/NVIDIA/cudf/pull/23172 +* Set `RUN_SERIAL` property on streaming tests by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/23188 +* Handle Arrow bool masks in DataFrame getitem by @dttutty in https://github.com/NVIDIA/cudf/pull/23195 +* Supply stream and mr to column synthesizers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/23209 +* Fix out-of-bounds access in segmented top-k with partial offsets by @ttnghia in https://github.com/NVIDIA/cudf/pull/23105 +* Remove duplicate return in ndsh apply_inner_join by @fallintoplace in https://github.com/NVIDIA/cudf/pull/23104 +* Fix cuco::extent type for comparator_helper key_set in distinct_count by @davidwendt in https://github.com/NVIDIA/cudf/pull/23236 +* Fix string reductions and isin across string dtype flavors by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23214 +* fix(pre-commit): normalize SPDX handling by @bdice in https://github.com/NVIDIA/cudf/pull/23228 +* Fix cudf.pandas arithmetic test failures (pow null identities, RangeIndex range-preserving binops, Index proxy array priority) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23176 +* Fix benchmark generate_input create_ascii_string_column function signature by @davidwendt in https://github.com/NVIDIA/cudf/pull/23239 +* Raise OutOfBoundsDatetime for datetime conversions that exceed the target unit's range by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23216 +* Fix cudf.pandas series accessor failures across numpy/nullable/arrow dtypes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23056 +* Fix Series.isin dtype and value semantics for masked (nullable) dtypes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23057 +* Replace thrust::reduce_by_key with cub::DeviceReduce::ReduceByKey in tdigest generation by @PointKernel in https://github.com/NVIDIA/cudf/pull/23204 +* Change cuco::extent type usage from size_type to size_t by @davidwendt in https://github.com/NVIDIA/cudf/pull/23259 +* Fix DataFrame.quantile and Series.quantile to match pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23058 +* Fix data_type ctor calls in the row_ir get_return_type function by @davidwendt in https://github.com/NVIDIA/cudf/pull/23237 +* Fix groupby size dtype for masked series and apply group order for sort=False by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23260 +* Rewrite GroupBy.nth as a pandas-compatible positional row filter by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23257 +* Keep the dtype family in GroupBy.rank results by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23258 +* Fix stale num_rows argument breaking mypy in cudf-polars duplicated-output path by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23303 +* Pin pyarrow<24 in the cudf and pylibcudf conda recipes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23319 +* Fix cudf.pandas isinstance checks for unproxied pandas subclasses and wrap groupby resamplers by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23273 +* Fix GroupBy.apply result assembly, UDF closure side effects, and empty-frame dtypes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23272 +* Avoid leaving multi-rank execution in a bad state in cudf-polars tests by @wence- in https://github.com/NVIDIA/cudf/pull/23340 +* Reconcile Index.union dtypes before the empty-operand short-circuits by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23318 +* Support the limit parameter in GroupBy.ffill and bfill by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23302 +* Stop forward-filling in GroupBy.pct_change by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23301 +* Return index labels from GroupBy idxmin/idxmax in agg and transform by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23298 +* Match pandas semantics for null reduction results and mode ordering by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23328 +* Fix incorrect DECIMAL128 groupby sums on Blackwell with CUDA 13.0 by @PointKernel in https://github.com/NVIDIA/cudf/pull/23229 +* Fix null_precedence access in preprocessed_table::create by @davidwendt in https://github.com/NVIDIA/cudf/pull/23238 +* Route GroupBy.transform size/cumcount/ngroup to their group-level implementations by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23297 +* Recognize named-aggregation lambdas as scans in GroupBy by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23300 +* Add timeout stacktrace utility for cudf-polars tests, remove pytest-timeout by @wence- in https://github.com/NVIDIA/cudf/pull/23332 +* Stop mutating the source column dtype in equivalent-type numerical casts by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23364 +* Mirror cudf.pandas class-level monkeypatches onto the real type by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23001 +* Restore the original index in GroupBy transform broadcasts and ngroup by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23296 +* Honor level and sort_remaining in sort_index(axis=1) by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23367 +* Fix lazy quantifier priority handling in Glushkov regex engine by @davidwendt in https://github.com/NVIDIA/cudf/pull/23381 +* Keep MultiIndex columns in GroupBy.agg for MultiIndex-column sources by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23366 +* Preserve masked integer dtypes in GroupBy cumsum/cumprod by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23299 +* Use new rapidsmpf utilities to handle cancellation in bloom filter awaitables by @wence- in https://github.com/NVIDIA/cudf/pull/23293 +* RMM Library Linking Error Bug Fix by @abigalekim in https://github.com/NVIDIA/cudf/pull/23341 +* Fix Parquet stats pinned host copy lifetime by @aryansri05 in https://github.com/NVIDIA/cudf/pull/23219 +* Add cuFile wheel dependencies for Python tests by @bdice in https://github.com/NVIDIA/cudf/pull/23392 +* Include sources.jar and javadoc.jar in the Java Maven repo output by @paul-aiyedun in https://github.com/NVIDIA/cudf/pull/23389 +* Preserve MultiIndex column fidelity through ColumnAccessor round trips by @galipremsagar in https://github.com/NVIDIA/cudf/pull/23365 +* [BUG] Fix ORC chunked writer root row statistics by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/23118 +* Fix flaky cudf-polars Quent integration test by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23482 ### 📖 Documentation -* Fix misleading Javadoc for Aggregation.nth(int) by @piotrrzysko in https://github.com/rapidsai/cudf/pull/22486 -* Remove wrong example from the docs by @findepi in https://github.com/rapidsai/cudf/pull/22575 -* Reorganize cuDF documentation to highlight all projects by @mroeschke in https://github.com/rapidsai/cudf/pull/22352 -* Drop caveat about strict channel priority by @jakirkham in https://github.com/rapidsai/cudf/pull/22688 -* Remove unneeded cudf classic documentation pages by @mroeschke in https://github.com/rapidsai/cudf/pull/22693 -* Fix duplicate and truncated copyright header in lists/detail/copying.hpp by @PointKernel in https://github.com/rapidsai/cudf/pull/22719 -* Fix cudf-polars MemoryResourceConfig example rendering by @bdice in https://github.com/rapidsai/cudf/pull/22751 -* Update cuDF Python user guides to current/pandas 3 behaviors by @mroeschke in https://github.com/rapidsai/cudf/pull/22720 -* Overhaul cudf Python developer docs by @mroeschke in https://github.com/rapidsai/cudf/pull/22689 -* Add blog links to appropriate cpp/examples readme files by @davidwendt in https://github.com/rapidsai/cudf/pull/22799 -* Document and test Parquet row-group output ordering for `read_parquet` by @qbacpey in https://github.com/rapidsai/cudf/pull/22733 -* Fix documentation for `cudf::lists::sort_lists` and the corresponding Java binding `listSortRows` by @ttnghia in https://github.com/rapidsai/cudf/pull/22916 -* Fix Python docs source links by @bdice in https://github.com/rapidsai/cudf/pull/22791 -* Clarify list sorting docs and replace stale disabled `sort_lists` test by @ttnghia in https://github.com/rapidsai/cudf/pull/22921 -* Document cudf.pandas Profiler usage in Python scripts and CLI line profiling by @gmartini2000 in https://github.com/rapidsai/cudf/pull/21991 -* Update docs to use NVIDIA Sphinx theme by @bdice in https://github.com/rapidsai/cudf/pull/23035 -* Add instructions on running PDS-H by @Matt711 in https://github.com/rapidsai/cudf/pull/23025 -* Fix doxygen to include appropriate include path names in generated output by @davidwendt in https://github.com/rapidsai/cudf/pull/23080 -* Add instructions on running PDS-H for pandas by @Matt711 in https://github.com/rapidsai/cudf/pull/23087 -* Add file descriptions for doxygen output by @davidwendt in https://github.com/rapidsai/cudf/pull/23083 -* Clarify C++ review approval requirements by @bdice in https://github.com/rapidsai/cudf/pull/23092 -* Align cuDF documentation branding by @bdice in https://github.com/rapidsai/cudf/pull/23079 -* Update outdated overflow/size guidance in libcudf developer guide by @PointKernel in https://github.com/rapidsai/cudf/pull/23127 -* Fix doxygen file descriptions by moving file/brief out of addtogroup blocks by @davidwendt in https://github.com/rapidsai/cudf/pull/23125 -* Document the include-what-you-use convention in developer and review guidelines by @PointKernel in https://github.com/rapidsai/cudf/pull/23245 +* Fix misleading Javadoc for Aggregation.nth(int) by @piotrrzysko in https://github.com/NVIDIA/cudf/pull/22486 +* Remove wrong example from the docs by @findepi in https://github.com/NVIDIA/cudf/pull/22575 +* Reorganize cuDF documentation to highlight all projects by @mroeschke in https://github.com/NVIDIA/cudf/pull/22352 +* Drop caveat about strict channel priority by @jakirkham in https://github.com/NVIDIA/cudf/pull/22688 +* Remove unneeded cudf classic documentation pages by @mroeschke in https://github.com/NVIDIA/cudf/pull/22693 +* Fix duplicate and truncated copyright header in lists/detail/copying.hpp by @PointKernel in https://github.com/NVIDIA/cudf/pull/22719 +* Fix cudf-polars MemoryResourceConfig example rendering by @bdice in https://github.com/NVIDIA/cudf/pull/22751 +* Update cuDF Python user guides to current/pandas 3 behaviors by @mroeschke in https://github.com/NVIDIA/cudf/pull/22720 +* Overhaul cudf Python developer docs by @mroeschke in https://github.com/NVIDIA/cudf/pull/22689 +* Add blog links to appropriate cpp/examples readme files by @davidwendt in https://github.com/NVIDIA/cudf/pull/22799 +* Document and test Parquet row-group output ordering for `read_parquet` by @qbacpey in https://github.com/NVIDIA/cudf/pull/22733 +* Fix documentation for `cudf::lists::sort_lists` and the corresponding Java binding `listSortRows` by @ttnghia in https://github.com/NVIDIA/cudf/pull/22916 +* Fix Python docs source links by @bdice in https://github.com/NVIDIA/cudf/pull/22791 +* Clarify list sorting docs and replace stale disabled `sort_lists` test by @ttnghia in https://github.com/NVIDIA/cudf/pull/22921 +* Document cudf.pandas Profiler usage in Python scripts and CLI line profiling by @gmartini2000 in https://github.com/NVIDIA/cudf/pull/21991 +* Update docs to use NVIDIA Sphinx theme by @bdice in https://github.com/NVIDIA/cudf/pull/23035 +* Add instructions on running PDS-H by @Matt711 in https://github.com/NVIDIA/cudf/pull/23025 +* Fix doxygen to include appropriate include path names in generated output by @davidwendt in https://github.com/NVIDIA/cudf/pull/23080 +* Add instructions on running PDS-H for pandas by @Matt711 in https://github.com/NVIDIA/cudf/pull/23087 +* Add file descriptions for doxygen output by @davidwendt in https://github.com/NVIDIA/cudf/pull/23083 +* Clarify C++ review approval requirements by @bdice in https://github.com/NVIDIA/cudf/pull/23092 +* Align cuDF documentation branding by @bdice in https://github.com/NVIDIA/cudf/pull/23079 +* Update outdated overflow/size guidance in libcudf developer guide by @PointKernel in https://github.com/NVIDIA/cudf/pull/23127 +* Fix doxygen file descriptions by moving file/brief out of addtogroup blocks by @davidwendt in https://github.com/NVIDIA/cudf/pull/23125 +* Document the include-what-you-use convention in developer and review guidelines by @PointKernel in https://github.com/NVIDIA/cudf/pull/23245 ### 🚀 New Features -* Write variable bit-width keys for Parquet dictionary encoded pages by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22279 -* [FEA] ANSI SQL Operator JIT Support (1) : Refactor ROW IR by @lamarrr in https://github.com/rapidsai/cudf/pull/22511 -* Forward merge release/26.06 into main by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22603 -* [FEA] LTO IR Support (1) - Introduce LibRTCX by @lamarrr in https://github.com/rapidsai/cudf/pull/21625 -* Extend SUM_WITH_OVERFLOW reduction to signed integers and decimals by @PointKernel in https://github.com/rapidsai/cudf/pull/22465 -* [FEA] LTO IR Support (2) - Cleanup CUDF's JIT Sources by @lamarrr in https://github.com/rapidsai/cudf/pull/22641 -* Rename build/probe to right/left in filtered_join by @PointKernel in https://github.com/rapidsai/cudf/pull/22464 -* Parquet utilities to fetch footer and page index buffers from multiple sources by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22613 -* Add hybrid scan multifile reader basics by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22616 -* [FEA] ANSI SQL Operator JIT Support (2) : Implement Operator Library by @lamarrr in https://github.com/rapidsai/cudf/pull/22514 -* Support reading from `file://` URIs in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/22749 -* Create new libcudf_streaming C++ library by @vyasr in https://github.com/rapidsai/cudf/pull/22747 -* Add multifile row group pruning with stats and byte ranges by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22715 -* Add Python/Cython bindings for libcudf_streaming by @vyasr in https://github.com/rapidsai/cudf/pull/22760 -* [FEA] LTO IR Support (3) - Replace JITIFY usage with LIBRTCX by @lamarrr in https://github.com/rapidsai/cudf/pull/22654 -* [BUG][FEA] Implement Static Linking for RTCX's Dependencies by @lamarrr in https://github.com/rapidsai/cudf/pull/22823 -* Add filter_join_indices_output_size by @PointKernel in https://github.com/rapidsai/cudf/pull/22694 -* Expose max_dictionary_size and dictionary_policy in Java/JNI Parquet writer bindings by @thirtiseven in https://github.com/rapidsai/cudf/pull/22656 -* Add multi-column ORDER BY support to range rolling windows by @pramodsatya in https://github.com/rapidsai/cudf/pull/22627 -* Prepare librtcx for standalone usage by @vyasr in https://github.com/rapidsai/cudf/pull/22856 -* Field extraction for Parquet VARIANT columns (scalar + nested objects) by @vuule in https://github.com/rapidsai/cudf/pull/22416 -* Add hybrid scan multifile single-step materializers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22793 -* Add array-index ("[N]") support to VARIANT field extraction by @vuule in https://github.com/rapidsai/cudf/pull/22877 -* Add row mask builder APIs for multifile hybrid scan by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22792 -* Add multifile hybrid scan pass construction API by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22794 -* [MLIR] Scaffold MLIR based UDF backend by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/22766 -* Split libcudf-streaming tests and benchmarks into a dedicated conda package by @pentschev in https://github.com/rapidsai/cudf/pull/22871 -* Added Concurrent Polynomial Transform Benchmark by @lamarrr in https://github.com/rapidsai/cudf/pull/19199 -* Support converting column to a python list in pylibcudf by @Matt711 in https://github.com/rapidsai/cudf/pull/19446 -* [FEA] ANSI SQL Operator JIT Support (3) : Refactor operator library by @lamarrr in https://github.com/rapidsai/cudf/pull/22836 -* Support `cudf-polars` `epoch` by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/18443 -* Support `cudf-polars` `total_xxx` datetime extraction methods by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/18171 -* Add SUM_OVERFLOW in sort groupby by @PointKernel in https://github.com/rapidsai/cudf/pull/22832 -* Skill to compare performance of a branch or PR with main by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22725 -* Add dynamic join key prefilter planning by @pentschev in https://github.com/rapidsai/cudf/pull/22995 -* Add hybrid scan multifile two-step materializers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22833 -* Add hybrid scan multifile chunked materializers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/23013 -* [MLIR] MaskedType: core type, constructor, accessors by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/22884 -* Prepend source index column in Parquet reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22879 -* [FEA] Improve cuDF's Kernel Cache Directory Resolution by @lamarrr in https://github.com/rapidsai/cudf/pull/22975 -* [FEA] ANSI SQL Operator JIT Support (4) : Implement IR for ANSI and other JIT Extensions by @lamarrr in https://github.com/rapidsai/cudf/pull/22602 -* Pin Polars<1.43 by @Matt711 in https://github.com/rapidsai/cudf/pull/22993 -* [FEA] Support input of dictionary-encoded types in Transforms by @lamarrr in https://github.com/rapidsai/cudf/pull/21552 -* [FEA] LTO IR Support (4) - Implement LTO Transform Kernels by @lamarrr in https://github.com/rapidsai/cudf/pull/22680 -* Add config option for JIT parquet filtering by @Matt711 in https://github.com/rapidsai/cudf/pull/21460 -* [MLIR] MaskedType: NA handling by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/22885 -* Implement ``item`` aggregation for ``LazyFrame.pivot`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22941 -* Support fused arithmetic expressions in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/23136 -* JNI support for multiple order-by columns for RANGE-based window functions by @mythrocks in https://github.com/rapidsai/cudf/pull/22778 -* Add Java Variant extraction support by @nartal1 in https://github.com/rapidsai/cudf/pull/23069 -* Add Java Bindings for Hybrid Scan Parquet Reader by @paul-aiyedun in https://github.com/rapidsai/cudf/pull/22456 -* Add cuDF Java option for loading libraries from a user provided directory by @paul-aiyedun in https://github.com/rapidsai/cudf/pull/22592 -* Propagate ``OrderScheme`` metadata through aliases and timestamp truncation by @rjzamora in https://github.com/rapidsai/cudf/pull/23148 -* Prepend row index column in Parquet reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/23077 -* Add memory_resources and reusable validation foundations by @bdice in https://github.com/rapidsai/cudf/pull/23028 -* Select Parquet columns by field ID by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22955 -* Prepend source and row index columns in Hybrid scan reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22878 -* Add Glushkov regex fast-path to libcudf by @lingyany-nv in https://github.com/rapidsai/cudf/pull/21936 -* Add memory resource parameter to cudf::hash_join constructors by @PointKernel in https://github.com/rapidsai/cudf/pull/23223 -* Add ``adjust_ordering`` utility by @rjzamora in https://github.com/rapidsai/cudf/pull/22628 -* Support non-``Col`` shuffle keys in streaming ``Join`` (version 2) by @rjzamora in https://github.com/rapidsai/cudf/pull/23262 -* Add fixed-size rolling window support to cudf-polars by @tolleybot in https://github.com/rapidsai/cudf/pull/21964 -* Add cuDF JAR build support for all Maven classifiers by @paul-aiyedun in https://github.com/rapidsai/cudf/pull/23261 -* Add plan optimizer pass that extracts join keys and uses them to construct pre-filters for inner joins by @pentschev in https://github.com/rapidsai/cudf/pull/22996 -* Experimental Parquet reader supports retention vectors by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/23402 -* Add array indexing support to Parquet variant field extraction by @vuule in https://github.com/rapidsai/cudf/pull/22895 -* Upgrade to nvcomp 5.3.0.16 by @robertmaynard in https://github.com/rapidsai/cudf/pull/23456 +* Write variable bit-width keys for Parquet dictionary encoded pages by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22279 +* [FEA] ANSI SQL Operator JIT Support (1) : Refactor ROW IR by @lamarrr in https://github.com/NVIDIA/cudf/pull/22511 +* Forward merge release/26.06 into main by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22603 +* [FEA] LTO IR Support (1) - Introduce LibRTCX by @lamarrr in https://github.com/NVIDIA/cudf/pull/21625 +* Extend SUM_WITH_OVERFLOW reduction to signed integers and decimals by @PointKernel in https://github.com/NVIDIA/cudf/pull/22465 +* [FEA] LTO IR Support (2) - Cleanup CUDF's JIT Sources by @lamarrr in https://github.com/NVIDIA/cudf/pull/22641 +* Rename build/probe to right/left in filtered_join by @PointKernel in https://github.com/NVIDIA/cudf/pull/22464 +* Parquet utilities to fetch footer and page index buffers from multiple sources by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22613 +* Add hybrid scan multifile reader basics by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22616 +* [FEA] ANSI SQL Operator JIT Support (2) : Implement Operator Library by @lamarrr in https://github.com/NVIDIA/cudf/pull/22514 +* Support reading from `file://` URIs in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/22749 +* Create new libcudf_streaming C++ library by @vyasr in https://github.com/NVIDIA/cudf/pull/22747 +* Add multifile row group pruning with stats and byte ranges by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22715 +* Add Python/Cython bindings for libcudf_streaming by @vyasr in https://github.com/NVIDIA/cudf/pull/22760 +* [FEA] LTO IR Support (3) - Replace JITIFY usage with LIBRTCX by @lamarrr in https://github.com/NVIDIA/cudf/pull/22654 +* [BUG][FEA] Implement Static Linking for RTCX's Dependencies by @lamarrr in https://github.com/NVIDIA/cudf/pull/22823 +* Add filter_join_indices_output_size by @PointKernel in https://github.com/NVIDIA/cudf/pull/22694 +* Expose max_dictionary_size and dictionary_policy in Java/JNI Parquet writer bindings by @thirtiseven in https://github.com/NVIDIA/cudf/pull/22656 +* Add multi-column ORDER BY support to range rolling windows by @pramodsatya in https://github.com/NVIDIA/cudf/pull/22627 +* Prepare librtcx for standalone usage by @vyasr in https://github.com/NVIDIA/cudf/pull/22856 +* Field extraction for Parquet VARIANT columns (scalar + nested objects) by @vuule in https://github.com/NVIDIA/cudf/pull/22416 +* Add hybrid scan multifile single-step materializers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22793 +* Add array-index ("[N]") support to VARIANT field extraction by @vuule in https://github.com/NVIDIA/cudf/pull/22877 +* Add row mask builder APIs for multifile hybrid scan by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22792 +* Add multifile hybrid scan pass construction API by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22794 +* [MLIR] Scaffold MLIR based UDF backend by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/22766 +* Split libcudf-streaming tests and benchmarks into a dedicated conda package by @pentschev in https://github.com/NVIDIA/cudf/pull/22871 +* Added Concurrent Polynomial Transform Benchmark by @lamarrr in https://github.com/NVIDIA/cudf/pull/19199 +* Support converting column to a python list in pylibcudf by @Matt711 in https://github.com/NVIDIA/cudf/pull/19446 +* [FEA] ANSI SQL Operator JIT Support (3) : Refactor operator library by @lamarrr in https://github.com/NVIDIA/cudf/pull/22836 +* Support `cudf-polars` `epoch` by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/18443 +* Support `cudf-polars` `total_xxx` datetime extraction methods by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/18171 +* Add SUM_OVERFLOW in sort groupby by @PointKernel in https://github.com/NVIDIA/cudf/pull/22832 +* Skill to compare performance of a branch or PR with main by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22725 +* Add dynamic join key prefilter planning by @pentschev in https://github.com/NVIDIA/cudf/pull/22995 +* Add hybrid scan multifile two-step materializers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22833 +* Add hybrid scan multifile chunked materializers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/23013 +* [MLIR] MaskedType: core type, constructor, accessors by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/22884 +* Prepend source index column in Parquet reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22879 +* [FEA] Improve cuDF's Kernel Cache Directory Resolution by @lamarrr in https://github.com/NVIDIA/cudf/pull/22975 +* [FEA] ANSI SQL Operator JIT Support (4) : Implement IR for ANSI and other JIT Extensions by @lamarrr in https://github.com/NVIDIA/cudf/pull/22602 +* Pin Polars<1.43 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22993 +* [FEA] Support input of dictionary-encoded types in Transforms by @lamarrr in https://github.com/NVIDIA/cudf/pull/21552 +* [FEA] LTO IR Support (4) - Implement LTO Transform Kernels by @lamarrr in https://github.com/NVIDIA/cudf/pull/22680 +* Add config option for JIT parquet filtering by @Matt711 in https://github.com/NVIDIA/cudf/pull/21460 +* [MLIR] MaskedType: NA handling by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/22885 +* Implement ``item`` aggregation for ``LazyFrame.pivot`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22941 +* Support fused arithmetic expressions in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/23136 +* JNI support for multiple order-by columns for RANGE-based window functions by @mythrocks in https://github.com/NVIDIA/cudf/pull/22778 +* Add Java Variant extraction support by @nartal1 in https://github.com/NVIDIA/cudf/pull/23069 +* Add Java Bindings for Hybrid Scan Parquet Reader by @paul-aiyedun in https://github.com/NVIDIA/cudf/pull/22456 +* Add cuDF Java option for loading libraries from a user provided directory by @paul-aiyedun in https://github.com/NVIDIA/cudf/pull/22592 +* Propagate ``OrderScheme`` metadata through aliases and timestamp truncation by @rjzamora in https://github.com/NVIDIA/cudf/pull/23148 +* Prepend row index column in Parquet reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/23077 +* Add memory_resources and reusable validation foundations by @bdice in https://github.com/NVIDIA/cudf/pull/23028 +* Select Parquet columns by field ID by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22955 +* Prepend source and row index columns in Hybrid scan reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22878 +* Add Glushkov regex fast-path to libcudf by @lingyany-nv in https://github.com/NVIDIA/cudf/pull/21936 +* Add memory resource parameter to cudf::hash_join constructors by @PointKernel in https://github.com/NVIDIA/cudf/pull/23223 +* Add ``adjust_ordering`` utility by @rjzamora in https://github.com/NVIDIA/cudf/pull/22628 +* Support non-``Col`` shuffle keys in streaming ``Join`` (version 2) by @rjzamora in https://github.com/NVIDIA/cudf/pull/23262 +* Add fixed-size rolling window support to cudf-polars by @tolleybot in https://github.com/NVIDIA/cudf/pull/21964 +* Add cuDF JAR build support for all Maven classifiers by @paul-aiyedun in https://github.com/NVIDIA/cudf/pull/23261 +* Add plan optimizer pass that extracts join keys and uses them to construct pre-filters for inner joins by @pentschev in https://github.com/NVIDIA/cudf/pull/22996 +* Experimental Parquet reader supports retention vectors by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/23402 +* Add array indexing support to Parquet variant field extraction by @vuule in https://github.com/NVIDIA/cudf/pull/22895 +* Upgrade to nvcomp 5.3.0.16 by @robertmaynard in https://github.com/NVIDIA/cudf/pull/23456 ### 🛠️ Improvements -* Trim CI temporarily for pandas3 feature branch by @galipremsagar in https://github.com/rapidsai/cudf/pull/20892 -* Ignore UserWarning from pandas 3.0 about obj.round on temporal types by @mroeschke in https://github.com/rapidsai/cudf/pull/20900 -* Remove expect_warning_if in series/test_np_ufuncs.py by @mroeschke in https://github.com/rapidsai/cudf/pull/20899 -* Remove unneeded pandas 3 FutureWarning filters from test_binops.py/test_apply.py by @mroeschke in https://github.com/rapidsai/cudf/pull/20917 -* Remove expected FutureWarning in test_date_range_raise_unsupported for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/20931 -* Remove most pandas 2.x version checks for pandas 3.0 by @mroeschke in https://github.com/rapidsai/cudf/pull/20932 -* Remove more pandas<3 version checks in cuDF by @mroeschke in https://github.com/rapidsai/cudf/pull/20948 -* Drop warnings in `test_reductions.py` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20911 -* Modify/Remove cuDF classic pandas-2 related xfails by @mroeschke in https://github.com/rapidsai/cudf/pull/20988 -* Modify unit tests that use deprecated copy keyword in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21020 -* Fix miscellaneous cuDF test failures from pandas 3 pt 2 by @mroeschke in https://github.com/rapidsai/cudf/pull/21005 -* Fix miscellaneous cuDF test failures from pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21001 -* Implement .groups return dictionary keys FutureWarning for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21021 -* Chage more misc cuDF classic unit tests for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21031 -* Fix DatetimeIndex docstrings for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21035 -* Fix Numba row udf compat with pandas.StringDtype by @mroeschke in https://github.com/rapidsai/cudf/pull/21063 -* Adjust cuDF unit tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21080 -* Bump to `pandas-3.0rc2` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21097 -* Adjust parquet unit tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21109 -* Adjust reset_index and categorial tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21111 -* Adjust string accessor, setitem, constructor tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21127 -* Adjust json, csv tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21128 -* Fix pandas3 related test failures in `test_fillna` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21142 -* Avoid asserting object in test_csv for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21150 -* Fully assert the result for test_join_multi in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21149 -* Adjust categorical tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21148 -* Switch to release version of pandas 3.0 by @galipremsagar in https://github.com/rapidsai/cudf/pull/21121 -* Fix pytest failures in `test_explode` and `test_repr` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21158 -* Adjust misc tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21171 -* Fix most tests that raise a Pandas4Warning by @mroeschke in https://github.com/rapidsai/cudf/pull/21167 -* Adjust more misc tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21193 -* Fix test_index_append_error for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21211 -* Fix docstrings for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21209 -* Forward merge main into pandas3 by @Matt711 in https://github.com/rapidsai/cudf/pull/21215 -* Forward merge main into pandas3 by @Matt711 in https://github.com/rapidsai/cudf/pull/21233 -* Forward merge main into pandas3 by @Matt711 in https://github.com/rapidsai/cudf/pull/21288 -* Bump pandas 3 branch's latest version to 3.0.1 by @mroeschke in https://github.com/rapidsai/cudf/pull/21593 -* Fix some minor, pandas 3 branch backport issues by @mroeschke in https://github.com/rapidsai/cudf/pull/21595 -* Fix `strftime` callers in `pandas3` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21624 -* Fix groupby value_counts pytest by @galipremsagar in https://github.com/rapidsai/cudf/pull/21689 -* Fix all udf pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21731 -* Fix all pytest failures in `test_apply.py` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21835 -* Fix pandas failures in fillna tests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21830 -* Update test_attributes for new better cudf/pandas alignment by @vyasr in https://github.com/rapidsai/cudf/pull/21833 -* Use arrow type to fix test issue by @vyasr in https://github.com/rapidsai/cudf/pull/21878 -* Remove now unnecessary env var flags by @vyasr in https://github.com/rapidsai/cudf/pull/21877 -* Fix type inconsistencies in `diff` pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21893 -* Update `loc` pytests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21983 -* Fix reduction test failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/21986 -* Bump latest pandas to 3.0.2 by @mroeschke in https://github.com/rapidsai/cudf/pull/21987 -* Avoid pytz in test_holidays_within_dates by @mroeschke in https://github.com/rapidsai/cudf/pull/22139 -* Test different pandas option in test_options_mode for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22160 -* Pass writable numpy array to pytorch third party test with pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22161 -* Drop all xpassing tests in pandas-3 by @galipremsagar in https://github.com/rapidsai/cudf/pull/22163 -* Fix custreamz tests for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22195 -* Fix some DatetimeIndex doctests for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22198 -* Adjust pylibcudf test_read_csv_parse_options for pandas 3 large string by @mroeschke in https://github.com/rapidsai/cudf/pull/22200 -* Fix more misc cuDF unit tests for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22201 -* Fix cudf interval test for pandas 3 and numpy < 2 by @mroeschke in https://github.com/rapidsai/cudf/pull/22205 -* Disable csv tests temporarily & fix dask_cudf failure by @galipremsagar in https://github.com/rapidsai/cudf/pull/22203 -* Remove unnecessary older pandas checks in pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22215 -* Change pandas links from version/2.3.3 to stable by @mroeschke in https://github.com/rapidsai/cudf/pull/22218 -* Move skippable pandas tests in run-pandas-test.sh to conftest-patch.py by @mroeschke in https://github.com/rapidsai/cudf/pull/22260 -* Remove pandas tests that no longer exist by @galipremsagar in https://github.com/rapidsai/cudf/pull/22262 -* Implement `Flags` in `cudf` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22225 -* Consolidate all pandas test skips to new NODEIDS_TO_SKIP dictionary by @mroeschke in https://github.com/rapidsai/cudf/pull/22282 -* Drop passing tests from `conftest-patch.py` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22298 -* Consolidate all pandas test to NODEIDS_THAT_FAIL dictionary by @mroeschke in https://github.com/rapidsai/cudf/pull/22311 -* Add remaining pandas test failures to NODEIDS_THAT_FAIL for pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22339 -* Remove dtype workaround in `Column.create` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22396 -* Add ability to set `freq` on `DatetimeIndex` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22440 -* Bump pandas to 3.0.3 by @mroeschke in https://github.com/rapidsai/cudf/pull/22484 -* Handle all dtypes in `convert_dtypes` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22202 -* Update conftest-patch with latest xfails by @galipremsagar in https://github.com/rapidsai/cudf/pull/22489 -* Enable full ci for `pandas3` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22502 -* Use `cudf::tables_equal` API in libcudf examples by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22428 -* Add additional AI review guidelines for pylibcudf/cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/22340 -* Require CMake 4.0 by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22492 -* Bump run_cudf_polars_pytests timeout by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22505 -* Add `Node.get_stable_plan_id` to cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22513 -* Capture command line on pds-h benchmark runner by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22508 -* Fix mypy issue by @wence- in https://github.com/rapidsai/cudf/pull/22549 -* Add nvtx annotations to more cudf-polars IO by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22573 -* Reduce some input data sizes for cudf_polars unit tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22562 -* pylibcudf read_parquet accepts pre-materialized FileMetadata by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22587 -* Add compression type axis to the parquet reader benchmarks by @qbacpey in https://github.com/rapidsai/cudf/pull/22471 -* Excercise assert_collect_raises with engine fixture by @mroeschke in https://github.com/rapidsai/cudf/pull/22563 -* Require engine argument in assert_ir_translation_raises in cudf_polars tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22565 -* Pass memory resource to exec_policy_nosync in remaining modules by @bdice in https://github.com/rapidsai/cudf/pull/22045 -* Pass engine everywhere in assert_sink_ir_translation_raises by @mroeschke in https://github.com/rapidsai/cudf/pull/22609 -* Add pylibucdf bindings for `FileMetadata.row_groups` by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22599 -* Expose libcudf tables_equal in pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/22611 -* Use CCCL overflow utils instead of custom overflow logic by @PointKernel in https://github.com/rapidsai/cudf/pull/22593 -* Improve devcontainer startup times by @trxcllnt in https://github.com/rapidsai/cudf/pull/22594 -* Reduce mutex scoped region in parquet io utilities by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22550 -* Clone rapidsmpf before devcontainer env creation by @bdice in https://github.com/rapidsai/cudf/pull/22579 -* Minimize Dask resource acquisition in cudf_polars tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22646 -* Convert conftest-patch.py to a pytest plugin for pandas testing by @mroeschke in https://github.com/rapidsai/cudf/pull/22643 -* Clean up Polars unit tests setup by @mroeschke in https://github.com/rapidsai/cudf/pull/22605 -* Use `nvcompGetStatusString` when available by @vuule in https://github.com/rapidsai/cudf/pull/22566 -* Add LZ4 to ORC/Parquet IO benchmark compression axes by @vuule in https://github.com/rapidsai/cudf/pull/22564 -* Add row group size/stripe options to Parquet and ORC benchmarks by @qbacpey in https://github.com/rapidsai/cudf/pull/22503 -* Workaround nvcc compiler hangs in libcudf debug build by @davidwendt in https://github.com/rapidsai/cudf/pull/22675 -* Remove deprecated mark_join overloads without join_prefilter by @PointKernel in https://github.com/rapidsai/cudf/pull/22670 -* Add SECURITY.md by @jameslamb in https://github.com/rapidsai/cudf/pull/22684 -* Various cuDF CI fixes by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22677 -* Reduce verbosity of cudf-polars-polars tests by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22543 -* Drop the GIL while making datasources in read_parquet_footers by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22685 -* Use CUB bulk for row hashing by @bdice in https://github.com/rapidsai/cudf/pull/22632 -* Minimize Ray resource acquisition in cudf_polars tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22661 -* Adjust cudf-polars test timeouts by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22691 -* Parallelize statistics collection by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22645 -* Reduce MixedInnerJoinTest.LargeDataMultiBlockCoordination to a single type by @davidwendt in https://github.com/rapidsai/cudf/pull/22698 -* Add noinline to n_table_comparator::operator() by @davidwendt in https://github.com/rapidsai/cudf/pull/22699 -* Add debug-cudf-pandas skill for fixing pandas test compatibility failures by @vyasr in https://github.com/rapidsai/cudf/pull/22625 -* Remove stale python filterwarnings and skipped tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22706 -* PERF: Skip header check in parquet reader by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22679 -* docs: fix linked locations for notebooks after docs refactor by @gforsyth in https://github.com/rapidsai/cudf/pull/22710 -* Pass value to timeout parameters in subprocess/threading APIs in cudf_polars tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22669 -* Drop the GIL in more places for parquet reads by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22692 -* Add cuDF agent skill by @kjmph in https://github.com/rapidsai/cudf/pull/22721 -* Add peak_memory_usage tracking to nvbench benchmarks by @lhlRahman in https://github.com/rapidsai/cudf/pull/22548 -* Add more patterns to libcudf regex benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/22213 -* Remove deprecated cudf::detail::copy_if_safe by @davidwendt in https://github.com/rapidsai/cudf/pull/22682 -* More miscellaneous cuDF Python testing cleanups by @mroeschke in https://github.com/rapidsai/cudf/pull/22713 -* Remove rapidsmpf from all dependencies by @bdice in https://github.com/rapidsai/cudf/pull/22655 -* Fix various pandas issues by @vyasr in https://github.com/rapidsai/cudf/pull/22705 -* Remove duplicated logic that does parquet metadata read by @Matt711 in https://github.com/rapidsai/cudf/pull/22741 -* fix(ci): fix configuration for breaking change notification workflow by @jameslamb in https://github.com/rapidsai/cudf/pull/22732 -* Define pandas and polars compatibility testing versions in dependencies.yaml by @Matt711 in https://github.com/rapidsai/cudf/pull/22189 -* Drop stale tests in `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/22618 -* Remove deprecated filtered_join set_as_build_table API by @PointKernel in https://github.com/rapidsai/cudf/pull/22671 -* Support `polars.Expr.truncate` by @mroeschke in https://github.com/rapidsai/cudf/pull/22754 -* Asynchronously fetch byte ranges from multiple sources by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22586 -* fix(ci): add missing permissions to pandas-tests workflow by @vyasr in https://github.com/rapidsai/cudf/pull/22761 -* Improve Parquet case-insensitive column selection semantics by @qbacpey in https://github.com/rapidsai/cudf/pull/22729 -* Unpin `numba-cuda` upper bound on `main` by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/22730 -* Condense `Decimal32/64/128Column` to `DecimalColumn`, remove other unused code by @mroeschke in https://github.com/rapidsai/cudf/pull/22738 -* Bump polars upper bound to <1.42 by @madsbk in https://github.com/rapidsai/cudf/pull/22745 -* Refactor dynamic Scan node lowering by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22758 -* Add benchmarks for reshaping operations by @Matt711 in https://github.com/rapidsai/cudf/pull/21212 -* Rename build/probe to left/right in mark_join by @PointKernel in https://github.com/rapidsai/cudf/pull/22723 -* Update `cudf.pandas` skill with recent changes by @galipremsagar in https://github.com/rapidsai/cudf/pull/22787 -* Rename build/probe to right/left in mixed_join_semi by @PointKernel in https://github.com/rapidsai/cudf/pull/22790 -* refactor(cudf_polars): use cudf_streaming for cudf-specific streaming imports by @vyasr in https://github.com/rapidsai/cudf/pull/22786 -* Make publish job search for abi3 wheels by @vyasr in https://github.com/rapidsai/cudf/pull/22815 -* Add fused expression IR tests by @Matt711 in https://github.com/rapidsai/cudf/pull/22805 -* refactor(rtcx): decouple librtcx from cudf and add standalone CMake build by @vyasr in https://github.com/rapidsai/cudf/pull/22744 -* Add remaining cudf_streaming tests and benchmarks by @vyasr in https://github.com/rapidsai/cudf/pull/22814 +* Trim CI temporarily for pandas3 feature branch by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20892 +* Ignore UserWarning from pandas 3.0 about obj.round on temporal types by @mroeschke in https://github.com/NVIDIA/cudf/pull/20900 +* Remove expect_warning_if in series/test_np_ufuncs.py by @mroeschke in https://github.com/NVIDIA/cudf/pull/20899 +* Remove unneeded pandas 3 FutureWarning filters from test_binops.py/test_apply.py by @mroeschke in https://github.com/NVIDIA/cudf/pull/20917 +* Remove expected FutureWarning in test_date_range_raise_unsupported for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/20931 +* Remove most pandas 2.x version checks for pandas 3.0 by @mroeschke in https://github.com/NVIDIA/cudf/pull/20932 +* Remove more pandas<3 version checks in cuDF by @mroeschke in https://github.com/NVIDIA/cudf/pull/20948 +* Drop warnings in `test_reductions.py` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20911 +* Modify/Remove cuDF classic pandas-2 related xfails by @mroeschke in https://github.com/NVIDIA/cudf/pull/20988 +* Modify unit tests that use deprecated copy keyword in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21020 +* Fix miscellaneous cuDF test failures from pandas 3 pt 2 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21005 +* Fix miscellaneous cuDF test failures from pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21001 +* Implement .groups return dictionary keys FutureWarning for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21021 +* Chage more misc cuDF classic unit tests for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21031 +* Fix DatetimeIndex docstrings for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21035 +* Fix Numba row udf compat with pandas.StringDtype by @mroeschke in https://github.com/NVIDIA/cudf/pull/21063 +* Adjust cuDF unit tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21080 +* Bump to `pandas-3.0rc2` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21097 +* Adjust parquet unit tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21109 +* Adjust reset_index and categorial tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21111 +* Adjust string accessor, setitem, constructor tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21127 +* Adjust json, csv tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21128 +* Fix pandas3 related test failures in `test_fillna` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21142 +* Avoid asserting object in test_csv for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21150 +* Fully assert the result for test_join_multi in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21149 +* Adjust categorical tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21148 +* Switch to release version of pandas 3.0 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21121 +* Fix pytest failures in `test_explode` and `test_repr` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21158 +* Adjust misc tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21171 +* Fix most tests that raise a Pandas4Warning by @mroeschke in https://github.com/NVIDIA/cudf/pull/21167 +* Adjust more misc tests for existing pandas object types in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21193 +* Fix test_index_append_error for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21211 +* Fix docstrings for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21209 +* Forward merge main into pandas3 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21215 +* Forward merge main into pandas3 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21233 +* Forward merge main into pandas3 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21288 +* Bump pandas 3 branch's latest version to 3.0.1 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21593 +* Fix some minor, pandas 3 branch backport issues by @mroeschke in https://github.com/NVIDIA/cudf/pull/21595 +* Fix `strftime` callers in `pandas3` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21624 +* Fix groupby value_counts pytest by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21689 +* Fix all udf pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21731 +* Fix all pytest failures in `test_apply.py` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21835 +* Fix pandas failures in fillna tests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21830 +* Update test_attributes for new better cudf/pandas alignment by @vyasr in https://github.com/NVIDIA/cudf/pull/21833 +* Use arrow type to fix test issue by @vyasr in https://github.com/NVIDIA/cudf/pull/21878 +* Remove now unnecessary env var flags by @vyasr in https://github.com/NVIDIA/cudf/pull/21877 +* Fix type inconsistencies in `diff` pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21893 +* Update `loc` pytests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21983 +* Fix reduction test failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21986 +* Bump latest pandas to 3.0.2 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21987 +* Avoid pytz in test_holidays_within_dates by @mroeschke in https://github.com/NVIDIA/cudf/pull/22139 +* Test different pandas option in test_options_mode for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22160 +* Pass writable numpy array to pytorch third party test with pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22161 +* Drop all xpassing tests in pandas-3 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22163 +* Fix custreamz tests for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22195 +* Fix some DatetimeIndex doctests for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22198 +* Adjust pylibcudf test_read_csv_parse_options for pandas 3 large string by @mroeschke in https://github.com/NVIDIA/cudf/pull/22200 +* Fix more misc cuDF unit tests for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22201 +* Fix cudf interval test for pandas 3 and numpy < 2 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22205 +* Disable csv tests temporarily & fix dask_cudf failure by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22203 +* Remove unnecessary older pandas checks in pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22215 +* Change pandas links from version/2.3.3 to stable by @mroeschke in https://github.com/NVIDIA/cudf/pull/22218 +* Move skippable pandas tests in run-pandas-test.sh to conftest-patch.py by @mroeschke in https://github.com/NVIDIA/cudf/pull/22260 +* Remove pandas tests that no longer exist by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22262 +* Implement `Flags` in `cudf` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22225 +* Consolidate all pandas test skips to new NODEIDS_TO_SKIP dictionary by @mroeschke in https://github.com/NVIDIA/cudf/pull/22282 +* Drop passing tests from `conftest-patch.py` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22298 +* Consolidate all pandas test to NODEIDS_THAT_FAIL dictionary by @mroeschke in https://github.com/NVIDIA/cudf/pull/22311 +* Add remaining pandas test failures to NODEIDS_THAT_FAIL for pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22339 +* Remove dtype workaround in `Column.create` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22396 +* Add ability to set `freq` on `DatetimeIndex` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22440 +* Bump pandas to 3.0.3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22484 +* Handle all dtypes in `convert_dtypes` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22202 +* Update conftest-patch with latest xfails by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22489 +* Enable full ci for `pandas3` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22502 +* Use `cudf::tables_equal` API in libcudf examples by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22428 +* Add additional AI review guidelines for pylibcudf/cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/22340 +* Require CMake 4.0 by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22492 +* Bump run_cudf_polars_pytests timeout by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22505 +* Add `Node.get_stable_plan_id` to cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22513 +* Capture command line on pds-h benchmark runner by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22508 +* Fix mypy issue by @wence- in https://github.com/NVIDIA/cudf/pull/22549 +* Add nvtx annotations to more cudf-polars IO by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22573 +* Reduce some input data sizes for cudf_polars unit tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22562 +* pylibcudf read_parquet accepts pre-materialized FileMetadata by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22587 +* Add compression type axis to the parquet reader benchmarks by @qbacpey in https://github.com/NVIDIA/cudf/pull/22471 +* Excercise assert_collect_raises with engine fixture by @mroeschke in https://github.com/NVIDIA/cudf/pull/22563 +* Require engine argument in assert_ir_translation_raises in cudf_polars tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22565 +* Pass memory resource to exec_policy_nosync in remaining modules by @bdice in https://github.com/NVIDIA/cudf/pull/22045 +* Pass engine everywhere in assert_sink_ir_translation_raises by @mroeschke in https://github.com/NVIDIA/cudf/pull/22609 +* Add pylibucdf bindings for `FileMetadata.row_groups` by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22599 +* Expose libcudf tables_equal in pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/22611 +* Use CCCL overflow utils instead of custom overflow logic by @PointKernel in https://github.com/NVIDIA/cudf/pull/22593 +* Improve devcontainer startup times by @trxcllnt in https://github.com/NVIDIA/cudf/pull/22594 +* Reduce mutex scoped region in parquet io utilities by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22550 +* Clone rapidsmpf before devcontainer env creation by @bdice in https://github.com/NVIDIA/cudf/pull/22579 +* Minimize Dask resource acquisition in cudf_polars tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22646 +* Convert conftest-patch.py to a pytest plugin for pandas testing by @mroeschke in https://github.com/NVIDIA/cudf/pull/22643 +* Clean up Polars unit tests setup by @mroeschke in https://github.com/NVIDIA/cudf/pull/22605 +* Use `nvcompGetStatusString` when available by @vuule in https://github.com/NVIDIA/cudf/pull/22566 +* Add LZ4 to ORC/Parquet IO benchmark compression axes by @vuule in https://github.com/NVIDIA/cudf/pull/22564 +* Add row group size/stripe options to Parquet and ORC benchmarks by @qbacpey in https://github.com/NVIDIA/cudf/pull/22503 +* Workaround nvcc compiler hangs in libcudf debug build by @davidwendt in https://github.com/NVIDIA/cudf/pull/22675 +* Remove deprecated mark_join overloads without join_prefilter by @PointKernel in https://github.com/NVIDIA/cudf/pull/22670 +* Add SECURITY.md by @jameslamb in https://github.com/NVIDIA/cudf/pull/22684 +* Various cuDF CI fixes by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22677 +* Reduce verbosity of cudf-polars-polars tests by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22543 +* Drop the GIL while making datasources in read_parquet_footers by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22685 +* Use CUB bulk for row hashing by @bdice in https://github.com/NVIDIA/cudf/pull/22632 +* Minimize Ray resource acquisition in cudf_polars tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22661 +* Adjust cudf-polars test timeouts by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22691 +* Parallelize statistics collection by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22645 +* Reduce MixedInnerJoinTest.LargeDataMultiBlockCoordination to a single type by @davidwendt in https://github.com/NVIDIA/cudf/pull/22698 +* Add noinline to n_table_comparator::operator() by @davidwendt in https://github.com/NVIDIA/cudf/pull/22699 +* Add debug-cudf-pandas skill for fixing pandas test compatibility failures by @vyasr in https://github.com/NVIDIA/cudf/pull/22625 +* Remove stale python filterwarnings and skipped tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22706 +* PERF: Skip header check in parquet reader by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22679 +* docs: fix linked locations for notebooks after docs refactor by @gforsyth in https://github.com/NVIDIA/cudf/pull/22710 +* Pass value to timeout parameters in subprocess/threading APIs in cudf_polars tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22669 +* Drop the GIL in more places for parquet reads by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22692 +* Add cuDF agent skill by @kjmph in https://github.com/NVIDIA/cudf/pull/22721 +* Add peak_memory_usage tracking to nvbench benchmarks by @lhlRahman in https://github.com/NVIDIA/cudf/pull/22548 +* Add more patterns to libcudf regex benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/22213 +* Remove deprecated cudf::detail::copy_if_safe by @davidwendt in https://github.com/NVIDIA/cudf/pull/22682 +* More miscellaneous cuDF Python testing cleanups by @mroeschke in https://github.com/NVIDIA/cudf/pull/22713 +* Remove rapidsmpf from all dependencies by @bdice in https://github.com/NVIDIA/cudf/pull/22655 +* Fix various pandas issues by @vyasr in https://github.com/NVIDIA/cudf/pull/22705 +* Remove duplicated logic that does parquet metadata read by @Matt711 in https://github.com/NVIDIA/cudf/pull/22741 +* fix(ci): fix configuration for breaking change notification workflow by @jameslamb in https://github.com/NVIDIA/cudf/pull/22732 +* Define pandas and polars compatibility testing versions in dependencies.yaml by @Matt711 in https://github.com/NVIDIA/cudf/pull/22189 +* Drop stale tests in `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22618 +* Remove deprecated filtered_join set_as_build_table API by @PointKernel in https://github.com/NVIDIA/cudf/pull/22671 +* Support `polars.Expr.truncate` by @mroeschke in https://github.com/NVIDIA/cudf/pull/22754 +* Asynchronously fetch byte ranges from multiple sources by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22586 +* fix(ci): add missing permissions to pandas-tests workflow by @vyasr in https://github.com/NVIDIA/cudf/pull/22761 +* Improve Parquet case-insensitive column selection semantics by @qbacpey in https://github.com/NVIDIA/cudf/pull/22729 +* Unpin `numba-cuda` upper bound on `main` by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/22730 +* Condense `Decimal32/64/128Column` to `DecimalColumn`, remove other unused code by @mroeschke in https://github.com/NVIDIA/cudf/pull/22738 +* Bump polars upper bound to <1.42 by @madsbk in https://github.com/NVIDIA/cudf/pull/22745 +* Refactor dynamic Scan node lowering by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22758 +* Add benchmarks for reshaping operations by @Matt711 in https://github.com/NVIDIA/cudf/pull/21212 +* Rename build/probe to left/right in mark_join by @PointKernel in https://github.com/NVIDIA/cudf/pull/22723 +* Update `cudf.pandas` skill with recent changes by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22787 +* Rename build/probe to right/left in mixed_join_semi by @PointKernel in https://github.com/NVIDIA/cudf/pull/22790 +* refactor(cudf_polars): use cudf_streaming for cudf-specific streaming imports by @vyasr in https://github.com/NVIDIA/cudf/pull/22786 +* Make publish job search for abi3 wheels by @vyasr in https://github.com/NVIDIA/cudf/pull/22815 +* Add fused expression IR tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/22805 +* refactor(rtcx): decouple librtcx from cudf and add standalone CMake build by @vyasr in https://github.com/NVIDIA/cudf/pull/22744 +* Add remaining cudf_streaming tests and benchmarks by @vyasr in https://github.com/NVIDIA/cudf/pull/22814 * Treat ` -` as a single line terminator for `^`/`$` under EXT_NEWLINE by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22763 -* Add rolling window tests that exercise overflow-like edge cases for SUM by @wence- in https://github.com/rapidsai/cudf/pull/22660 -* Add ``send_chunk`` utility for repeated trace-then-send pattern by @rjzamora in https://github.com/rapidsai/cudf/pull/22651 -* Add unstable decorator to cudf-polars by @madsbk in https://github.com/rapidsai/cudf/pull/22804 -* Add CUDA error checks after every kernel launch by @PointKernel in https://github.com/rapidsai/cudf/pull/22755 -* Add `FusedScan` and nxtx annotations for `FusedScan` and `SplitScan` by @Matt711 in https://github.com/rapidsai/cudf/pull/22838 -* Improve performance of strings count_matches utility with specific expression patterns by @davidwendt in https://github.com/rapidsai/cudf/pull/22429 -* Increase polars IR coverage in tests by @Matt711 in https://github.com/rapidsai/cudf/pull/22847 -* Derive per-chunk memory reservation from scan statistics by @Matt711 in https://github.com/rapidsai/cudf/pull/22846 -* Reenable benchmarks when rapidsmpf supports MPI by @vyasr in https://github.com/rapidsai/cudf/pull/22853 -* Build `libcudf_streaming` comm benchmarks with partial MPI/UCXX availability by @pentschev in https://github.com/rapidsai/cudf/pull/22860 -* Don't capture GPU hardware information during CPU benchmarking runs by @mroeschke in https://github.com/rapidsai/cudf/pull/22862 -* Use google cloud read-only Maven Central mirror by @davidwendt in https://github.com/rapidsai/cudf/pull/22875 -* Collect CPU hardware info during benchmarking by @mroeschke in https://github.com/rapidsai/cudf/pull/22873 -* Pin matplotlib<3.11 for pandas-tests CI job by @vyasr in https://github.com/rapidsai/cudf/pull/22882 -* ci: reduce timeout for cudf_streaming C++ tests to 5m by @vyasr in https://github.com/rapidsai/cudf/pull/22880 -* Improve parquet footer reading performance by speculatively reading footer bytes by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22782 -* Add cudf_streaming to clang-format IncludeCategories by @vyasr in https://github.com/rapidsai/cudf/pull/22881 -* refactor: switch to `rapids-artifact-name` for consistent artifact naming by @gforsyth in https://github.com/rapidsai/cudf/pull/22327 -* Refactor streaming scan construction into `StreamingScan` static methods by @Matt711 in https://github.com/rapidsai/cudf/pull/22889 -* Revert unreviewed VARIANT array-index extraction changes by @vuule in https://github.com/rapidsai/cudf/pull/22893 -* chore(cudf.pandas): remove stale xfail entries for tests no longer in pandas by @vyasr in https://github.com/rapidsai/cudf/pull/22883 -* Use conservative column-size estimates in cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/22796 -* fix(artifacts): correctly label non-arch-dependent but cuda-dependent artifacts by @gforsyth in https://github.com/rapidsai/cudf/pull/22902 -* Fix Parquet column mis-resolution when `allow_mismatched_pq_schemas` is enabled by @qbacpey in https://github.com/rapidsai/cudf/pull/22803 -* Rename build/probe to right/left in key_remapping by @PointKernel in https://github.com/rapidsai/cudf/pull/22789 -* CI Add third-party testing of cuml by @betatim in https://github.com/rapidsai/cudf/pull/21897 -* Expand chunked Parquet writer benchmark also add formatted throughput by @Matt711 in https://github.com/rapidsai/cudf/pull/19049 -* Update pre-commit hooks by @vyasr in https://github.com/rapidsai/cudf/pull/22905 -* Expose the table_view version of unique/distinct_count in pylibcudf by @wence- in https://github.com/rapidsai/cudf/pull/22487 -* Revert xfail_strict=false for pandas tests by @vyasr in https://github.com/rapidsai/cudf/pull/22913 -* Cudf_streaming dedup tests from rapidsmpf by @nirandaperera in https://github.com/rapidsai/cudf/pull/22858 -* Add nvbench for libcudf dictionary concatenate by @davidwendt in https://github.com/rapidsai/cudf/pull/22798 -* Adapt to rapidsmpf changes by @wence- in https://github.com/rapidsai/cudf/pull/22919 -* Build and test with CUDA 13.3.0 by @bdice in https://github.com/rapidsai/cudf/pull/22825 -* Fix deprecation warning in transform_polynomials_concurrent benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/22937 -* Test only oldest and newest supported polars versions in PR CI by @Matt711 in https://github.com/rapidsai/cudf/pull/22939 -* Default the chunked Parquet reader's `pass_read_limit` to reduce surprise OOM by @qbacpey in https://github.com/rapidsai/cudf/pull/22779 -* Add reproduce-ci agent skill for local CI failure reproduction by @vyasr in https://github.com/rapidsai/cudf/pull/22872 -* Adding cudf streaming dist tests by @nirandaperera in https://github.com/rapidsai/cudf/pull/22864 -* Hide absorbed rapids_logger from cudf exports by @bdice in https://github.com/rapidsai/cudf/pull/22538 -* fix(cudf.pandas): remove stale xfail entries from pandas-testing-plugin by @vyasr in https://github.com/rapidsai/cudf/pull/22904 -* Update CONTRIBUTING guide and `build.sh` script to include cudf streaming libraries by @Matt711 in https://github.com/rapidsai/cudf/pull/22943 -* Add string_view -> string normalization in pylibcudf test utilities by @vyasr in https://github.com/rapidsai/cudf/pull/22958 -* [FEA] Use the conda package for croaring instead of fetching via CPM by @vyasr in https://github.com/rapidsai/cudf/pull/22953 -* Enable deprecation warnings by default in build.sh by @vyasr in https://github.com/rapidsai/cudf/pull/22959 -* Use main shared-workflows branch by @bdice in https://github.com/rapidsai/cudf/pull/22960 -* java: use GCS mirror for Maven Central, other small changes by @jameslamb in https://github.com/rapidsai/cudf/pull/22969 -* Relax librdkafka pinning by @vyasr in https://github.com/rapidsai/cudf/pull/22961 -* Pin Cython <3.3.0a0 to avoid prerelease regression by @vyasr in https://github.com/rapidsai/cudf/pull/22977 -* Use matplotlib-base instead of matplotlib in conda test deps by @vyasr in https://github.com/rapidsai/cudf/pull/22974 -* Add row-group statistics availability check by @qbacpey in https://github.com/rapidsai/cudf/pull/22664 -* Fix mypy arg-type for from_buffers after pyarrow-stubs bump by @madsbk in https://github.com/rapidsai/cudf/pull/22979 -* Make HostColumnVector constructors for non-nested types public by @paul-aiyedun in https://github.com/rapidsai/cudf/pull/22498 -* Add the predicate in physical plan explain output by @Matt711 in https://github.com/rapidsai/cudf/pull/22984 -* Adjust verbosity of cudf-polars-polars-tests by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22980 -* Omit Parquet min/max statistics for float/double columns containing NaN by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22818 -* Add a I/O partition planning information to benchmark runner by @Matt711 in https://github.com/rapidsai/cudf/pull/22945 -* Add additional regex gtests for contains, count, findall, and replace by @davidwendt in https://github.com/rapidsai/cudf/pull/22874 -* Support building and testing cudf-java on JDK 17/21 by @igorpeshansky in https://github.com/rapidsai/cudf/pull/23006 -* Add pandas-compatible args and caching to RangeIndex.to_numpy by @rpathade in https://github.com/rapidsai/cudf/pull/21896 -* Add regex-flags member variable to internal libcudf reprog class by @davidwendt in https://github.com/rapidsai/cudf/pull/22994 -* Remove nogil from pylibcudf view()/mutable_view() and hoist calls out of nogil blocks by @vyasr in https://github.com/rapidsai/cudf/pull/23002 -* Fuse multi-column range window offset generation by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22863 -* PERF/API: Specify datasource size to avoid HEAD requests for S3 endpoints by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22739 -* Improve performance for concatenate on dictionary columns by @davidwendt in https://github.com/rapidsai/cudf/pull/22807 -* Bound mypy pre-commit dependencies by @vyasr in https://github.com/rapidsai/cudf/pull/23020 -* Move lists_column_device_view and structs_column_device_view to cudf public namespace by @davidwendt in https://github.com/rapidsai/cudf/pull/23008 -* Upgrade nanoarrow to 0.8.0 by @vyasr in https://github.com/rapidsai/cudf/pull/23019 -* Cull unused CudaStreamPoolConfig and dependents by @wence- in https://github.com/rapidsai/cudf/pull/23026 -* fix(csv): correct to_csv quoting documentation and expose QUOTE_NONE by @a-hirota in https://github.com/rapidsai/cudf/pull/21517 -* Avoid unnecessary copy when `TableChunk` data is already packed by @Matt711 in https://github.com/rapidsai/cudf/pull/23034 -* Update CuPy and NumPy dependency specs by @bdice in https://github.com/rapidsai/cudf/pull/22962 -* Document and Migrate host_span vs span public APIs by @joannamooon in https://github.com/rapidsai/cudf/pull/22588 -* Rename `sum_with_overflow` aggregation to `sum_overflow` and deprecate old APIs by @PointKernel in https://github.com/rapidsai/cudf/pull/23014 -* Add braces to prefetch_noexcept location var to fix clang-tidy reported error by @davidwendt in https://github.com/rapidsai/cudf/pull/23066 -* Update dictionary::get_index to use cub FindIf by @davidwendt in https://github.com/rapidsai/cudf/pull/22841 -* Skewed-Length String Dataset Benchmark by @abigalekim in https://github.com/rapidsai/cudf/pull/22876 -* Add additional regex pytests derived from cudf-spark integration tests by @davidwendt in https://github.com/rapidsai/cudf/pull/23065 -* cudf-classic and CI cleanups for NumPy/CuPy 2/14 version bump by @mroeschke in https://github.com/rapidsai/cudf/pull/23053 -* Enforce east const via clang-format QualifierAlignment by @PointKernel in https://github.com/rapidsai/cudf/pull/22834 -* cudf-polars `PythonScan` support with rank-aware IO sources by @madsbk in https://github.com/rapidsai/cudf/pull/22867 -* Add roles to cudf-polars benchmark run outputs. by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23068 -* Improve scan partitioning planning to prefer sizes closer to target by @Matt711 in https://github.com/rapidsai/cudf/pull/23011 -* Improve chunk sampling in dynamic actors by @rjzamora in https://github.com/rapidsai/cudf/pull/23051 -* fix(deps): try to dynamically load nvrtc from wheel by @gforsyth in https://github.com/rapidsai/cudf/pull/23089 -* Add copyright header to .clang-format and enforce it via pre-commit by @nethum529 in https://github.com/rapidsai/cudf/pull/23073 -* Replace numeric_scalar parameters in cudf::strings::slice_strings with std::optional by @davidwendt in https://github.com/rapidsai/cudf/pull/23031 -* Adopt cuco's parametric bloom filter policy in Parquet and streaming by @PointKernel in https://github.com/rapidsai/cudf/pull/23049 -* Prefetch parquet metadata for scan tasks by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22700 -* switch from 'cuda-python' to specific components (e.g. 'cuda-bindings') by @jameslamb in https://github.com/rapidsai/cudf/pull/23091 -* Emit Quent traces for Engine, Lifecycle, and Plans by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22337 -* Add row-level JSON schema mismatch diagnostics by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22915 -* Add Quent trace export to .gitignore by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23152 -* Support single-partition `pl.Expr.is_empty` by @mroeschke in https://github.com/rapidsai/cudf/pull/23170 -* Support single-partition `pl.Expr.gather_every` by @mroeschke in https://github.com/rapidsai/cudf/pull/23166 -* Support single-partition `pl.Expr.drop_nans` by @mroeschke in https://github.com/rapidsai/cudf/pull/23165 -* Support single-partition `pl.Expr.cum_count` by @mroeschke in https://github.com/rapidsai/cudf/pull/23163 -* Support `pl.Expr.str.escape_regex` by @mroeschke in https://github.com/rapidsai/cudf/pull/23171 -* Support `pl.Expr.dt.century` by @mroeschke in https://github.com/rapidsai/cudf/pull/23153 -* Support single-partition `pl.Expr.has_nulls` by @mroeschke in https://github.com/rapidsai/cudf/pull/23168 -* Add agent skill for cudf_polars expressions by @mroeschke in https://github.com/rapidsai/cudf/pull/23078 -* Support single-partition `pl.Expr.extend_constant` by @mroeschke in https://github.com/rapidsai/cudf/pull/23167 -* Refactor host_span to use cuda::std::span internally and remove span_base by @PointKernel in https://github.com/rapidsai/cudf/pull/23072 -* Support single-partition `pl.Expr.product` in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/23169 -* Support `pl.Expr.dt.millennium` by @mroeschke in https://github.com/rapidsai/cudf/pull/23154 -* Support `pl.Expr.dt.days_in_month` by @mroeschke in https://github.com/rapidsai/cudf/pull/23156 -* Support `pl.Expr.dt.quarter` by @mroeschke in https://github.com/rapidsai/cudf/pull/23157 -* Propagate parent nulls into struct children in mergeAndSetValidity by @rishic3 in https://github.com/rapidsai/cudf/pull/22650 -* Support `pl.Expr.rechunk` by @mroeschke in https://github.com/rapidsai/cudf/pull/23192 -* Support remote storage URLs for PDS-H validation by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23202 -* Support `pl.Expr.clip` by @mroeschke in https://github.com/rapidsai/cudf/pull/23162 -* Update benchmark data-generator to allow nullable with no nulls by @davidwendt in https://github.com/rapidsai/cudf/pull/23198 -* Rename some benchmark .cu files to .cpp by @davidwendt in https://github.com/rapidsai/cudf/pull/23186 -* Plumb statistics object into ProgressThread for all engines by @wence- in https://github.com/rapidsai/cudf/pull/23082 -* Support `pl.Expr.reinterpret/degrees/radians` by @mroeschke in https://github.com/rapidsai/cudf/pull/23210 -* Pin pyarrow < 25 in Polars tests by @mroeschke in https://github.com/rapidsai/cudf/pull/23218 -* Optimize RLE decoding by staging in shared memory by @vyasr in https://github.com/rapidsai/cudf/pull/23090 -* Fix IO benchmark naming consistency by @Tjindl in https://github.com/rapidsai/cudf/pull/23180 -* Error when parquet decode string corruption detected by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/23203 -* Add dictionary benchmarks for match-keys and join by @davidwendt in https://github.com/rapidsai/cudf/pull/23185 -* Use rtcx from the librtcx repository by @vyasr in https://github.com/rapidsai/cudf/pull/23137 -* Support `pl.Expr.cot/log1p` by @mroeschke in https://github.com/rapidsai/cudf/pull/23191 -* Start async device writes and default host reads eagerly instead of deferring I/O to future wait by @Emillock in https://github.com/rapidsai/cudf/pull/23231 -* Move function definitions from generate_input.hpp to generate_input.cu by @davidwendt in https://github.com/rapidsai/cudf/pull/23199 -* Add fast-path logic to some libcudf regex APIs by @davidwendt in https://github.com/rapidsai/cudf/pull/22178 -* Fix parquet preprocess strings page bounds check for pages with zero rows by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/23241 -* Support `pl.Expr.truncate/unique_counts` by @mroeschke in https://github.com/rapidsai/cudf/pull/23244 -* Use batched memcpy instead of a custom copy kernel by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/23266 -* Add multifile dictionary pruning support for hybrid scan parquet reader by @qbacpey in https://github.com/rapidsai/cudf/pull/22866 -* Replace docs.rapids.ai links that 400 by @mroeschke in https://github.com/rapidsai/cudf/pull/23270 -* Support `pl.Expr.index_of/search_sorted` by @mroeschke in https://github.com/rapidsai/cudf/pull/23252 -* Support `pl.Expr.arg_max/arg_min/arg_unique/arg_sort` by @mroeschke in https://github.com/rapidsai/cudf/pull/23212 -* Support `pl.Expr.round_sig_figs/sign` by @mroeschke in https://github.com/rapidsai/cudf/pull/23250 -* Add `engine.execute()` returning a `PersistedQueryResult` by @madsbk in https://github.com/rapidsai/cudf/pull/23114 -* No longer link Arrow in CSV and JSON tests by @mroeschke in https://github.com/rapidsai/cudf/pull/23283 -* Support `pl.Expr.replace/replace_strict` by @mroeschke in https://github.com/rapidsai/cudf/pull/23253 -* cudf-polars: preserve zero-column row counts by @madsbk in https://github.com/rapidsai/cudf/pull/23234 -* Improve column selection validation under mismatched PQ schemas by @qbacpey in https://github.com/rapidsai/cudf/pull/23000 -* Support `pl.argwhere/pl.Expr.arg_true` by @mroeschke in https://github.com/rapidsai/cudf/pull/23284 -* Update upstream cuml tests run by @jcrist in https://github.com/rapidsai/cudf/pull/23335 -* Empty commit to trigger a build by @bdice in https://github.com/rapidsai/cudf/pull/23376 -* Remove rmm::device_buffer forward declaration from types.hpp by @davidwendt in https://github.com/rapidsai/cudf/pull/23373 -* Bump cudf-polars test suite timeouts by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23399 -* Bump cudf-polars CI timeouts by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23410 -* Disable join filter pushdown rewrites by default by @TomAugspurger in https://github.com/rapidsai/cudf/pull/23407 -* Add profitability guards for join filter pushdown by @pentschev in https://github.com/rapidsai/cudf/pull/22997 +` as a single line terminator for `^`/`$` under EXT_NEWLINE by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22763 +* Add rolling window tests that exercise overflow-like edge cases for SUM by @wence- in https://github.com/NVIDIA/cudf/pull/22660 +* Add ``send_chunk`` utility for repeated trace-then-send pattern by @rjzamora in https://github.com/NVIDIA/cudf/pull/22651 +* Add unstable decorator to cudf-polars by @madsbk in https://github.com/NVIDIA/cudf/pull/22804 +* Add CUDA error checks after every kernel launch by @PointKernel in https://github.com/NVIDIA/cudf/pull/22755 +* Add `FusedScan` and nxtx annotations for `FusedScan` and `SplitScan` by @Matt711 in https://github.com/NVIDIA/cudf/pull/22838 +* Improve performance of strings count_matches utility with specific expression patterns by @davidwendt in https://github.com/NVIDIA/cudf/pull/22429 +* Increase polars IR coverage in tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/22847 +* Derive per-chunk memory reservation from scan statistics by @Matt711 in https://github.com/NVIDIA/cudf/pull/22846 +* Reenable benchmarks when rapidsmpf supports MPI by @vyasr in https://github.com/NVIDIA/cudf/pull/22853 +* Build `libcudf_streaming` comm benchmarks with partial MPI/UCXX availability by @pentschev in https://github.com/NVIDIA/cudf/pull/22860 +* Don't capture GPU hardware information during CPU benchmarking runs by @mroeschke in https://github.com/NVIDIA/cudf/pull/22862 +* Use google cloud read-only Maven Central mirror by @davidwendt in https://github.com/NVIDIA/cudf/pull/22875 +* Collect CPU hardware info during benchmarking by @mroeschke in https://github.com/NVIDIA/cudf/pull/22873 +* Pin matplotlib<3.11 for pandas-tests CI job by @vyasr in https://github.com/NVIDIA/cudf/pull/22882 +* ci: reduce timeout for cudf_streaming C++ tests to 5m by @vyasr in https://github.com/NVIDIA/cudf/pull/22880 +* Improve parquet footer reading performance by speculatively reading footer bytes by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22782 +* Add cudf_streaming to clang-format IncludeCategories by @vyasr in https://github.com/NVIDIA/cudf/pull/22881 +* refactor: switch to `rapids-artifact-name` for consistent artifact naming by @gforsyth in https://github.com/NVIDIA/cudf/pull/22327 +* Refactor streaming scan construction into `StreamingScan` static methods by @Matt711 in https://github.com/NVIDIA/cudf/pull/22889 +* Revert unreviewed VARIANT array-index extraction changes by @vuule in https://github.com/NVIDIA/cudf/pull/22893 +* chore(cudf.pandas): remove stale xfail entries for tests no longer in pandas by @vyasr in https://github.com/NVIDIA/cudf/pull/22883 +* Use conservative column-size estimates in cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/22796 +* fix(artifacts): correctly label non-arch-dependent but cuda-dependent artifacts by @gforsyth in https://github.com/NVIDIA/cudf/pull/22902 +* Fix Parquet column mis-resolution when `allow_mismatched_pq_schemas` is enabled by @qbacpey in https://github.com/NVIDIA/cudf/pull/22803 +* Rename build/probe to right/left in key_remapping by @PointKernel in https://github.com/NVIDIA/cudf/pull/22789 +* CI Add third-party testing of cuml by @betatim in https://github.com/NVIDIA/cudf/pull/21897 +* Expand chunked Parquet writer benchmark also add formatted throughput by @Matt711 in https://github.com/NVIDIA/cudf/pull/19049 +* Update pre-commit hooks by @vyasr in https://github.com/NVIDIA/cudf/pull/22905 +* Expose the table_view version of unique/distinct_count in pylibcudf by @wence- in https://github.com/NVIDIA/cudf/pull/22487 +* Revert xfail_strict=false for pandas tests by @vyasr in https://github.com/NVIDIA/cudf/pull/22913 +* Cudf_streaming dedup tests from rapidsmpf by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22858 +* Add nvbench for libcudf dictionary concatenate by @davidwendt in https://github.com/NVIDIA/cudf/pull/22798 +* Adapt to rapidsmpf changes by @wence- in https://github.com/NVIDIA/cudf/pull/22919 +* Build and test with CUDA 13.3.0 by @bdice in https://github.com/NVIDIA/cudf/pull/22825 +* Fix deprecation warning in transform_polynomials_concurrent benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/22937 +* Test only oldest and newest supported polars versions in PR CI by @Matt711 in https://github.com/NVIDIA/cudf/pull/22939 +* Default the chunked Parquet reader's `pass_read_limit` to reduce surprise OOM by @qbacpey in https://github.com/NVIDIA/cudf/pull/22779 +* Add reproduce-ci agent skill for local CI failure reproduction by @vyasr in https://github.com/NVIDIA/cudf/pull/22872 +* Adding cudf streaming dist tests by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22864 +* Hide absorbed rapids_logger from cudf exports by @bdice in https://github.com/NVIDIA/cudf/pull/22538 +* fix(cudf.pandas): remove stale xfail entries from pandas-testing-plugin by @vyasr in https://github.com/NVIDIA/cudf/pull/22904 +* Update CONTRIBUTING guide and `build.sh` script to include cudf streaming libraries by @Matt711 in https://github.com/NVIDIA/cudf/pull/22943 +* Add string_view -> string normalization in pylibcudf test utilities by @vyasr in https://github.com/NVIDIA/cudf/pull/22958 +* [FEA] Use the conda package for croaring instead of fetching via CPM by @vyasr in https://github.com/NVIDIA/cudf/pull/22953 +* Enable deprecation warnings by default in build.sh by @vyasr in https://github.com/NVIDIA/cudf/pull/22959 +* Use main shared-workflows branch by @bdice in https://github.com/NVIDIA/cudf/pull/22960 +* java: use GCS mirror for Maven Central, other small changes by @jameslamb in https://github.com/NVIDIA/cudf/pull/22969 +* Relax librdkafka pinning by @vyasr in https://github.com/NVIDIA/cudf/pull/22961 +* Pin Cython <3.3.0a0 to avoid prerelease regression by @vyasr in https://github.com/NVIDIA/cudf/pull/22977 +* Use matplotlib-base instead of matplotlib in conda test deps by @vyasr in https://github.com/NVIDIA/cudf/pull/22974 +* Add row-group statistics availability check by @qbacpey in https://github.com/NVIDIA/cudf/pull/22664 +* Fix mypy arg-type for from_buffers after pyarrow-stubs bump by @madsbk in https://github.com/NVIDIA/cudf/pull/22979 +* Make HostColumnVector constructors for non-nested types public by @paul-aiyedun in https://github.com/NVIDIA/cudf/pull/22498 +* Add the predicate in physical plan explain output by @Matt711 in https://github.com/NVIDIA/cudf/pull/22984 +* Adjust verbosity of cudf-polars-polars-tests by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22980 +* Omit Parquet min/max statistics for float/double columns containing NaN by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22818 +* Add a I/O partition planning information to benchmark runner by @Matt711 in https://github.com/NVIDIA/cudf/pull/22945 +* Add additional regex gtests for contains, count, findall, and replace by @davidwendt in https://github.com/NVIDIA/cudf/pull/22874 +* Support building and testing cudf-java on JDK 17/21 by @igorpeshansky in https://github.com/NVIDIA/cudf/pull/23006 +* Add pandas-compatible args and caching to RangeIndex.to_numpy by @rpathade in https://github.com/NVIDIA/cudf/pull/21896 +* Add regex-flags member variable to internal libcudf reprog class by @davidwendt in https://github.com/NVIDIA/cudf/pull/22994 +* Remove nogil from pylibcudf view()/mutable_view() and hoist calls out of nogil blocks by @vyasr in https://github.com/NVIDIA/cudf/pull/23002 +* Fuse multi-column range window offset generation by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22863 +* PERF/API: Specify datasource size to avoid HEAD requests for S3 endpoints by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22739 +* Improve performance for concatenate on dictionary columns by @davidwendt in https://github.com/NVIDIA/cudf/pull/22807 +* Bound mypy pre-commit dependencies by @vyasr in https://github.com/NVIDIA/cudf/pull/23020 +* Move lists_column_device_view and structs_column_device_view to cudf public namespace by @davidwendt in https://github.com/NVIDIA/cudf/pull/23008 +* Upgrade nanoarrow to 0.8.0 by @vyasr in https://github.com/NVIDIA/cudf/pull/23019 +* Cull unused CudaStreamPoolConfig and dependents by @wence- in https://github.com/NVIDIA/cudf/pull/23026 +* fix(csv): correct to_csv quoting documentation and expose QUOTE_NONE by @a-hirota in https://github.com/NVIDIA/cudf/pull/21517 +* Avoid unnecessary copy when `TableChunk` data is already packed by @Matt711 in https://github.com/NVIDIA/cudf/pull/23034 +* Update CuPy and NumPy dependency specs by @bdice in https://github.com/NVIDIA/cudf/pull/22962 +* Document and Migrate host_span vs span public APIs by @joannamooon in https://github.com/NVIDIA/cudf/pull/22588 +* Rename `sum_with_overflow` aggregation to `sum_overflow` and deprecate old APIs by @PointKernel in https://github.com/NVIDIA/cudf/pull/23014 +* Add braces to prefetch_noexcept location var to fix clang-tidy reported error by @davidwendt in https://github.com/NVIDIA/cudf/pull/23066 +* Update dictionary::get_index to use cub FindIf by @davidwendt in https://github.com/NVIDIA/cudf/pull/22841 +* Skewed-Length String Dataset Benchmark by @abigalekim in https://github.com/NVIDIA/cudf/pull/22876 +* Add additional regex pytests derived from cudf-spark integration tests by @davidwendt in https://github.com/NVIDIA/cudf/pull/23065 +* cudf-classic and CI cleanups for NumPy/CuPy 2/14 version bump by @mroeschke in https://github.com/NVIDIA/cudf/pull/23053 +* Enforce east const via clang-format QualifierAlignment by @PointKernel in https://github.com/NVIDIA/cudf/pull/22834 +* cudf-polars `PythonScan` support with rank-aware IO sources by @madsbk in https://github.com/NVIDIA/cudf/pull/22867 +* Add roles to cudf-polars benchmark run outputs. by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23068 +* Improve scan partitioning planning to prefer sizes closer to target by @Matt711 in https://github.com/NVIDIA/cudf/pull/23011 +* Improve chunk sampling in dynamic actors by @rjzamora in https://github.com/NVIDIA/cudf/pull/23051 +* fix(deps): try to dynamically load nvrtc from wheel by @gforsyth in https://github.com/NVIDIA/cudf/pull/23089 +* Add copyright header to .clang-format and enforce it via pre-commit by @nethum529 in https://github.com/NVIDIA/cudf/pull/23073 +* Replace numeric_scalar parameters in cudf::strings::slice_strings with std::optional by @davidwendt in https://github.com/NVIDIA/cudf/pull/23031 +* Adopt cuco's parametric bloom filter policy in Parquet and streaming by @PointKernel in https://github.com/NVIDIA/cudf/pull/23049 +* Prefetch parquet metadata for scan tasks by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22700 +* switch from 'cuda-python' to specific components (e.g. 'cuda-bindings') by @jameslamb in https://github.com/NVIDIA/cudf/pull/23091 +* Emit Quent traces for Engine, Lifecycle, and Plans by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22337 +* Add row-level JSON schema mismatch diagnostics by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22915 +* Add Quent trace export to .gitignore by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23152 +* Support single-partition `pl.Expr.is_empty` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23170 +* Support single-partition `pl.Expr.gather_every` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23166 +* Support single-partition `pl.Expr.drop_nans` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23165 +* Support single-partition `pl.Expr.cum_count` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23163 +* Support `pl.Expr.str.escape_regex` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23171 +* Support `pl.Expr.dt.century` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23153 +* Support single-partition `pl.Expr.has_nulls` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23168 +* Add agent skill for cudf_polars expressions by @mroeschke in https://github.com/NVIDIA/cudf/pull/23078 +* Support single-partition `pl.Expr.extend_constant` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23167 +* Refactor host_span to use cuda::std::span internally and remove span_base by @PointKernel in https://github.com/NVIDIA/cudf/pull/23072 +* Support single-partition `pl.Expr.product` in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/23169 +* Support `pl.Expr.dt.millennium` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23154 +* Support `pl.Expr.dt.days_in_month` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23156 +* Support `pl.Expr.dt.quarter` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23157 +* Propagate parent nulls into struct children in mergeAndSetValidity by @rishic3 in https://github.com/NVIDIA/cudf/pull/22650 +* Support `pl.Expr.rechunk` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23192 +* Support remote storage URLs for PDS-H validation by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23202 +* Support `pl.Expr.clip` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23162 +* Update benchmark data-generator to allow nullable with no nulls by @davidwendt in https://github.com/NVIDIA/cudf/pull/23198 +* Rename some benchmark .cu files to .cpp by @davidwendt in https://github.com/NVIDIA/cudf/pull/23186 +* Plumb statistics object into ProgressThread for all engines by @wence- in https://github.com/NVIDIA/cudf/pull/23082 +* Support `pl.Expr.reinterpret/degrees/radians` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23210 +* Pin pyarrow < 25 in Polars tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/23218 +* Optimize RLE decoding by staging in shared memory by @vyasr in https://github.com/NVIDIA/cudf/pull/23090 +* Fix IO benchmark naming consistency by @Tjindl in https://github.com/NVIDIA/cudf/pull/23180 +* Error when parquet decode string corruption detected by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/23203 +* Add dictionary benchmarks for match-keys and join by @davidwendt in https://github.com/NVIDIA/cudf/pull/23185 +* Use rtcx from the librtcx repository by @vyasr in https://github.com/NVIDIA/cudf/pull/23137 +* Support `pl.Expr.cot/log1p` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23191 +* Start async device writes and default host reads eagerly instead of deferring I/O to future wait by @Emillock in https://github.com/NVIDIA/cudf/pull/23231 +* Move function definitions from generate_input.hpp to generate_input.cu by @davidwendt in https://github.com/NVIDIA/cudf/pull/23199 +* Add fast-path logic to some libcudf regex APIs by @davidwendt in https://github.com/NVIDIA/cudf/pull/22178 +* Fix parquet preprocess strings page bounds check for pages with zero rows by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/23241 +* Support `pl.Expr.truncate/unique_counts` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23244 +* Use batched memcpy instead of a custom copy kernel by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/23266 +* Add multifile dictionary pruning support for hybrid scan parquet reader by @qbacpey in https://github.com/NVIDIA/cudf/pull/22866 +* Replace docs.rapids.ai links that 400 by @mroeschke in https://github.com/NVIDIA/cudf/pull/23270 +* Support `pl.Expr.index_of/search_sorted` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23252 +* Support `pl.Expr.arg_max/arg_min/arg_unique/arg_sort` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23212 +* Support `pl.Expr.round_sig_figs/sign` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23250 +* Add `engine.execute()` returning a `PersistedQueryResult` by @madsbk in https://github.com/NVIDIA/cudf/pull/23114 +* No longer link Arrow in CSV and JSON tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/23283 +* Support `pl.Expr.replace/replace_strict` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23253 +* cudf-polars: preserve zero-column row counts by @madsbk in https://github.com/NVIDIA/cudf/pull/23234 +* Improve column selection validation under mismatched PQ schemas by @qbacpey in https://github.com/NVIDIA/cudf/pull/23000 +* Support `pl.argwhere/pl.Expr.arg_true` by @mroeschke in https://github.com/NVIDIA/cudf/pull/23284 +* Update upstream cuml tests run by @jcrist in https://github.com/NVIDIA/cudf/pull/23335 +* Empty commit to trigger a build by @bdice in https://github.com/NVIDIA/cudf/pull/23376 +* Remove rmm::device_buffer forward declaration from types.hpp by @davidwendt in https://github.com/NVIDIA/cudf/pull/23373 +* Bump cudf-polars test suite timeouts by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23399 +* Bump cudf-polars CI timeouts by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23410 +* Disable join filter pushdown rewrites by default by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/23407 +* Add profitability guards for join filter pushdown by @pentschev in https://github.com/NVIDIA/cudf/pull/22997 ## New Contributors -* @piotrrzysko made their first contribution in https://github.com/rapidsai/cudf/pull/22486 -* @findepi made their first contribution in https://github.com/rapidsai/cudf/pull/22575 -* @0ax1 made their first contribution in https://github.com/rapidsai/cudf/pull/22620 -* @kjmph made their first contribution in https://github.com/rapidsai/cudf/pull/22721 -* @lhlRahman made their first contribution in https://github.com/rapidsai/cudf/pull/22548 -* @betatim made their first contribution in https://github.com/rapidsai/cudf/pull/21897 -* @anthonyguerrera made their first contribution in https://github.com/rapidsai/cudf/pull/22436 -* @igorpeshansky made their first contribution in https://github.com/rapidsai/cudf/pull/23006 -* @rpathade made their first contribution in https://github.com/rapidsai/cudf/pull/21896 -* @utkarshparekh made their first contribution in https://github.com/rapidsai/cudf/pull/23021 -* @gmartini2000 made their first contribution in https://github.com/rapidsai/cudf/pull/21991 -* @joannamooon made their first contribution in https://github.com/rapidsai/cudf/pull/22588 -* @dttutty made their first contribution in https://github.com/rapidsai/cudf/pull/23195 -* @Emillock made their first contribution in https://github.com/rapidsai/cudf/pull/23231 -* @lingyany-nv made their first contribution in https://github.com/rapidsai/cudf/pull/21936 -* @tolleybot made their first contribution in https://github.com/rapidsai/cudf/pull/21964 - -**Full Changelog**: https://github.com/rapidsai/cudf/compare/v26.08.00a...release/26.08 +* @piotrrzysko made their first contribution in https://github.com/NVIDIA/cudf/pull/22486 +* @findepi made their first contribution in https://github.com/NVIDIA/cudf/pull/22575 +* @0ax1 made their first contribution in https://github.com/NVIDIA/cudf/pull/22620 +* @kjmph made their first contribution in https://github.com/NVIDIA/cudf/pull/22721 +* @lhlRahman made their first contribution in https://github.com/NVIDIA/cudf/pull/22548 +* @betatim made their first contribution in https://github.com/NVIDIA/cudf/pull/21897 +* @anthonyguerrera made their first contribution in https://github.com/NVIDIA/cudf/pull/22436 +* @igorpeshansky made their first contribution in https://github.com/NVIDIA/cudf/pull/23006 +* @rpathade made their first contribution in https://github.com/NVIDIA/cudf/pull/21896 +* @utkarshparekh made their first contribution in https://github.com/NVIDIA/cudf/pull/23021 +* @gmartini2000 made their first contribution in https://github.com/NVIDIA/cudf/pull/21991 +* @joannamooon made their first contribution in https://github.com/NVIDIA/cudf/pull/22588 +* @dttutty made their first contribution in https://github.com/NVIDIA/cudf/pull/23195 +* @Emillock made their first contribution in https://github.com/NVIDIA/cudf/pull/23231 +* @lingyany-nv made their first contribution in https://github.com/NVIDIA/cudf/pull/21936 +* @tolleybot made their first contribution in https://github.com/NVIDIA/cudf/pull/21964 + +**Full Changelog**: https://github.com/NVIDIA/cudf/compare/v26.08.00a...release/26.08 # cudf 26.06.00 (3 Jun 2026) ### 🚨 Breaking Changes -* Undeprecate the byte-pair-encoding APIs by @davidwendt in https://github.com/rapidsai/cudf/pull/21760 -* [Multi-GPU Polars] Introduce Ray mode for multi-GPU cudf-polars execution by @madsbk in https://github.com/rapidsai/cudf/pull/21746 -* Get rid of relaxed constexpr across libcudf by @PointKernel in https://github.com/rapidsai/cudf/pull/21703 -* [Multi-GPU Polars] Use current rmm resource in SPMD mode by @madsbk in https://github.com/rapidsai/cudf/pull/21842 -* Remove obsolete statistics infrastructure by @rjzamora in https://github.com/rapidsai/cudf/pull/21857 -* [Multi-GPU Polars] Create engines directly instead of factory functions by @madsbk in https://github.com/rapidsai/cudf/pull/21898 -* Handle integers in floor division and power AST operators by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21831 -* Enforce cudf_polars `cardinality_factor` and `scheduler` deprecations by @mroeschke in https://github.com/rapidsai/cudf/pull/21988 -* [Multi-GPU Polars] Unify streaming engine options by @madsbk in https://github.com/rapidsai/cudf/pull/21930 -* [Multi-GPU Polars] Split PDSH utils into legacy and new frontend paths by @madsbk in https://github.com/rapidsai/cudf/pull/21941 -* Remove CUDAStreamPolicy enum and simplify CUDA stream policy by @vyasr in https://github.com/rapidsai/cudf/pull/22086 -* [Multi-GPU Polars] Bind workers to topology-local hardware by @madsbk in https://github.com/rapidsai/cudf/pull/22113 -* [FEA] Support Multi-Output JIT Transforms by @lamarrr in https://github.com/rapidsai/cudf/pull/21704 -* Migrate RMM usage to CCCL MR design by @bdice in https://github.com/rapidsai/cudf/pull/22008 -* Refactor cudf-polars plugin for Polars' test suite by @madsbk in https://github.com/rapidsai/cudf/pull/22301 -* Remove legacy Dask-based streaming backends by @madsbk in https://github.com/rapidsai/cudf/pull/22358 -* Make RapidsMPF the default runtime for cudf_polars streaming executor by @mroeschke in https://github.com/rapidsai/cudf/pull/22281 -* Bump minimum Polars version to 1.35 by @mroeschke in https://github.com/rapidsai/cudf/pull/22459 -* Introduce a process-wide singleton engine for `.collect(engine="gpu")` by @madsbk in https://github.com/rapidsai/cudf/pull/22410 -* Remove cudf-polars[rapidsmpf] pip extra & numpy as a [test] dependency; add [dask] pip extra by @mroeschke in https://github.com/rapidsai/cudf/pull/22480 -* Untangle ``target_partition_size`` and ``broadcast_join_limit`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22411 -* Replace `--executor` with extended `--frontend` choices in cudf-polars benchmarks by @madsbk in https://github.com/rapidsai/cudf/pull/22504 -* Clean up legacy test scaffolding in cudf-polars by @madsbk in https://github.com/rapidsai/cudf/pull/22535 -* [cudf_polars] Reorganize package layout by @madsbk in https://github.com/rapidsai/cudf/pull/22491 -* Move ``collectives`` module by @rjzamora in https://github.com/rapidsai/cudf/pull/22578 +* Undeprecate the byte-pair-encoding APIs by @davidwendt in https://github.com/NVIDIA/cudf/pull/21760 +* [Multi-GPU Polars] Introduce Ray mode for multi-GPU cudf-polars execution by @madsbk in https://github.com/NVIDIA/cudf/pull/21746 +* Get rid of relaxed constexpr across libcudf by @PointKernel in https://github.com/NVIDIA/cudf/pull/21703 +* [Multi-GPU Polars] Use current rmm resource in SPMD mode by @madsbk in https://github.com/NVIDIA/cudf/pull/21842 +* Remove obsolete statistics infrastructure by @rjzamora in https://github.com/NVIDIA/cudf/pull/21857 +* [Multi-GPU Polars] Create engines directly instead of factory functions by @madsbk in https://github.com/NVIDIA/cudf/pull/21898 +* Handle integers in floor division and power AST operators by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21831 +* Enforce cudf_polars `cardinality_factor` and `scheduler` deprecations by @mroeschke in https://github.com/NVIDIA/cudf/pull/21988 +* [Multi-GPU Polars] Unify streaming engine options by @madsbk in https://github.com/NVIDIA/cudf/pull/21930 +* [Multi-GPU Polars] Split PDSH utils into legacy and new frontend paths by @madsbk in https://github.com/NVIDIA/cudf/pull/21941 +* Remove CUDAStreamPolicy enum and simplify CUDA stream policy by @vyasr in https://github.com/NVIDIA/cudf/pull/22086 +* [Multi-GPU Polars] Bind workers to topology-local hardware by @madsbk in https://github.com/NVIDIA/cudf/pull/22113 +* [FEA] Support Multi-Output JIT Transforms by @lamarrr in https://github.com/NVIDIA/cudf/pull/21704 +* Migrate RMM usage to CCCL MR design by @bdice in https://github.com/NVIDIA/cudf/pull/22008 +* Refactor cudf-polars plugin for Polars' test suite by @madsbk in https://github.com/NVIDIA/cudf/pull/22301 +* Remove legacy Dask-based streaming backends by @madsbk in https://github.com/NVIDIA/cudf/pull/22358 +* Make RapidsMPF the default runtime for cudf_polars streaming executor by @mroeschke in https://github.com/NVIDIA/cudf/pull/22281 +* Bump minimum Polars version to 1.35 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22459 +* Introduce a process-wide singleton engine for `.collect(engine="gpu")` by @madsbk in https://github.com/NVIDIA/cudf/pull/22410 +* Remove cudf-polars[rapidsmpf] pip extra & numpy as a [test] dependency; add [dask] pip extra by @mroeschke in https://github.com/NVIDIA/cudf/pull/22480 +* Untangle ``target_partition_size`` and ``broadcast_join_limit`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22411 +* Replace `--executor` with extended `--frontend` choices in cudf-polars benchmarks by @madsbk in https://github.com/NVIDIA/cudf/pull/22504 +* Clean up legacy test scaffolding in cudf-polars by @madsbk in https://github.com/NVIDIA/cudf/pull/22535 +* [cudf_polars] Reorganize package layout by @madsbk in https://github.com/NVIDIA/cudf/pull/22491 +* Move ``collectives`` module by @rjzamora in https://github.com/NVIDIA/cudf/pull/22578 ### 🐛 Bug Fixes -* Fix TypeError when gathering on empty indices by @jberg5 in https://github.com/rapidsai/cudf/pull/21705 -* CPU-only importable pdsh benchmark file by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21791 -* Change more Rapidsmpf Shuffler.wait_on to Shuffler.wait by @mroeschke in https://github.com/rapidsai/cudf/pull/21798 -* Add missing headers to reader_impl_chunking_utils.cu by @bdice in https://github.com/rapidsai/cudf/pull/21784 -* Fix TPC-DS query validation failures due to nulls_last mismatch by @Matt711 in https://github.com/rapidsai/cudf/pull/21814 -* IWYU to fix latest CCCL compilation by @vyasr in https://github.com/rapidsai/cudf/pull/21839 -* Fix additional dictionary tests to handle unordered keys by @davidwendt in https://github.com/rapidsai/cudf/pull/21773 -* Add missing includes for cuda::std::abs by @PointKernel in https://github.com/rapidsai/cudf/pull/21845 -* Fix segment calculation in TPC-DS Q54 by @Matt711 in https://github.com/rapidsai/cudf/pull/21829 -* Workaround `sum(nulls)` difference between DuckDB and Polars in TPC-DS Q64 by @Matt711 in https://github.com/rapidsai/cudf/pull/21826 -* Fix ambiguous stream constructor by @bdice in https://github.com/rapidsai/cudf/pull/21881 -* Ensure cudf.pandas proxy object tests populate test-local type maps by @mroeschke in https://github.com/rapidsai/cudf/pull/21879 -* Dont allow rtxpro6000 runners to pick up CI jobs by @Matt711 in https://github.com/rapidsai/cudf/pull/21954 -* Fix union actor deadlock when input branches share a fanout by @Matt711 in https://github.com/rapidsai/cudf/pull/21949 -* Fix expression decomposition when mixing fusable and non-fusable reductions by @Matt711 in https://github.com/rapidsai/cudf/pull/21822 -* Fix type mismatch in groupby-count with multiple partitions by @Matt711 in https://github.com/rapidsai/cudf/pull/21934 -* Return null instead of nan for pl.Expr.mean with rapidsmpf by @mroeschke in https://github.com/rapidsai/cudf/pull/21805 -* Fix stream-ordering bugs related to pool streams by @vuule in https://github.com/rapidsai/cudf/pull/21908 -* Fix `data_alloc_size` query bugs by @rjzamora in https://github.com/rapidsai/cudf/pull/21955 -* Skip pinned memory tests on unsupported systems by @rjzamora in https://github.com/rapidsai/cudf/pull/21976 -* Fix null_count incorrectly marked as pointwise by @vyasr in https://github.com/rapidsai/cudf/pull/21995 -* Add sort_keys to benchmark validation for complex sort expressions by @Matt711 in https://github.com/rapidsai/cudf/pull/21817 -* Avoid invalid pwise join when dynamic-planning is enabled by @rjzamora in https://github.com/rapidsai/cudf/pull/21977 -* Validate PDS-DS Queries Q24, 47, 49, 94 by @Matt711 in https://github.com/rapidsai/cudf/pull/22007 -* Fix validation failures in TPC-DS Q70 and Q79 by @Matt711 in https://github.com/rapidsai/cudf/pull/21820 -* Fix OOM in PDS-DS Q78 by @Matt711 in https://github.com/rapidsai/cudf/pull/22009 -* Workaround unsupported unary function in a groupby context in PDS-DS Q94 by @Matt711 in https://github.com/rapidsai/cudf/pull/22013 -* Fix unreachable else branch in gather bitmask logic by @eternallyproud in https://github.com/rapidsai/cudf/pull/21946 -* Fix cuda error when sorting empty pl.concat result by @jberg5 in https://github.com/rapidsai/cudf/pull/21825 -* Cast groupby sum of integers result to schema in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21990 -* Fix RTX PRO 6000 Blackwell CI by @bdice in https://github.com/rapidsai/cudf/pull/21999 -* Exclude `value_counts` as a pointwise UnaryFunction in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/22001 -* Fix ast return_type_functor to handle decimal types with non-zero scale by @davidwendt in https://github.com/rapidsai/cudf/pull/21996 -* Fix remote IO in cudf-polars pdsh benchmark by @ncclementi in https://github.com/rapidsai/cudf/pull/22090 -* Fix deprecation warnings for set_as_build_table by @davidwendt in https://github.com/rapidsai/cudf/pull/22087 -* Fix libcudf gather segfault in set_all_valid_null_masks by @davidwendt in https://github.com/rapidsai/cudf/pull/22092 -* Fix stable ID for ``DataFrameScan`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22091 -* Fix PDS-H decimal validation failures by @Matt711 in https://github.com/rapidsai/cudf/pull/22107 -* Optimize PDS-DS Q74 by @Matt711 in https://github.com/rapidsai/cudf/pull/22109 -* Pass BufferResource for stream lifetime in rapidsmpf integration layer by @vyasr in https://github.com/rapidsai/cudf/pull/22110 -* Pass required `br` argument to `TableChunk.from_pylibcudf_table` by @pentschev in https://github.com/rapidsai/cudf/pull/22116 -* Fix missing rapidsmpf hiding real `ImportError` in benchmark scripts by @pentschev in https://github.com/rapidsai/cudf/pull/22114 -* Fix partitioning metadata preservation for ``GroupBy`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22111 -* Expand CSE placeholders during HStack lowering by @rjzamora in https://github.com/rapidsai/cudf/pull/21796 -* Fix lists::segmented_gather to return empty for empty input by @davidwendt in https://github.com/rapidsai/cudf/pull/22115 -* Ensure insert_finished() is called on error paths for streaming collectives by @Matt711 in https://github.com/rapidsai/cudf/pull/22142 -* Skip flaky upstream polars deadlock test by @Matt711 in https://github.com/rapidsai/cudf/pull/22182 -* Fix CSE HStack lowering to respect with_columns semantics by @Matt711 in https://github.com/rapidsai/cudf/pull/22184 -* CUDA 13.2 support: prefer `__syncthreads()` to `block.sync()` for shared memory fencing, fix compiler errors in C++ tests by @jameslamb in https://github.com/rapidsai/cudf/pull/22152 -* Prevent memory corruption in ORC reader by @vuule in https://github.com/rapidsai/cudf/pull/22186 -* Pin to `pyarrow<24` in type checking environment by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22230 -* Pin PyArrow to <24 by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22236 -* Increase tolerance in `test_groupby_categorical_key` by @pentschev in https://github.com/rapidsai/cudf/pull/22249 -* Set memory limit for DaskEngine by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22242 -* Revert Date casts in pdsh benchmarks by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22232 -* Fix flaky tracing test in cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22012 -* Pin polars version in type-checking environment by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22256 -* Fix nvbench handling of memory-resource objects by @davidwendt in https://github.com/rapidsai/cudf/pull/22257 -* Temporarily increase max days without success to 40 days by @pentschev in https://github.com/rapidsai/cudf/pull/22264 -* fix(cmake): exclude zstd, roaring, and cuco from install by @vyasr in https://github.com/rapidsai/cudf/pull/22263 -* Rescale timestamp stats to the target precision in parquet predicate pushdown by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22166 -* Multi-rank sinks: enforce directory output for streaming engines by @madsbk in https://github.com/rapidsai/cudf/pull/22285 -* Prevent potential overflow errors in the CSV reader by @vuule in https://github.com/rapidsai/cudf/pull/22237 -* Unsnap throws for malformed copy element by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22283 -* Preserve LIST element field ids in Parquet output by @res-life in https://github.com/rapidsai/cudf/pull/22143 -* Hardcode disabled network bindings by @pentschev in https://github.com/rapidsai/cudf/pull/22253 -* Fix cudaErrorIllegalAddress in concatenate_list_elements when inner list has 0 rows by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22147 -* Add rapidsmpf as a test dependency of py_test_cudf_polars by @Matt711 in https://github.com/rapidsai/cudf/pull/22316 -* Fix malformed pages in PQ byte stream split decoder by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22280 -* Revise ``Sort`` lowering and ``sort_actor`` assumptions by @rjzamora in https://github.com/rapidsai/cudf/pull/22315 -* Fix use-after-free of host_vector when used with cuda_memcpy_async by @davidwendt in https://github.com/rapidsai/cudf/pull/22321 -* Fix use-after-free in `memory_stats_logger` by @PointKernel in https://github.com/rapidsai/cudf/pull/22333 -* Fix CCCL compilation errors by @bdice in https://github.com/rapidsai/cudf/pull/22349 -* Fix CSV reader `delim_whitespace` header handling to match Pandas by @vuule in https://github.com/rapidsai/cudf/pull/22239 -* Fix more use-after-free cases found in libcudf by @davidwendt in https://github.com/rapidsai/cudf/pull/22332 -* Handle 0-row input in `ContainsAny`, `JsonDecode`, and `JsonEncode` by @madsbk in https://github.com/rapidsai/cudf/pull/22362 -* Fix StatsCollector.serialize to use value equality instead of object identity by @Matt711 in https://github.com/rapidsai/cudf/pull/22366 -* Pass managed pool MR explicitly in NDSH parquet data generation by @vuule in https://github.com/rapidsai/cudf/pull/22344 -* Fix compile warnings in libcudf examples by @davidwendt in https://github.com/rapidsai/cudf/pull/22335 -* Multi-rank fixes for cudf-polars streaming by @madsbk in https://github.com/rapidsai/cudf/pull/22361 -* Fix reading of large CSV files (>64MB) by @vuule in https://github.com/rapidsai/cudf/pull/22375 -* Validate PDS-DS Q1 by @Matt711 in https://github.com/rapidsai/cudf/pull/22389 -* Fix a crash in the ORC reader with malformed stripe footers by @vuule in https://github.com/rapidsai/cudf/pull/22383 -* Correctly handle blocks with "block byte size" fields in the Avro reader by @vuule in https://github.com/rapidsai/cudf/pull/22387 -* Fix `to_array` to return non-corrupted data by @galipremsagar in https://github.com/rapidsai/cudf/pull/22342 -* Use thread pool to submit hybrid scan host IO tasks by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21992 -* Fix pdsh script dropping records by @galipremsagar in https://github.com/rapidsai/cudf/pull/22412 -* Handle sign-extension while decoding Parquet decimal stats by @pramodsatya in https://github.com/rapidsai/cudf/pull/22402 -* Fix MERGE_M2 for extreme finite partial means by @wjxiz1992 in https://github.com/rapidsai/cudf/pull/22393 -* [JAVA] Fix ColumnWriterOptions parquet field id placement on outer list/binary/map by @res-life in https://github.com/rapidsai/cudf/pull/22422 -* Align pdsh benchmarks and library defaults by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22399 -* Zero-initialize `is_quoted_flags` buffer in the CSV reader by @vuule in https://github.com/rapidsai/cudf/pull/22386 -* Include find_package_root everywhere it's used by @vyasr in https://github.com/rapidsai/cudf/pull/22460 -* Fix race condition in page header decoder by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22458 -* Deprecate the multi-patterns cudf::strings::replace_re API by @davidwendt in https://github.com/rapidsai/cudf/pull/22380 -* Move replicated-output dedup to the Dask and Ray frontends by @Matt711 in https://github.com/rapidsai/cudf/pull/22394 -* Fix assertion failures in `assert_tpch_result_equal` due to float sort ambiguity by @Matt711 in https://github.com/rapidsai/cudf/pull/22378 -* Validate TPC-DS Q8 by @Matt711 in https://github.com/rapidsai/cudf/pull/22473 -* Revert PR #22490 (Split PR devcontainer CI into pip and conda jobs) by @bdice in https://github.com/rapidsai/cudf/pull/22497 -* Fix ORC reader 1-second error for negative timestamps with non-UTC writer timezone by @vuule in https://github.com/rapidsai/cudf/pull/22179 -* Fix `to_cupy(dtype=...)` on non-numeric columns by @galipremsagar in https://github.com/rapidsai/cudf/pull/22485 -* Avoid allocating over the batch size limit in the JSON reader by @vuule in https://github.com/rapidsai/cudf/pull/22481 -* Guarantee `insert_finished()` on bulk AllGather by @Matt711 in https://github.com/rapidsai/cudf/pull/22516 -* Fix potential malformed headers in parquet delta decoder by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22275 -* Fix JSON reader guards for scatter validity, validation, and max nesting depth by @karthikeyann in https://github.com/rapidsai/cudf/pull/22452 -* [release/26.06] Remove stale import of `deleted assert_collect_raises` by @madsbk in https://github.com/rapidsai/cudf/pull/22558 -* Fix use-after-destroy and stream ordering in Parquet IO utils by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22529 -* Serialize engine config in new pdsh benchmark CLI by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22572 -* Fix memcheck error in json checked-token-level utility by @davidwendt in https://github.com/rapidsai/cudf/pull/22571 -* Patch Arrow to set CMAKE_POLICY_VERSION_MINIMUM for RapidJSON by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22582 -* Remove unnecessary max token count check in JSON tokenizer by @shrshi in https://github.com/rapidsai/cudf/pull/22589 -* Fix silent row drops in multi-GPU joins with computed key expressions by @Matt711 in https://github.com/rapidsai/cudf/pull/22318 -* Adapt ast conversion for literals by @wence- in https://github.com/rapidsai/cudf/pull/22623 -* Backport #22551 by @wence- in https://github.com/rapidsai/cudf/pull/22636 -* Fix default ``target_partition_size`` singleton engine by @rjzamora in https://github.com/rapidsai/cudf/pull/22638 +* Fix TypeError when gathering on empty indices by @jberg5 in https://github.com/NVIDIA/cudf/pull/21705 +* CPU-only importable pdsh benchmark file by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21791 +* Change more Rapidsmpf Shuffler.wait_on to Shuffler.wait by @mroeschke in https://github.com/NVIDIA/cudf/pull/21798 +* Add missing headers to reader_impl_chunking_utils.cu by @bdice in https://github.com/NVIDIA/cudf/pull/21784 +* Fix TPC-DS query validation failures due to nulls_last mismatch by @Matt711 in https://github.com/NVIDIA/cudf/pull/21814 +* IWYU to fix latest CCCL compilation by @vyasr in https://github.com/NVIDIA/cudf/pull/21839 +* Fix additional dictionary tests to handle unordered keys by @davidwendt in https://github.com/NVIDIA/cudf/pull/21773 +* Add missing includes for cuda::std::abs by @PointKernel in https://github.com/NVIDIA/cudf/pull/21845 +* Fix segment calculation in TPC-DS Q54 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21829 +* Workaround `sum(nulls)` difference between DuckDB and Polars in TPC-DS Q64 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21826 +* Fix ambiguous stream constructor by @bdice in https://github.com/NVIDIA/cudf/pull/21881 +* Ensure cudf.pandas proxy object tests populate test-local type maps by @mroeschke in https://github.com/NVIDIA/cudf/pull/21879 +* Dont allow rtxpro6000 runners to pick up CI jobs by @Matt711 in https://github.com/NVIDIA/cudf/pull/21954 +* Fix union actor deadlock when input branches share a fanout by @Matt711 in https://github.com/NVIDIA/cudf/pull/21949 +* Fix expression decomposition when mixing fusable and non-fusable reductions by @Matt711 in https://github.com/NVIDIA/cudf/pull/21822 +* Fix type mismatch in groupby-count with multiple partitions by @Matt711 in https://github.com/NVIDIA/cudf/pull/21934 +* Return null instead of nan for pl.Expr.mean with rapidsmpf by @mroeschke in https://github.com/NVIDIA/cudf/pull/21805 +* Fix stream-ordering bugs related to pool streams by @vuule in https://github.com/NVIDIA/cudf/pull/21908 +* Fix `data_alloc_size` query bugs by @rjzamora in https://github.com/NVIDIA/cudf/pull/21955 +* Skip pinned memory tests on unsupported systems by @rjzamora in https://github.com/NVIDIA/cudf/pull/21976 +* Fix null_count incorrectly marked as pointwise by @vyasr in https://github.com/NVIDIA/cudf/pull/21995 +* Add sort_keys to benchmark validation for complex sort expressions by @Matt711 in https://github.com/NVIDIA/cudf/pull/21817 +* Avoid invalid pwise join when dynamic-planning is enabled by @rjzamora in https://github.com/NVIDIA/cudf/pull/21977 +* Validate PDS-DS Queries Q24, 47, 49, 94 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22007 +* Fix validation failures in TPC-DS Q70 and Q79 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21820 +* Fix OOM in PDS-DS Q78 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22009 +* Workaround unsupported unary function in a groupby context in PDS-DS Q94 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22013 +* Fix unreachable else branch in gather bitmask logic by @eternallyproud in https://github.com/NVIDIA/cudf/pull/21946 +* Fix cuda error when sorting empty pl.concat result by @jberg5 in https://github.com/NVIDIA/cudf/pull/21825 +* Cast groupby sum of integers result to schema in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21990 +* Fix RTX PRO 6000 Blackwell CI by @bdice in https://github.com/NVIDIA/cudf/pull/21999 +* Exclude `value_counts` as a pointwise UnaryFunction in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/22001 +* Fix ast return_type_functor to handle decimal types with non-zero scale by @davidwendt in https://github.com/NVIDIA/cudf/pull/21996 +* Fix remote IO in cudf-polars pdsh benchmark by @ncclementi in https://github.com/NVIDIA/cudf/pull/22090 +* Fix deprecation warnings for set_as_build_table by @davidwendt in https://github.com/NVIDIA/cudf/pull/22087 +* Fix libcudf gather segfault in set_all_valid_null_masks by @davidwendt in https://github.com/NVIDIA/cudf/pull/22092 +* Fix stable ID for ``DataFrameScan`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22091 +* Fix PDS-H decimal validation failures by @Matt711 in https://github.com/NVIDIA/cudf/pull/22107 +* Optimize PDS-DS Q74 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22109 +* Pass BufferResource for stream lifetime in rapidsmpf integration layer by @vyasr in https://github.com/NVIDIA/cudf/pull/22110 +* Pass required `br` argument to `TableChunk.from_pylibcudf_table` by @pentschev in https://github.com/NVIDIA/cudf/pull/22116 +* Fix missing rapidsmpf hiding real `ImportError` in benchmark scripts by @pentschev in https://github.com/NVIDIA/cudf/pull/22114 +* Fix partitioning metadata preservation for ``GroupBy`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22111 +* Expand CSE placeholders during HStack lowering by @rjzamora in https://github.com/NVIDIA/cudf/pull/21796 +* Fix lists::segmented_gather to return empty for empty input by @davidwendt in https://github.com/NVIDIA/cudf/pull/22115 +* Ensure insert_finished() is called on error paths for streaming collectives by @Matt711 in https://github.com/NVIDIA/cudf/pull/22142 +* Skip flaky upstream polars deadlock test by @Matt711 in https://github.com/NVIDIA/cudf/pull/22182 +* Fix CSE HStack lowering to respect with_columns semantics by @Matt711 in https://github.com/NVIDIA/cudf/pull/22184 +* CUDA 13.2 support: prefer `__syncthreads()` to `block.sync()` for shared memory fencing, fix compiler errors in C++ tests by @jameslamb in https://github.com/NVIDIA/cudf/pull/22152 +* Prevent memory corruption in ORC reader by @vuule in https://github.com/NVIDIA/cudf/pull/22186 +* Pin to `pyarrow<24` in type checking environment by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22230 +* Pin PyArrow to <24 by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22236 +* Increase tolerance in `test_groupby_categorical_key` by @pentschev in https://github.com/NVIDIA/cudf/pull/22249 +* Set memory limit for DaskEngine by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22242 +* Revert Date casts in pdsh benchmarks by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22232 +* Fix flaky tracing test in cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22012 +* Pin polars version in type-checking environment by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22256 +* Fix nvbench handling of memory-resource objects by @davidwendt in https://github.com/NVIDIA/cudf/pull/22257 +* Temporarily increase max days without success to 40 days by @pentschev in https://github.com/NVIDIA/cudf/pull/22264 +* fix(cmake): exclude zstd, roaring, and cuco from install by @vyasr in https://github.com/NVIDIA/cudf/pull/22263 +* Rescale timestamp stats to the target precision in parquet predicate pushdown by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22166 +* Multi-rank sinks: enforce directory output for streaming engines by @madsbk in https://github.com/NVIDIA/cudf/pull/22285 +* Prevent potential overflow errors in the CSV reader by @vuule in https://github.com/NVIDIA/cudf/pull/22237 +* Unsnap throws for malformed copy element by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22283 +* Preserve LIST element field ids in Parquet output by @res-life in https://github.com/NVIDIA/cudf/pull/22143 +* Hardcode disabled network bindings by @pentschev in https://github.com/NVIDIA/cudf/pull/22253 +* Fix cudaErrorIllegalAddress in concatenate_list_elements when inner list has 0 rows by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22147 +* Add rapidsmpf as a test dependency of py_test_cudf_polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/22316 +* Fix malformed pages in PQ byte stream split decoder by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22280 +* Revise ``Sort`` lowering and ``sort_actor`` assumptions by @rjzamora in https://github.com/NVIDIA/cudf/pull/22315 +* Fix use-after-free of host_vector when used with cuda_memcpy_async by @davidwendt in https://github.com/NVIDIA/cudf/pull/22321 +* Fix use-after-free in `memory_stats_logger` by @PointKernel in https://github.com/NVIDIA/cudf/pull/22333 +* Fix CCCL compilation errors by @bdice in https://github.com/NVIDIA/cudf/pull/22349 +* Fix CSV reader `delim_whitespace` header handling to match Pandas by @vuule in https://github.com/NVIDIA/cudf/pull/22239 +* Fix more use-after-free cases found in libcudf by @davidwendt in https://github.com/NVIDIA/cudf/pull/22332 +* Handle 0-row input in `ContainsAny`, `JsonDecode`, and `JsonEncode` by @madsbk in https://github.com/NVIDIA/cudf/pull/22362 +* Fix StatsCollector.serialize to use value equality instead of object identity by @Matt711 in https://github.com/NVIDIA/cudf/pull/22366 +* Pass managed pool MR explicitly in NDSH parquet data generation by @vuule in https://github.com/NVIDIA/cudf/pull/22344 +* Fix compile warnings in libcudf examples by @davidwendt in https://github.com/NVIDIA/cudf/pull/22335 +* Multi-rank fixes for cudf-polars streaming by @madsbk in https://github.com/NVIDIA/cudf/pull/22361 +* Fix reading of large CSV files (>64MB) by @vuule in https://github.com/NVIDIA/cudf/pull/22375 +* Validate PDS-DS Q1 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22389 +* Fix a crash in the ORC reader with malformed stripe footers by @vuule in https://github.com/NVIDIA/cudf/pull/22383 +* Correctly handle blocks with "block byte size" fields in the Avro reader by @vuule in https://github.com/NVIDIA/cudf/pull/22387 +* Fix `to_array` to return non-corrupted data by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22342 +* Use thread pool to submit hybrid scan host IO tasks by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21992 +* Fix pdsh script dropping records by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22412 +* Handle sign-extension while decoding Parquet decimal stats by @pramodsatya in https://github.com/NVIDIA/cudf/pull/22402 +* Fix MERGE_M2 for extreme finite partial means by @wjxiz1992 in https://github.com/NVIDIA/cudf/pull/22393 +* [JAVA] Fix ColumnWriterOptions parquet field id placement on outer list/binary/map by @res-life in https://github.com/NVIDIA/cudf/pull/22422 +* Align pdsh benchmarks and library defaults by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22399 +* Zero-initialize `is_quoted_flags` buffer in the CSV reader by @vuule in https://github.com/NVIDIA/cudf/pull/22386 +* Include find_package_root everywhere it's used by @vyasr in https://github.com/NVIDIA/cudf/pull/22460 +* Fix race condition in page header decoder by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22458 +* Deprecate the multi-patterns cudf::strings::replace_re API by @davidwendt in https://github.com/NVIDIA/cudf/pull/22380 +* Move replicated-output dedup to the Dask and Ray frontends by @Matt711 in https://github.com/NVIDIA/cudf/pull/22394 +* Fix assertion failures in `assert_tpch_result_equal` due to float sort ambiguity by @Matt711 in https://github.com/NVIDIA/cudf/pull/22378 +* Validate TPC-DS Q8 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22473 +* Revert PR #22490 (Split PR devcontainer CI into pip and conda jobs) by @bdice in https://github.com/NVIDIA/cudf/pull/22497 +* Fix ORC reader 1-second error for negative timestamps with non-UTC writer timezone by @vuule in https://github.com/NVIDIA/cudf/pull/22179 +* Fix `to_cupy(dtype=...)` on non-numeric columns by @galipremsagar in https://github.com/NVIDIA/cudf/pull/22485 +* Avoid allocating over the batch size limit in the JSON reader by @vuule in https://github.com/NVIDIA/cudf/pull/22481 +* Guarantee `insert_finished()` on bulk AllGather by @Matt711 in https://github.com/NVIDIA/cudf/pull/22516 +* Fix potential malformed headers in parquet delta decoder by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22275 +* Fix JSON reader guards for scatter validity, validation, and max nesting depth by @karthikeyann in https://github.com/NVIDIA/cudf/pull/22452 +* [release/26.06] Remove stale import of `deleted assert_collect_raises` by @madsbk in https://github.com/NVIDIA/cudf/pull/22558 +* Fix use-after-destroy and stream ordering in Parquet IO utils by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22529 +* Serialize engine config in new pdsh benchmark CLI by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22572 +* Fix memcheck error in json checked-token-level utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/22571 +* Patch Arrow to set CMAKE_POLICY_VERSION_MINIMUM for RapidJSON by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22582 +* Remove unnecessary max token count check in JSON tokenizer by @shrshi in https://github.com/NVIDIA/cudf/pull/22589 +* Fix silent row drops in multi-GPU joins with computed key expressions by @Matt711 in https://github.com/NVIDIA/cudf/pull/22318 +* Adapt ast conversion for literals by @wence- in https://github.com/NVIDIA/cudf/pull/22623 +* Backport #22551 by @wence- in https://github.com/NVIDIA/cudf/pull/22636 +* Fix default ``target_partition_size`` singleton engine by @rjzamora in https://github.com/NVIDIA/cudf/pull/22638 ### 📖 Documentation -* Temporarily `nitpick_ignore_regex` pandas sphinx references by @mroeschke in https://github.com/rapidsai/cudf/pull/21774 -* Fix Doxygen `@param` entries in `/src` by @vuule in https://github.com/rapidsai/cudf/pull/21764 -* Fix Doxygen `@param` entries in `/include` by @vuule in https://github.com/rapidsai/cudf/pull/21762 -* Add developer guideline for constexpr and device code by @PointKernel in https://github.com/rapidsai/cudf/pull/21965 -* Update dictionary section in developer guide by @davidwendt in https://github.com/rapidsai/cudf/pull/21979 -* Fix apostrophe in CHANGELOG.md by @davidwendt in https://github.com/rapidsai/cudf/pull/22055 -* Fix doxygen format for contains and datetime functions by @davidwendt in https://github.com/rapidsai/cudf/pull/22151 -* Fix associativity example in pandas-comparison docs by @vyasr in https://github.com/rapidsai/cudf/pull/22196 -* Overhaul cudf-polars docs for new streaming multi-GPU engines by @madsbk in https://github.com/rapidsai/cudf/pull/22252 -* Update cudf-polars benchmarks for new default engine by @btepera in https://github.com/rapidsai/cudf/pull/22619 +* Temporarily `nitpick_ignore_regex` pandas sphinx references by @mroeschke in https://github.com/NVIDIA/cudf/pull/21774 +* Fix Doxygen `@param` entries in `/src` by @vuule in https://github.com/NVIDIA/cudf/pull/21764 +* Fix Doxygen `@param` entries in `/include` by @vuule in https://github.com/NVIDIA/cudf/pull/21762 +* Add developer guideline for constexpr and device code by @PointKernel in https://github.com/NVIDIA/cudf/pull/21965 +* Update dictionary section in developer guide by @davidwendt in https://github.com/NVIDIA/cudf/pull/21979 +* Fix apostrophe in CHANGELOG.md by @davidwendt in https://github.com/NVIDIA/cudf/pull/22055 +* Fix doxygen format for contains and datetime functions by @davidwendt in https://github.com/NVIDIA/cudf/pull/22151 +* Fix associativity example in pandas-comparison docs by @vyasr in https://github.com/NVIDIA/cudf/pull/22196 +* Overhaul cudf-polars docs for new streaming multi-GPU engines by @madsbk in https://github.com/NVIDIA/cudf/pull/22252 +* Update cudf-polars benchmarks for new default engine by @btepera in https://github.com/NVIDIA/cudf/pull/22619 ### 🚀 New Features -* Create public libcudf gather API with a negative_index_policy parameter by @davidwendt in https://github.com/rapidsai/cudf/pull/21739 -* Add join selectivity benchmarks by @PointKernel in https://github.com/rapidsai/cudf/pull/21775 -* Parquet readers support case-insensitive column names by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21700 -* Add agent skills for cudf by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21737 -* Add JIT cache management functions to pylibcudf by @Matt711 in https://github.com/rapidsai/cudf/pull/21795 -* Add a `rebind_stream` API to set streams for all buffers in a column by @vuule in https://github.com/rapidsai/cudf/pull/21940 -* Add ``sort_actor`` to cudf-polars + rapidsmpf by @rjzamora in https://github.com/rapidsai/cudf/pull/21690 -* Add hybrid scan API to construct row group passes by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21895 -* Add support for ignorecase flag in regex functions by @davidwendt in https://github.com/rapidsai/cudf/pull/21861 -* Add pre-filtering support for mark join by @PointKernel in https://github.com/rapidsai/cudf/pull/21865 -* Fuse drop_nulls into n_unique by @vyasr in https://github.com/rapidsai/cudf/pull/22014 -* Upgrade to nvcomp 5.2.0.10 (and 5.2.0.13 for wheels) by @bdice in https://github.com/rapidsai/cudf/pull/22127 -* Support null_count decomposition in multi-partition Select by @Matt711 in https://github.com/rapidsai/cudf/pull/22126 -* Add API to count number of deleted rows across deletion vector(s) by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21963 -* Expose getters for ColumnWriterOptions isBinary and Parquet field id by @res-life in https://github.com/rapidsai/cudf/pull/22188 -* Implement cudf roaring bitmap by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22133 -* Python bindings for hybrid scan API to construct row group passes by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21918 -* Support multi-partition groupby variance / standard deviation aggregations by @Matt711 in https://github.com/rapidsai/cudf/pull/21962 -* Implement `cudf::apply_deletion_mask` API by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22144 -* Add cudf::strings::count API for literal strings by @davidwendt in https://github.com/rapidsai/cudf/pull/22288 -* Add CodeRabbit configuration and AI review guidelines by @bdice in https://github.com/rapidsai/cudf/pull/22176 -* JNI bindings for `strings::contains(column)` by @mythrocks in https://github.com/rapidsai/cudf/pull/22003 -* Support for `cudf::strings::replace()` where the `targets` and `repls` are columns by @mythrocks in https://github.com/rapidsai/cudf/pull/22132 -* Add skip axis to all join benchmarks by @PointKernel in https://github.com/rapidsai/cudf/pull/22241 -* Add decimal128 to groupby_max_cardinality benchmark by @PointKernel in https://github.com/rapidsai/cudf/pull/22162 -* Python bindings and pytests for `cudf::apply_deletion_mask` by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22145 -* Add basic support for VARIANT type to the Parquet reader by @vuule in https://github.com/rapidsai/cudf/pull/22310 -* Add ``LocalRepartitioner`` utility by @rjzamora in https://github.com/rapidsai/cudf/pull/22439 -* Add arrow bloom filter policy by @PointKernel in https://github.com/rapidsai/cudf/pull/22415 -* JNI support for SUM_WITH_OVERFLOW aggregation by @mythrocks in https://github.com/rapidsai/cudf/pull/22404 -* Implement streaming window functions in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/22191 -* Add `streaming_groupby` for stateful streaming aggregation by @PointKernel in https://github.com/rapidsai/cudf/pull/21924 -* Add partitioned probe support for hash joins by @PointKernel in https://github.com/rapidsai/cudf/pull/22108 +* Create public libcudf gather API with a negative_index_policy parameter by @davidwendt in https://github.com/NVIDIA/cudf/pull/21739 +* Add join selectivity benchmarks by @PointKernel in https://github.com/NVIDIA/cudf/pull/21775 +* Parquet readers support case-insensitive column names by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21700 +* Add agent skills for cudf by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21737 +* Add JIT cache management functions to pylibcudf by @Matt711 in https://github.com/NVIDIA/cudf/pull/21795 +* Add a `rebind_stream` API to set streams for all buffers in a column by @vuule in https://github.com/NVIDIA/cudf/pull/21940 +* Add ``sort_actor`` to cudf-polars + rapidsmpf by @rjzamora in https://github.com/NVIDIA/cudf/pull/21690 +* Add hybrid scan API to construct row group passes by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21895 +* Add support for ignorecase flag in regex functions by @davidwendt in https://github.com/NVIDIA/cudf/pull/21861 +* Add pre-filtering support for mark join by @PointKernel in https://github.com/NVIDIA/cudf/pull/21865 +* Fuse drop_nulls into n_unique by @vyasr in https://github.com/NVIDIA/cudf/pull/22014 +* Upgrade to nvcomp 5.2.0.10 (and 5.2.0.13 for wheels) by @bdice in https://github.com/NVIDIA/cudf/pull/22127 +* Support null_count decomposition in multi-partition Select by @Matt711 in https://github.com/NVIDIA/cudf/pull/22126 +* Add API to count number of deleted rows across deletion vector(s) by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21963 +* Expose getters for ColumnWriterOptions isBinary and Parquet field id by @res-life in https://github.com/NVIDIA/cudf/pull/22188 +* Implement cudf roaring bitmap by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22133 +* Python bindings for hybrid scan API to construct row group passes by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21918 +* Support multi-partition groupby variance / standard deviation aggregations by @Matt711 in https://github.com/NVIDIA/cudf/pull/21962 +* Implement `cudf::apply_deletion_mask` API by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22144 +* Add cudf::strings::count API for literal strings by @davidwendt in https://github.com/NVIDIA/cudf/pull/22288 +* Add CodeRabbit configuration and AI review guidelines by @bdice in https://github.com/NVIDIA/cudf/pull/22176 +* JNI bindings for `strings::contains(column)` by @mythrocks in https://github.com/NVIDIA/cudf/pull/22003 +* Support for `cudf::strings::replace()` where the `targets` and `repls` are columns by @mythrocks in https://github.com/NVIDIA/cudf/pull/22132 +* Add skip axis to all join benchmarks by @PointKernel in https://github.com/NVIDIA/cudf/pull/22241 +* Add decimal128 to groupby_max_cardinality benchmark by @PointKernel in https://github.com/NVIDIA/cudf/pull/22162 +* Python bindings and pytests for `cudf::apply_deletion_mask` by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22145 +* Add basic support for VARIANT type to the Parquet reader by @vuule in https://github.com/NVIDIA/cudf/pull/22310 +* Add ``LocalRepartitioner`` utility by @rjzamora in https://github.com/NVIDIA/cudf/pull/22439 +* Add arrow bloom filter policy by @PointKernel in https://github.com/NVIDIA/cudf/pull/22415 +* JNI support for SUM_WITH_OVERFLOW aggregation by @mythrocks in https://github.com/NVIDIA/cudf/pull/22404 +* Implement streaming window functions in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/22191 +* Add `streaming_groupby` for stateful streaming aggregation by @PointKernel in https://github.com/NVIDIA/cudf/pull/21924 +* Add partitioned probe support for hash joins by @PointKernel in https://github.com/NVIDIA/cudf/pull/22108 ### 🛠️ Improvements -* Adapt to rapidsmpf async shuffle changes by @wence- in https://github.com/rapidsai/cudf/pull/21787 -* Remove CSV reader warnings emitted in unit tests by @vuule in https://github.com/rapidsai/cudf/pull/21794 -* Remove test_infer_objects_no_reference from cudf_pandas xfail list by @mroeschke in https://github.com/rapidsai/cudf/pull/21818 -* Change to use non-detail APIs in some libcudf benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/21821 -* Deduplicate libcudf examples CMake files by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21809 -* Merge release/26.04 into main by @davidwendt in https://github.com/rapidsai/cudf/pull/21815 -* Support drop_nulls unary function in expression decomposition by @quasiben in https://github.com/rapidsai/cudf/pull/21837 -* [Multi-GPU Polars] Ray mode in PDSH benchmarks by @madsbk in https://github.com/rapidsai/cudf/pull/21811 -* Improve build time using transform instead of tabulate by @davidwendt in https://github.com/rapidsai/cudf/pull/21793 -* Use conda packages instead of pip packages in test_narwhals & remove xpassing test_series_setitem from pandas tests by @mroeschke in https://github.com/rapidsai/cudf/pull/21862 -* Ensure Lineariser channels in scan_node are shutdown on error in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21854 -* Add missing includes for `` and `` by @bdice in https://github.com/rapidsai/cudf/pull/21859 -* [Multi-GPU Polars] SPMD mode works without `rrun` by @madsbk in https://github.com/rapidsai/cudf/pull/21851 -* cudf-polars tracing improvements by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21789 -* Fix utf8-to-codepoint utility to handle out-of-range unicode by @davidwendt in https://github.com/rapidsai/cudf/pull/21823 -* Use `PyBuffer_FillInfo` for `HostBuffer`'s buffer by @jakirkham in https://github.com/rapidsai/cudf/pull/21855 -* Fix deprecation warning in JNI for parquet_reader_options::builder.names() by @davidwendt in https://github.com/rapidsai/cudf/pull/21868 -* Add noarch Python channel to cudf.pandas third party tests conda solve by @mroeschke in https://github.com/rapidsai/cudf/pull/21873 -* [Multi-GPU Polars] Introduce `StreamingEngine` base class and `SPMDEngine` by @madsbk in https://github.com/rapidsai/cudf/pull/21867 -* Rewrite TPC-DS Q14 plan to workaround Polars optimizer CSE limitation by @quasiben in https://github.com/rapidsai/cudf/pull/21885 -* Fix CPU PDS* runs by @quasiben in https://github.com/rapidsai/cudf/pull/21899 -* Remove unneeded CUDF_EXPORT from some cudf/detail headers by @davidwendt in https://github.com/rapidsai/cudf/pull/21693 -* Increase cpp-memcheck test timeout by @davidwendt in https://github.com/rapidsai/cudf/pull/21901 -* Run cudf_polars unit tests with RapidsMPF by @mroeschke in https://github.com/rapidsai/cudf/pull/21807 -* Maintain column sorted metadata in result groupby keys in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21871 -* Consolidate/simplify cudf.pandas unit testing script by @mroeschke in https://github.com/rapidsai/cudf/pull/21892 -* Add scoped_range to cudf::benchmark for nvtx ranges in benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/21902 -* Fix mypy pinning on rmm (26.06) by @bdice in https://github.com/rapidsai/cudf/pull/21935 -* Main release/26.04 into main by @mroeschke in https://github.com/rapidsai/cudf/pull/21958 -* Remove paths from cudf-polars Scan Trace properties by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21927 -* Use pip index to fetch polars versions in CI scripts by @mroeschke in https://github.com/rapidsai/cudf/pull/21950 -* Explicitly cancel outstanding tasks in `fanout_node_unbounded` by @mroeschke in https://github.com/rapidsai/cudf/pull/21853 -* Ensure nodes are `del`'d during errors in `run_actor_network` in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21850 -* Ensure rapidsmpf RmmResourceAdaptor is unset after actor network is run in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21856 -* [cudf_polars] Enabling pinned memory for pdsh runs by @nirandaperera in https://github.com/rapidsai/cudf/pull/21880 -* Optimize Polars TPC-DS q50 implementation by @beckernick in https://github.com/rapidsai/cudf/pull/21884 -* Use public gather in libcudf benchmarks and gtests by @davidwendt in https://github.com/rapidsai/cudf/pull/21903 -* Support dynamic error messages in CUDF_EXPECTS and CUDF_FAIL macros by @kingcrimsontianyu in https://github.com/rapidsai/cudf/pull/21900 -* Remove unconditional large left table skip in mark join benchmarks by @PointKernel in https://github.com/rapidsai/cudf/pull/21909 -* [cudf_polars] [MINOR] Configurable dask worker memory in benchmarks by @nirandaperera in https://github.com/rapidsai/cudf/pull/21972 -* Use asyncio.TaskGroup instead of .gather in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21858 -* Replace thrust counting iterators with cuda::counting_iterator by @PointKernel in https://github.com/rapidsai/cudf/pull/21718 -* `review-cudf` skill checks for functions defined in headers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21971 -* Expose ability to hash_partition based on a separate key table by @wence- in https://github.com/rapidsai/cudf/pull/21730 -* Use ruff to disallow asyncio.gather in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21985 -* Remove parquet statistics filter validation logic by @Matt711 in https://github.com/rapidsai/cudf/pull/21736 -* Remove unneeded include of some detail headers in non-internal libcudf code by @davidwendt in https://github.com/rapidsai/cudf/pull/21907 -* Support regex named capture groups in contains, count, match, findall by @davidwendt in https://github.com/rapidsai/cudf/pull/21848 -* [WIP] Better path handling for s3 by @quasiben in https://github.com/rapidsai/cudf/pull/22005 -* Remove cudf::detail::target_type from groupby gtests by @davidwendt in https://github.com/rapidsai/cudf/pull/21932 -* Clean up log messages in Parquet and ORC unit tests by @vuule in https://github.com/rapidsai/cudf/pull/21797 -* Optimize TPC-DS query plans for streaming executor (q4, q23, q64, q75, q78) by @vyasr in https://github.com/rapidsai/cudf/pull/22011 -* [Multi-GPU Polars] Introduce a new Dask frontend by @madsbk in https://github.com/rapidsai/cudf/pull/21812 -* Disable "native" rapidsmpf parquet reader by default by @rjzamora in https://github.com/rapidsai/cudf/pull/22023 -* Pass memory resource explicitly to remove implicit default mr usage (Part 2) by @karthikeyann in https://github.com/rapidsai/cudf/pull/22029 -* Pass memory resource explicitly to remove implicit default mr usage (Part 3) by @karthikeyann in https://github.com/rapidsai/cudf/pull/22030 -* Pass memory resource explicitly to remove implicit default mr usage (Part 1) by @karthikeyann in https://github.com/rapidsai/cudf/pull/22028 -* [Multi-GPU Polars] Fix SPMD bootstrap race with session-scoped communicator by @madsbk in https://github.com/rapidsai/cudf/pull/22015 -* Split hash join definitions to reduce build time by @PointKernel in https://github.com/rapidsai/cudf/pull/21804 -* Deduplicate parquet pass construction by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21923 -* Remove "using namespace cudf" from row_ir jit gtest by @davidwendt in https://github.com/rapidsai/cudf/pull/22069 -* Deprecate build side option for filtered_join by @PointKernel in https://github.com/rapidsai/cudf/pull/21982 -* Pass memory resource to exec_policy_nosync in lists module by @bdice in https://github.com/rapidsai/cudf/pull/22041 -* Pass memory resource to exec_policy_nosync in groupby module by @bdice in https://github.com/rapidsai/cudf/pull/22038 -* Remove more unneeded detail header includes from libcudf benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/22068 -* [Multi-GPU Polars] New frontends accept RMM config by @madsbk in https://github.com/rapidsai/cudf/pull/22052 -* Avoid storing all executor options in `StreamingSink` by @rjzamora in https://github.com/rapidsai/cudf/pull/22079 -* Update to clang 20.1.8 by @bdice in https://github.com/rapidsai/cudf/pull/22093 -* Use cudf::test::iterator utilities instead of make_counting_transform_iterator as appropriate by @davidwendt in https://github.com/rapidsai/cudf/pull/22071 -* Adopt workflow dispatch pattern for compute sanitizer workflows by @davidwendt in https://github.com/rapidsai/cudf/pull/22054 -* [Multi-GPU Polars] Remove all `pytest.skip` calls for unavailable GPU. by @madsbk in https://github.com/rapidsai/cudf/pull/22099 -* Optimize TPC-DS query plans for streaming executor (q80, q31, q11) by @vyasr in https://github.com/rapidsai/cudf/pull/22070 -* Preserve partitioning metadata for ``HStack`` nodes by @rjzamora in https://github.com/rapidsai/cudf/pull/22103 -* Pass memory resource to exec_policy_nosync in join module by @bdice in https://github.com/rapidsai/cudf/pull/22039 -* Fix unsanitized nulls from strings_column_wrapper inputs in gtests by @davidwendt in https://github.com/rapidsai/cudf/pull/22088 -* Use simpler iterators instead of make_counting_transform_iterator by @davidwendt in https://github.com/rapidsai/cudf/pull/22119 -* Remove redundant aggregation identity logic in shared memory groupby by @PointKernel in https://github.com/rapidsai/cudf/pull/22010 -* Optimize TPC-DS query plans for streaming executor (q9, q74) by @vyasr in https://github.com/rapidsai/cudf/pull/22121 -* Add more AST gtests for supported decimal operations by @davidwendt in https://github.com/rapidsai/cudf/pull/22097 -* Rename GroupedRollingWindow as GroupedWindow by @Matt711 in https://github.com/rapidsai/cudf/pull/22135 -* Pass memory resource to exec_policy_nosync in io module by @bdice in https://github.com/rapidsai/cudf/pull/22035 -* Pass memory resource to exec_policy_nosync in text module by @bdice in https://github.com/rapidsai/cudf/pull/22037 -* Pass memory resource to exec_policy_nosync in reductions and quantiles modules by @bdice in https://github.com/rapidsai/cudf/pull/22040 -* Pass memory resource to exec_policy_nosync in copying, rolling, and merge modules by @bdice in https://github.com/rapidsai/cudf/pull/22043 -* Pass memory resource to exec_policy_nosync in dictionary, interop, and replace modules by @bdice in https://github.com/rapidsai/cudf/pull/22044 -* Pass memory resource to exec_policy_nosync in sort, search, stream_compaction, and partitioning modules by @bdice in https://github.com/rapidsai/cudf/pull/22042 -* Pass memory resource to exec_policy_nosync in strings module by @bdice in https://github.com/rapidsai/cudf/pull/22036 -* Remove clang-format-off from nth_element_tests.cpp by @davidwendt in https://github.com/rapidsai/cudf/pull/22100 -* Add multiple_of utilities to cudf::test::iterators by @davidwendt in https://github.com/rapidsai/cudf/pull/22078 -* [Multi-GPU Polars] Add `--num-gpus` for the benchmarks by @madsbk in https://github.com/rapidsai/cudf/pull/22149 -* [Multi-GPU Polars] Unify `num_py_executors` default by @madsbk in https://github.com/rapidsai/cudf/pull/22168 -* Remove verbose from bind() in favor of exceptions in cudf-polars by @pentschev in https://github.com/rapidsai/cudf/pull/22169 -* [Multi-GPU Polars] Add GPU sharing detection by @madsbk in https://github.com/rapidsai/cudf/pull/22148 -* [Multi-GPU Polars] Reduce DaskEngine local cluster log verbosity by @madsbk in https://github.com/rapidsai/cudf/pull/22167 -* Use cudf::sequence instead of make_counting_transform for large gtests columns by @davidwendt in https://github.com/rapidsai/cudf/pull/22106 -* Add cudf-polars-codeowners to CI by @Matt711 in https://github.com/rapidsai/cudf/pull/22192 -* Run Polars unit tests with RapidsMPF by @mroeschke in https://github.com/rapidsai/cudf/pull/21677 -* Implement basic bloom pre-filtering in shuffle join by @wence- in https://github.com/rapidsai/cudf/pull/21931 -* Pass BufferResource to SpillableMessages by @vyasr in https://github.com/rapidsai/cudf/pull/22164 -* Add dedicated stream testing job by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22150 -* Automatically pin `numba-cuda` upper bound at release time in `update-version.sh` by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/21533 -* Enable --collect-traces with new frontends by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22199 -* Support decimal operators with different scales in AST by @davidwendt in https://github.com/rapidsai/cudf/pull/22122 -* fix(ci): remove `needs` dependency on job that doesn't exist in test yaml by @gforsyth in https://github.com/rapidsai/cudf/pull/22211 -* Remove unused device_memory_resource includes by @bdice in https://github.com/rapidsai/cudf/pull/22187 -* Deduplicate lower_ir_graph in cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22220 -* Pass `dask.datasets.timeseries(seed=)` in tests by @mroeschke in https://github.com/rapidsai/cudf/pull/22223 -* [Multi-GPU Polars] Gather statistics by @madsbk in https://github.com/rapidsai/cudf/pull/22210 -* Clean up numba extension code generation by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/22270 -* Ignore ResourceWarning from stumpy for Python 3.14 by @mroeschke in https://github.com/rapidsai/cudf/pull/22272 -* Update deselected polars tests by @wence- in https://github.com/rapidsai/cudf/pull/22265 -* Skip Python/pandas versions without supported wheels by @vyasr in https://github.com/rapidsai/cudf/pull/22273 -* Generalize ``NormalizedPartitioning`` class by @rjzamora in https://github.com/rapidsai/cudf/pull/22246 -* Use the new compute-matrix workflow for stream testing by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22240 -* Agent skill to build and test cudf java by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21894 -* Refactor DataSourceInfo by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22254 -* Record `engine_name` in cudf-polars benchamrks by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22269 -* [MINOR] Remove redundant vectors in `memcpy_batch_async` fast path by @nirandaperera in https://github.com/rapidsai/cudf/pull/22125 -* fix(cmake): exclude kvikio from install, build static, and fix some export issues introduced in 22263 by @vyasr in https://github.com/rapidsai/cudf/pull/22277 -* Skip two more tests by @vyasr in https://github.com/rapidsai/cudf/pull/22294 -* Refactor cudf-polars test suite onto pytest fixtures by @madsbk in https://github.com/rapidsai/cudf/pull/22212 -* Import nvcomp CMake configuration from rapids-cmake into cudf by @vyasr in https://github.com/rapidsai/cudf/pull/22306 -* Reset check-nightly-ci max days without success by @davidwendt in https://github.com/rapidsai/cudf/pull/22303 -* Lower the query plan on workers in cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22287 -* Resolve timezone alias links via tzdata.zi when loading transition tables by @vuule in https://github.com/rapidsai/cudf/pull/22293 -* Skip zero-sized default pinned pool by @bdice in https://github.com/rapidsai/cudf/pull/22292 -* RayEngine: support GPU oversubscription in tests by @madsbk in https://github.com/rapidsai/cudf/pull/22302 -* Address cuml RandomForestClassifier(max_depth=) deprecation by @mroeschke in https://github.com/rapidsai/cudf/pull/22324 -* Add DuckDB resource-limit options to benchmark runner by @Matt711 in https://github.com/rapidsai/cudf/pull/22266 -* Refactor cudf-polars test fixtures away from indirect parametrization by @madsbk in https://github.com/rapidsai/cudf/pull/22325 -* Add call to reset_current_device_resource in gtests fixtures by @davidwendt in https://github.com/rapidsai/cudf/pull/22267 -* cudf-polars: add `RayEngine._reset()` by @madsbk in https://github.com/rapidsai/cudf/pull/22348 -* `StreamingEngine._reset()` by @madsbk in https://github.com/rapidsai/cudf/pull/22364 -* Improve hstack lowering by @rjzamora in https://github.com/rapidsai/cudf/pull/22353 -* Replace `LD_PRELOAD` hack with compute-sanitizer by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22290 -* Run all nvbench benchmarks with timeout in smoketest by @bdice in https://github.com/rapidsai/cudf/pull/20538 -* Rename build/probe to right/left in hash_join and distinct_hash_join by @PointKernel in https://github.com/rapidsai/cudf/pull/22382 -* Use `token.rapids.nvidia.com` when issuing S3 bucket creds in devcontainers by @trxcllnt in https://github.com/rapidsai/cudf/pull/22338 -* Use static cudart by default by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/22397 -* Use cudaStream_t instead of cuda_stream_view in pylibcudf Cython by @vyasr in https://github.com/rapidsai/cudf/pull/22368 -* Use `language: script` for cudf-polars-ir-signatures pre-commit hook by @vyasr in https://github.com/rapidsai/cudf/pull/22384 -* Fix potential errors in Parquet page header decode by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22274 -* Refactor ``sort_actor`` to prepare for ``OrderScheme`` changes by @rjzamora in https://github.com/rapidsai/cudf/pull/22350 -* Run the cudf-polars test suite against `DaskEngine` and `RayEngine` by @madsbk in https://github.com/rapidsai/cudf/pull/22381 -* Move table_device_view function definitions from .cuh to .cu by @davidwendt in https://github.com/rapidsai/cudf/pull/22354 -* Fallback to `async-mr` for the multithreaded parquet example by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/22245 -* fix(ci): resolve all zizmor findings and add zizmor pre-commit checks by @gforsyth in https://github.com/rapidsai/cudf/pull/22343 -* Adopt ``OrderScheme`` metadata in cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/22291 -* Consolidate `evaluate_rapidsmpf` into `evaluate_streaming` in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/22417 -* Add ray run_constraints in cudf_polars conda recipe by @mroeschke in https://github.com/rapidsai/cudf/pull/22414 -* Improve installation hygiene of built and header-only dependencies by @vyasr in https://github.com/rapidsai/cudf/pull/22341 -* Run conda, cudf_polars CI tests with Ray by @mroeschke in https://github.com/rapidsai/cudf/pull/22420 -* Support `Buffer`'s in `HybridScanReader` methods needing `bytes`-like data by @jakirkham in https://github.com/rapidsai/cudf/pull/22345 -* Implement equality of two table_views by @wence- in https://github.com/rapidsai/cudf/pull/22319 -* fix(ci): use sha for the only allowlisted version of action-add-assignees by @gforsyth in https://github.com/rapidsai/cudf/pull/22453 -* Update default memory resource for cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/22426 +* Adapt to rapidsmpf async shuffle changes by @wence- in https://github.com/NVIDIA/cudf/pull/21787 +* Remove CSV reader warnings emitted in unit tests by @vuule in https://github.com/NVIDIA/cudf/pull/21794 +* Remove test_infer_objects_no_reference from cudf_pandas xfail list by @mroeschke in https://github.com/NVIDIA/cudf/pull/21818 +* Change to use non-detail APIs in some libcudf benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/21821 +* Deduplicate libcudf examples CMake files by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21809 +* Merge release/26.04 into main by @davidwendt in https://github.com/NVIDIA/cudf/pull/21815 +* Support drop_nulls unary function in expression decomposition by @quasiben in https://github.com/NVIDIA/cudf/pull/21837 +* [Multi-GPU Polars] Ray mode in PDSH benchmarks by @madsbk in https://github.com/NVIDIA/cudf/pull/21811 +* Improve build time using transform instead of tabulate by @davidwendt in https://github.com/NVIDIA/cudf/pull/21793 +* Use conda packages instead of pip packages in test_narwhals & remove xpassing test_series_setitem from pandas tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/21862 +* Ensure Lineariser channels in scan_node are shutdown on error in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21854 +* Add missing includes for `` and `` by @bdice in https://github.com/NVIDIA/cudf/pull/21859 +* [Multi-GPU Polars] SPMD mode works without `rrun` by @madsbk in https://github.com/NVIDIA/cudf/pull/21851 +* cudf-polars tracing improvements by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21789 +* Fix utf8-to-codepoint utility to handle out-of-range unicode by @davidwendt in https://github.com/NVIDIA/cudf/pull/21823 +* Use `PyBuffer_FillInfo` for `HostBuffer`'s buffer by @jakirkham in https://github.com/NVIDIA/cudf/pull/21855 +* Fix deprecation warning in JNI for parquet_reader_options::builder.names() by @davidwendt in https://github.com/NVIDIA/cudf/pull/21868 +* Add noarch Python channel to cudf.pandas third party tests conda solve by @mroeschke in https://github.com/NVIDIA/cudf/pull/21873 +* [Multi-GPU Polars] Introduce `StreamingEngine` base class and `SPMDEngine` by @madsbk in https://github.com/NVIDIA/cudf/pull/21867 +* Rewrite TPC-DS Q14 plan to workaround Polars optimizer CSE limitation by @quasiben in https://github.com/NVIDIA/cudf/pull/21885 +* Fix CPU PDS* runs by @quasiben in https://github.com/NVIDIA/cudf/pull/21899 +* Remove unneeded CUDF_EXPORT from some cudf/detail headers by @davidwendt in https://github.com/NVIDIA/cudf/pull/21693 +* Increase cpp-memcheck test timeout by @davidwendt in https://github.com/NVIDIA/cudf/pull/21901 +* Run cudf_polars unit tests with RapidsMPF by @mroeschke in https://github.com/NVIDIA/cudf/pull/21807 +* Maintain column sorted metadata in result groupby keys in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21871 +* Consolidate/simplify cudf.pandas unit testing script by @mroeschke in https://github.com/NVIDIA/cudf/pull/21892 +* Add scoped_range to cudf::benchmark for nvtx ranges in benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/21902 +* Fix mypy pinning on rmm (26.06) by @bdice in https://github.com/NVIDIA/cudf/pull/21935 +* Main release/26.04 into main by @mroeschke in https://github.com/NVIDIA/cudf/pull/21958 +* Remove paths from cudf-polars Scan Trace properties by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21927 +* Use pip index to fetch polars versions in CI scripts by @mroeschke in https://github.com/NVIDIA/cudf/pull/21950 +* Explicitly cancel outstanding tasks in `fanout_node_unbounded` by @mroeschke in https://github.com/NVIDIA/cudf/pull/21853 +* Ensure nodes are `del`'d during errors in `run_actor_network` in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21850 +* Ensure rapidsmpf RmmResourceAdaptor is unset after actor network is run in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21856 +* [cudf_polars] Enabling pinned memory for pdsh runs by @nirandaperera in https://github.com/NVIDIA/cudf/pull/21880 +* Optimize Polars TPC-DS q50 implementation by @beckernick in https://github.com/NVIDIA/cudf/pull/21884 +* Use public gather in libcudf benchmarks and gtests by @davidwendt in https://github.com/NVIDIA/cudf/pull/21903 +* Support dynamic error messages in CUDF_EXPECTS and CUDF_FAIL macros by @kingcrimsontianyu in https://github.com/NVIDIA/cudf/pull/21900 +* Remove unconditional large left table skip in mark join benchmarks by @PointKernel in https://github.com/NVIDIA/cudf/pull/21909 +* [cudf_polars] [MINOR] Configurable dask worker memory in benchmarks by @nirandaperera in https://github.com/NVIDIA/cudf/pull/21972 +* Use asyncio.TaskGroup instead of .gather in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21858 +* Replace thrust counting iterators with cuda::counting_iterator by @PointKernel in https://github.com/NVIDIA/cudf/pull/21718 +* `review-cudf` skill checks for functions defined in headers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21971 +* Expose ability to hash_partition based on a separate key table by @wence- in https://github.com/NVIDIA/cudf/pull/21730 +* Use ruff to disallow asyncio.gather in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21985 +* Remove parquet statistics filter validation logic by @Matt711 in https://github.com/NVIDIA/cudf/pull/21736 +* Remove unneeded include of some detail headers in non-internal libcudf code by @davidwendt in https://github.com/NVIDIA/cudf/pull/21907 +* Support regex named capture groups in contains, count, match, findall by @davidwendt in https://github.com/NVIDIA/cudf/pull/21848 +* [WIP] Better path handling for s3 by @quasiben in https://github.com/NVIDIA/cudf/pull/22005 +* Remove cudf::detail::target_type from groupby gtests by @davidwendt in https://github.com/NVIDIA/cudf/pull/21932 +* Clean up log messages in Parquet and ORC unit tests by @vuule in https://github.com/NVIDIA/cudf/pull/21797 +* Optimize TPC-DS query plans for streaming executor (q4, q23, q64, q75, q78) by @vyasr in https://github.com/NVIDIA/cudf/pull/22011 +* [Multi-GPU Polars] Introduce a new Dask frontend by @madsbk in https://github.com/NVIDIA/cudf/pull/21812 +* Disable "native" rapidsmpf parquet reader by default by @rjzamora in https://github.com/NVIDIA/cudf/pull/22023 +* Pass memory resource explicitly to remove implicit default mr usage (Part 2) by @karthikeyann in https://github.com/NVIDIA/cudf/pull/22029 +* Pass memory resource explicitly to remove implicit default mr usage (Part 3) by @karthikeyann in https://github.com/NVIDIA/cudf/pull/22030 +* Pass memory resource explicitly to remove implicit default mr usage (Part 1) by @karthikeyann in https://github.com/NVIDIA/cudf/pull/22028 +* [Multi-GPU Polars] Fix SPMD bootstrap race with session-scoped communicator by @madsbk in https://github.com/NVIDIA/cudf/pull/22015 +* Split hash join definitions to reduce build time by @PointKernel in https://github.com/NVIDIA/cudf/pull/21804 +* Deduplicate parquet pass construction by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21923 +* Remove "using namespace cudf" from row_ir jit gtest by @davidwendt in https://github.com/NVIDIA/cudf/pull/22069 +* Deprecate build side option for filtered_join by @PointKernel in https://github.com/NVIDIA/cudf/pull/21982 +* Pass memory resource to exec_policy_nosync in lists module by @bdice in https://github.com/NVIDIA/cudf/pull/22041 +* Pass memory resource to exec_policy_nosync in groupby module by @bdice in https://github.com/NVIDIA/cudf/pull/22038 +* Remove more unneeded detail header includes from libcudf benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/22068 +* [Multi-GPU Polars] New frontends accept RMM config by @madsbk in https://github.com/NVIDIA/cudf/pull/22052 +* Avoid storing all executor options in `StreamingSink` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22079 +* Update to clang 20.1.8 by @bdice in https://github.com/NVIDIA/cudf/pull/22093 +* Use cudf::test::iterator utilities instead of make_counting_transform_iterator as appropriate by @davidwendt in https://github.com/NVIDIA/cudf/pull/22071 +* Adopt workflow dispatch pattern for compute sanitizer workflows by @davidwendt in https://github.com/NVIDIA/cudf/pull/22054 +* [Multi-GPU Polars] Remove all `pytest.skip` calls for unavailable GPU. by @madsbk in https://github.com/NVIDIA/cudf/pull/22099 +* Optimize TPC-DS query plans for streaming executor (q80, q31, q11) by @vyasr in https://github.com/NVIDIA/cudf/pull/22070 +* Preserve partitioning metadata for ``HStack`` nodes by @rjzamora in https://github.com/NVIDIA/cudf/pull/22103 +* Pass memory resource to exec_policy_nosync in join module by @bdice in https://github.com/NVIDIA/cudf/pull/22039 +* Fix unsanitized nulls from strings_column_wrapper inputs in gtests by @davidwendt in https://github.com/NVIDIA/cudf/pull/22088 +* Use simpler iterators instead of make_counting_transform_iterator by @davidwendt in https://github.com/NVIDIA/cudf/pull/22119 +* Remove redundant aggregation identity logic in shared memory groupby by @PointKernel in https://github.com/NVIDIA/cudf/pull/22010 +* Optimize TPC-DS query plans for streaming executor (q9, q74) by @vyasr in https://github.com/NVIDIA/cudf/pull/22121 +* Add more AST gtests for supported decimal operations by @davidwendt in https://github.com/NVIDIA/cudf/pull/22097 +* Rename GroupedRollingWindow as GroupedWindow by @Matt711 in https://github.com/NVIDIA/cudf/pull/22135 +* Pass memory resource to exec_policy_nosync in io module by @bdice in https://github.com/NVIDIA/cudf/pull/22035 +* Pass memory resource to exec_policy_nosync in text module by @bdice in https://github.com/NVIDIA/cudf/pull/22037 +* Pass memory resource to exec_policy_nosync in reductions and quantiles modules by @bdice in https://github.com/NVIDIA/cudf/pull/22040 +* Pass memory resource to exec_policy_nosync in copying, rolling, and merge modules by @bdice in https://github.com/NVIDIA/cudf/pull/22043 +* Pass memory resource to exec_policy_nosync in dictionary, interop, and replace modules by @bdice in https://github.com/NVIDIA/cudf/pull/22044 +* Pass memory resource to exec_policy_nosync in sort, search, stream_compaction, and partitioning modules by @bdice in https://github.com/NVIDIA/cudf/pull/22042 +* Pass memory resource to exec_policy_nosync in strings module by @bdice in https://github.com/NVIDIA/cudf/pull/22036 +* Remove clang-format-off from nth_element_tests.cpp by @davidwendt in https://github.com/NVIDIA/cudf/pull/22100 +* Add multiple_of utilities to cudf::test::iterators by @davidwendt in https://github.com/NVIDIA/cudf/pull/22078 +* [Multi-GPU Polars] Add `--num-gpus` for the benchmarks by @madsbk in https://github.com/NVIDIA/cudf/pull/22149 +* [Multi-GPU Polars] Unify `num_py_executors` default by @madsbk in https://github.com/NVIDIA/cudf/pull/22168 +* Remove verbose from bind() in favor of exceptions in cudf-polars by @pentschev in https://github.com/NVIDIA/cudf/pull/22169 +* [Multi-GPU Polars] Add GPU sharing detection by @madsbk in https://github.com/NVIDIA/cudf/pull/22148 +* [Multi-GPU Polars] Reduce DaskEngine local cluster log verbosity by @madsbk in https://github.com/NVIDIA/cudf/pull/22167 +* Use cudf::sequence instead of make_counting_transform for large gtests columns by @davidwendt in https://github.com/NVIDIA/cudf/pull/22106 +* Add cudf-polars-codeowners to CI by @Matt711 in https://github.com/NVIDIA/cudf/pull/22192 +* Run Polars unit tests with RapidsMPF by @mroeschke in https://github.com/NVIDIA/cudf/pull/21677 +* Implement basic bloom pre-filtering in shuffle join by @wence- in https://github.com/NVIDIA/cudf/pull/21931 +* Pass BufferResource to SpillableMessages by @vyasr in https://github.com/NVIDIA/cudf/pull/22164 +* Add dedicated stream testing job by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22150 +* Automatically pin `numba-cuda` upper bound at release time in `update-version.sh` by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/21533 +* Enable --collect-traces with new frontends by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22199 +* Support decimal operators with different scales in AST by @davidwendt in https://github.com/NVIDIA/cudf/pull/22122 +* fix(ci): remove `needs` dependency on job that doesn't exist in test yaml by @gforsyth in https://github.com/NVIDIA/cudf/pull/22211 +* Remove unused device_memory_resource includes by @bdice in https://github.com/NVIDIA/cudf/pull/22187 +* Deduplicate lower_ir_graph in cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22220 +* Pass `dask.datasets.timeseries(seed=)` in tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/22223 +* [Multi-GPU Polars] Gather statistics by @madsbk in https://github.com/NVIDIA/cudf/pull/22210 +* Clean up numba extension code generation by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/22270 +* Ignore ResourceWarning from stumpy for Python 3.14 by @mroeschke in https://github.com/NVIDIA/cudf/pull/22272 +* Update deselected polars tests by @wence- in https://github.com/NVIDIA/cudf/pull/22265 +* Skip Python/pandas versions without supported wheels by @vyasr in https://github.com/NVIDIA/cudf/pull/22273 +* Generalize ``NormalizedPartitioning`` class by @rjzamora in https://github.com/NVIDIA/cudf/pull/22246 +* Use the new compute-matrix workflow for stream testing by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22240 +* Agent skill to build and test cudf java by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21894 +* Refactor DataSourceInfo by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22254 +* Record `engine_name` in cudf-polars benchamrks by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22269 +* [MINOR] Remove redundant vectors in `memcpy_batch_async` fast path by @nirandaperera in https://github.com/NVIDIA/cudf/pull/22125 +* fix(cmake): exclude kvikio from install, build static, and fix some export issues introduced in 22263 by @vyasr in https://github.com/NVIDIA/cudf/pull/22277 +* Skip two more tests by @vyasr in https://github.com/NVIDIA/cudf/pull/22294 +* Refactor cudf-polars test suite onto pytest fixtures by @madsbk in https://github.com/NVIDIA/cudf/pull/22212 +* Import nvcomp CMake configuration from rapids-cmake into cudf by @vyasr in https://github.com/NVIDIA/cudf/pull/22306 +* Reset check-nightly-ci max days without success by @davidwendt in https://github.com/NVIDIA/cudf/pull/22303 +* Lower the query plan on workers in cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22287 +* Resolve timezone alias links via tzdata.zi when loading transition tables by @vuule in https://github.com/NVIDIA/cudf/pull/22293 +* Skip zero-sized default pinned pool by @bdice in https://github.com/NVIDIA/cudf/pull/22292 +* RayEngine: support GPU oversubscription in tests by @madsbk in https://github.com/NVIDIA/cudf/pull/22302 +* Address cuml RandomForestClassifier(max_depth=) deprecation by @mroeschke in https://github.com/NVIDIA/cudf/pull/22324 +* Add DuckDB resource-limit options to benchmark runner by @Matt711 in https://github.com/NVIDIA/cudf/pull/22266 +* Refactor cudf-polars test fixtures away from indirect parametrization by @madsbk in https://github.com/NVIDIA/cudf/pull/22325 +* Add call to reset_current_device_resource in gtests fixtures by @davidwendt in https://github.com/NVIDIA/cudf/pull/22267 +* cudf-polars: add `RayEngine._reset()` by @madsbk in https://github.com/NVIDIA/cudf/pull/22348 +* `StreamingEngine._reset()` by @madsbk in https://github.com/NVIDIA/cudf/pull/22364 +* Improve hstack lowering by @rjzamora in https://github.com/NVIDIA/cudf/pull/22353 +* Replace `LD_PRELOAD` hack with compute-sanitizer by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22290 +* Run all nvbench benchmarks with timeout in smoketest by @bdice in https://github.com/NVIDIA/cudf/pull/20538 +* Rename build/probe to right/left in hash_join and distinct_hash_join by @PointKernel in https://github.com/NVIDIA/cudf/pull/22382 +* Use `token.rapids.nvidia.com` when issuing S3 bucket creds in devcontainers by @trxcllnt in https://github.com/NVIDIA/cudf/pull/22338 +* Use static cudart by default by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/22397 +* Use cudaStream_t instead of cuda_stream_view in pylibcudf Cython by @vyasr in https://github.com/NVIDIA/cudf/pull/22368 +* Use `language: script` for cudf-polars-ir-signatures pre-commit hook by @vyasr in https://github.com/NVIDIA/cudf/pull/22384 +* Fix potential errors in Parquet page header decode by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22274 +* Refactor ``sort_actor`` to prepare for ``OrderScheme`` changes by @rjzamora in https://github.com/NVIDIA/cudf/pull/22350 +* Run the cudf-polars test suite against `DaskEngine` and `RayEngine` by @madsbk in https://github.com/NVIDIA/cudf/pull/22381 +* Move table_device_view function definitions from .cuh to .cu by @davidwendt in https://github.com/NVIDIA/cudf/pull/22354 +* Fallback to `async-mr` for the multithreaded parquet example by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/22245 +* fix(ci): resolve all zizmor findings and add zizmor pre-commit checks by @gforsyth in https://github.com/NVIDIA/cudf/pull/22343 +* Adopt ``OrderScheme`` metadata in cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/22291 +* Consolidate `evaluate_rapidsmpf` into `evaluate_streaming` in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/22417 +* Add ray run_constraints in cudf_polars conda recipe by @mroeschke in https://github.com/NVIDIA/cudf/pull/22414 +* Improve installation hygiene of built and header-only dependencies by @vyasr in https://github.com/NVIDIA/cudf/pull/22341 +* Run conda, cudf_polars CI tests with Ray by @mroeschke in https://github.com/NVIDIA/cudf/pull/22420 +* Support `Buffer`'s in `HybridScanReader` methods needing `bytes`-like data by @jakirkham in https://github.com/NVIDIA/cudf/pull/22345 +* Implement equality of two table_views by @wence- in https://github.com/NVIDIA/cudf/pull/22319 +* fix(ci): use sha for the only allowlisted version of action-add-assignees by @gforsyth in https://github.com/NVIDIA/cudf/pull/22453 +* Update default memory resource for cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/22426 * fix(ci): add explicit `actions: write` permission for `telemetry-summarize` - by @gforsyth in https://github.com/rapidsai/cudf/pull/22479 -* Split PR devcontainer CI into pip and conda jobs by @bdice in https://github.com/rapidsai/cudf/pull/22490 -* Undo some CCCL workarounds fixed in the latest update by @davidwendt in https://github.com/rapidsai/cudf/pull/22475 -* Build and test with CUDA 13.2.0 by @bdice in https://github.com/rapidsai/cudf/pull/22463 -* Remove `wheel-tests-cudf-polars-with-rapidsmpf` in favor of existing `wheel-tests-cudf-polars` by @mroeschke in https://github.com/rapidsai/cudf/pull/22467 -* Remove anonymous namespaces from cudf headers by @PointKernel in https://github.com/rapidsai/cudf/pull/22418 -* Relax ``NormalizedPartitioning.from_keys`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22483 -* Reduce peak footprint of cudf-polars test memory usage by @wence- in https://github.com/rapidsai/cudf/pull/22493 -* Use basic ``OrderScheme`` metadata in ``sort_actor`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22477 -* Add `pinned_max_pool_size` and `unbounded_file_read_cache` to `StreamingOptions` by @madsbk in https://github.com/rapidsai/cudf/pull/22501 -* Implement our own to_thread offload for cudf-polars streaming execution by @wence- in https://github.com/rapidsai/cudf/pull/22474 -* Remove ``--broadcast-join-limit`` by @rjzamora in https://github.com/rapidsai/cudf/pull/22499 -* Revert `FD_GROUPBY_REWRITE` in TPC-DS benchmark queries by @Matt711 in https://github.com/rapidsai/cudf/pull/22525 -* Bump polars upper bound to <1.40 by @Matt711 in https://github.com/rapidsai/cudf/pull/22048 -* More Polars plan optimizations for TPC-DS by @Matt711 in https://github.com/rapidsai/cudf/pull/22395 -* Remove bad algorithmic behaviour when reserving collective IDs by @wence- in https://github.com/rapidsai/cudf/pull/22604 -* Add configuration hints when ``run_actor_network`` raises a memory error by @rjzamora in https://github.com/rapidsai/cudf/pull/22561 -* Simplify validation in cudf-polars benchmark by @wence- in https://github.com/rapidsai/cudf/pull/22600 -* skip CuPy 14.1.0 by @jameslamb in https://github.com/rapidsai/cudf/pull/22702 + by @gforsyth in https://github.com/NVIDIA/cudf/pull/22479 +* Split PR devcontainer CI into pip and conda jobs by @bdice in https://github.com/NVIDIA/cudf/pull/22490 +* Undo some CCCL workarounds fixed in the latest update by @davidwendt in https://github.com/NVIDIA/cudf/pull/22475 +* Build and test with CUDA 13.2.0 by @bdice in https://github.com/NVIDIA/cudf/pull/22463 +* Remove `wheel-tests-cudf-polars-with-rapidsmpf` in favor of existing `wheel-tests-cudf-polars` by @mroeschke in https://github.com/NVIDIA/cudf/pull/22467 +* Remove anonymous namespaces from cudf headers by @PointKernel in https://github.com/NVIDIA/cudf/pull/22418 +* Relax ``NormalizedPartitioning.from_keys`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22483 +* Reduce peak footprint of cudf-polars test memory usage by @wence- in https://github.com/NVIDIA/cudf/pull/22493 +* Use basic ``OrderScheme`` metadata in ``sort_actor`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22477 +* Add `pinned_max_pool_size` and `unbounded_file_read_cache` to `StreamingOptions` by @madsbk in https://github.com/NVIDIA/cudf/pull/22501 +* Implement our own to_thread offload for cudf-polars streaming execution by @wence- in https://github.com/NVIDIA/cudf/pull/22474 +* Remove ``--broadcast-join-limit`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/22499 +* Revert `FD_GROUPBY_REWRITE` in TPC-DS benchmark queries by @Matt711 in https://github.com/NVIDIA/cudf/pull/22525 +* Bump polars upper bound to <1.40 by @Matt711 in https://github.com/NVIDIA/cudf/pull/22048 +* More Polars plan optimizations for TPC-DS by @Matt711 in https://github.com/NVIDIA/cudf/pull/22395 +* Remove bad algorithmic behaviour when reserving collective IDs by @wence- in https://github.com/NVIDIA/cudf/pull/22604 +* Add configuration hints when ``run_actor_network`` raises a memory error by @rjzamora in https://github.com/NVIDIA/cudf/pull/22561 +* Simplify validation in cudf-polars benchmark by @wence- in https://github.com/NVIDIA/cudf/pull/22600 +* skip CuPy 14.1.0 by @jameslamb in https://github.com/NVIDIA/cudf/pull/22702 ## New Contributors -* @eternallyproud made their first contribution in https://github.com/rapidsai/cudf/pull/21946 -* @pramodsatya made their first contribution in https://github.com/rapidsai/cudf/pull/22402 +* @eternallyproud made their first contribution in https://github.com/NVIDIA/cudf/pull/21946 +* @pramodsatya made their first contribution in https://github.com/NVIDIA/cudf/pull/22402 -**Full Changelog**: https://github.com/rapidsai/cudf/compare/v26.06.00a...release/26.06 +**Full Changelog**: https://github.com/NVIDIA/cudf/compare/v26.06.00a...release/26.06 # cudf 26.04.00 (8 Apr 2026) ### 🚨 Breaking Changes -* Make offset vectors consistent across partitioning APIs by @PointKernel in https://github.com/rapidsai/cudf/pull/20937 -* Remove unneeded parameters from make_dictionary_column by @davidwendt in https://github.com/rapidsai/cudf/pull/21032 -* Deprecate nvtext::byte_pair_encoding APIs by @davidwendt in https://github.com/rapidsai/cudf/pull/21157 -* Require pyarrow 19 or greater by @vyasr in https://github.com/rapidsai/cudf/pull/21183 -* Accept device spans instead of buffers in hybrid scan APIs. by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21026 -* Deprecate nvtext::edit_distance_matrix API by @davidwendt in https://github.com/rapidsai/cudf/pull/21159 -* Enable parquet readers to select top-level cols by index by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21113 -* Hybrid scan APIs to accept an `mr` to allocate output tables by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21263 -* Add pylibcudf interface to cudf::round_decimal by @davidwendt in https://github.com/rapidsai/cudf/pull/21332 -* Use the new `make_table_chunks_available_or_wait` API from RapidsMPF by @madsbk in https://github.com/rapidsai/cudf/pull/21291 -* [BUGFIX] Make cudf's context mt-safe on initialization by @lamarrr in https://github.com/rapidsai/cudf/pull/21008 -* Enforce thread safety in sort-merge join APIs by @shrshi in https://github.com/rapidsai/cudf/pull/21314 -* Change dictionary columns to support unordered keys by @davidwendt in https://github.com/rapidsai/cudf/pull/21058 -* Add cudf::column_nans_to_nulls API by @davidwendt in https://github.com/rapidsai/cudf/pull/21331 -* Relocate `distinct_count` and `unique_count` from `stream_compaction` to `reduction` by @PointKernel in https://github.com/rapidsai/cudf/pull/21372 -* Change missing value repr for string from to None by @mroeschke in https://github.com/rapidsai/cudf/pull/21425 -* Make device_span an alias of cuda::std::span by @PointKernel in https://github.com/rapidsai/cudf/pull/21283 -* Preserve pandas nullable types in `dtype=` arguments by @mroeschke in https://github.com/rapidsai/cudf/pull/21499 -* [BUGFIX] Fix handling of scalar inputs in transform and row_ir by @lamarrr in https://github.com/rapidsai/cudf/pull/21443 -* Preserve pandas nullable types from pandas objects by @mroeschke in https://github.com/rapidsai/cudf/pull/21514 -* Remove unneeded stream/mr parameters from make_lists_column by @davidwendt in https://github.com/rapidsai/cudf/pull/21548 -* Bump minimum NumPy dependency to 1.26 by @mroeschke in https://github.com/rapidsai/cudf/pull/21607 -* Prune uneeded dependencies from pyproject.toml's and recipe.yml's by @mroeschke in https://github.com/rapidsai/cudf/pull/21608 -* Adapt to rapidsmpf changes splitting communicator and streaming context by @wence- in https://github.com/rapidsai/cudf/pull/21632 -* RapidsMPF SPMD mode by @madsbk in https://github.com/rapidsai/cudf/pull/21619 -* Improve approx_distinct_count to work with preallocated sketch by @PointKernel in https://github.com/rapidsai/cudf/pull/21396 -* [BUG] [cudf-polars] Remove rapidsmpf concat insert option by @nirandaperera in https://github.com/rapidsai/cudf/pull/21761 -* Bump librdkafka to >=2.11.1,<2.12.0 by @mroeschke in https://github.com/rapidsai/cudf/pull/21951 +* Make offset vectors consistent across partitioning APIs by @PointKernel in https://github.com/NVIDIA/cudf/pull/20937 +* Remove unneeded parameters from make_dictionary_column by @davidwendt in https://github.com/NVIDIA/cudf/pull/21032 +* Deprecate nvtext::byte_pair_encoding APIs by @davidwendt in https://github.com/NVIDIA/cudf/pull/21157 +* Require pyarrow 19 or greater by @vyasr in https://github.com/NVIDIA/cudf/pull/21183 +* Accept device spans instead of buffers in hybrid scan APIs. by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21026 +* Deprecate nvtext::edit_distance_matrix API by @davidwendt in https://github.com/NVIDIA/cudf/pull/21159 +* Enable parquet readers to select top-level cols by index by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21113 +* Hybrid scan APIs to accept an `mr` to allocate output tables by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21263 +* Add pylibcudf interface to cudf::round_decimal by @davidwendt in https://github.com/NVIDIA/cudf/pull/21332 +* Use the new `make_table_chunks_available_or_wait` API from RapidsMPF by @madsbk in https://github.com/NVIDIA/cudf/pull/21291 +* [BUGFIX] Make cudf's context mt-safe on initialization by @lamarrr in https://github.com/NVIDIA/cudf/pull/21008 +* Enforce thread safety in sort-merge join APIs by @shrshi in https://github.com/NVIDIA/cudf/pull/21314 +* Change dictionary columns to support unordered keys by @davidwendt in https://github.com/NVIDIA/cudf/pull/21058 +* Add cudf::column_nans_to_nulls API by @davidwendt in https://github.com/NVIDIA/cudf/pull/21331 +* Relocate `distinct_count` and `unique_count` from `stream_compaction` to `reduction` by @PointKernel in https://github.com/NVIDIA/cudf/pull/21372 +* Change missing value repr for string from to None by @mroeschke in https://github.com/NVIDIA/cudf/pull/21425 +* Make device_span an alias of cuda::std::span by @PointKernel in https://github.com/NVIDIA/cudf/pull/21283 +* Preserve pandas nullable types in `dtype=` arguments by @mroeschke in https://github.com/NVIDIA/cudf/pull/21499 +* [BUGFIX] Fix handling of scalar inputs in transform and row_ir by @lamarrr in https://github.com/NVIDIA/cudf/pull/21443 +* Preserve pandas nullable types from pandas objects by @mroeschke in https://github.com/NVIDIA/cudf/pull/21514 +* Remove unneeded stream/mr parameters from make_lists_column by @davidwendt in https://github.com/NVIDIA/cudf/pull/21548 +* Bump minimum NumPy dependency to 1.26 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21607 +* Prune uneeded dependencies from pyproject.toml's and recipe.yml's by @mroeschke in https://github.com/NVIDIA/cudf/pull/21608 +* Adapt to rapidsmpf changes splitting communicator and streaming context by @wence- in https://github.com/NVIDIA/cudf/pull/21632 +* RapidsMPF SPMD mode by @madsbk in https://github.com/NVIDIA/cudf/pull/21619 +* Improve approx_distinct_count to work with preallocated sketch by @PointKernel in https://github.com/NVIDIA/cudf/pull/21396 +* [BUG] [cudf-polars] Remove rapidsmpf concat insert option by @nirandaperera in https://github.com/NVIDIA/cudf/pull/21761 +* Bump librdkafka to >=2.11.1,<2.12.0 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21951 ### 🐛 Bug Fixes -* Add missing synchronisation in decode_delta_byte_array_kernel by @wence- in https://github.com/rapidsai/cudf/pull/21051 -* Fixes index-based filtering of left join indices by @shrshi in https://github.com/rapidsai/cudf/pull/20990 -* Fix issues in CSV reader benchmarks by @vuule in https://github.com/rapidsai/cudf/pull/21025 -* Fix .first/last cudf_polars expr with empty columns by @mroeschke in https://github.com/rapidsai/cudf/pull/21089 -* Switch date type in pandas PDSH queries by @Matt711 in https://github.com/rapidsai/cudf/pull/21101 -* Preserve partitioning information in `Cache` nodes by @rjzamora in https://github.com/rapidsai/cudf/pull/21098 -* Use Dask "context" to populate rapidsmpf context for proper statistics by @quasiben in https://github.com/rapidsai/cudf/pull/21126 -* Change output dtype to `int32` for datetime component methods by @Matt711 in https://github.com/rapidsai/cudf/pull/21099 -* Update column names in pdsh query 18 by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21119 -* Preserve NA values in decimal ArrowExtensionArray comparisons by @Matt711 in https://github.com/rapidsai/cudf/pull/21168 -* Handle empty format string in dt.strftime for GPU engine in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21143 -* Handle Greek final sigma in `str.lower()` by @Matt711 in https://github.com/rapidsai/cudf/pull/21154 -* Preserve type metadata in unary ops by @Matt711 in https://github.com/rapidsai/cudf/pull/21170 -* Fix the host-device access to device_span data in cumulative_centroid_weight_grouped by @vuule in https://github.com/rapidsai/cudf/pull/21163 -* Make JNI partition output consistent with cuDF by @zpuller in https://github.com/rapidsai/cudf/pull/21029 -* Ensure empty windows return NaN for rolling aggregations by @Matt711 in https://github.com/rapidsai/cudf/pull/20677 -* Fix `at`/`iat` indexers to reject list-like keys by @Matt711 in https://github.com/rapidsai/cudf/pull/21213 -* Vendor Pandas' to_xarray in cudf.pandas by @Matt711 in https://github.com/rapidsai/cudf/pull/21175 -* Raise error for string slicing on non-monotonic DatetimeIndex by @Matt711 in https://github.com/rapidsai/cudf/pull/21218 -* Add `__array_priority__` to proxy scalar types to match pandas by @Matt711 in https://github.com/rapidsai/cudf/pull/21235 -* Change find_first_set_bit_kernel to use BlockReduce instead of cg::reduce by @davidwendt in https://github.com/rapidsai/cudf/pull/21234 -* Add noinline declaration to secondary type-dispatching row-operators in Debug build by @davidwendt in https://github.com/rapidsai/cudf/pull/21197 -* Fix casting to StringDtype for ArrowExtensionArray duration and decimal types by @Matt711 in https://github.com/rapidsai/cudf/pull/21230 -* Fix `pl.Expr.concat_str` with boolean columns in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/21145 -* Only serialize column slice by @pentschev in https://github.com/rapidsai/cudf/pull/21301 -* Fix return type of str.ip2int to uint32 by @mroeschke in https://github.com/rapidsai/cudf/pull/21324 -* Fix integer overflow in `cudf::hash_partition` by @wence- in https://github.com/rapidsai/cudf/pull/21298 -* Remove pandas compatibility gating for dtypes in `datetime` and `timedelta` columns by @galipremsagar in https://github.com/rapidsai/cudf/pull/21252 -* Fix how cudf-polars tracing finds dataframe inputs by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20647 -* Use ColumnBase.create in from_arrow, follow-ups in pyarrow 19 decimal support by @mroeschke in https://github.com/rapidsai/cudf/pull/21317 -* Fix validation for datetimes coming from arrow by @vyasr in https://github.com/rapidsai/cudf/pull/21369 -* Fix n_workers detection in tpc-h benchmark by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21368 -* Fix list literal translation by @Matt711 in https://github.com/rapidsai/cudf/pull/20801 -* Fix exponentially weighted mean with leading NULLs by @Matt711 in https://github.com/rapidsai/cudf/pull/20556 -* Fix various issues in Python microbenchmarks by @vyasr in https://github.com/rapidsai/cudf/pull/21340 -* Allow np.dtype(str) to coerce to object type for cudf column construction by @mroeschke in https://github.com/rapidsai/cudf/pull/21403 -* Remove 4-byte padding requirement from the Brotli decompressor by @vuule in https://github.com/rapidsai/cudf/pull/21392 -* Replace `thrust::get` with `cuda::std::get` by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21416 -* Tests all APIs docstrings in doctests by @galipremsagar in https://github.com/rapidsai/cudf/pull/21404 -* [BUG] Fixed a bug in the 10min notebook by @drivanov in https://github.com/rapidsai/cudf/pull/21400 -* Fix empty string generation in get_json_object by @vyasr in https://github.com/rapidsai/cudf/pull/21418 -* Save duckdb pdsh results to output file by @Matt711 in https://github.com/rapidsai/cudf/pull/21391 -* Pin cupy<14 for wheel-tests-cudf and cudf-pandas unit tests by @Matt711 in https://github.com/rapidsai/cudf/pull/21469 -* Fix generic_scalar_device_view::value() to honor decimal scale by @davidwendt in https://github.com/rapidsai/cudf/pull/21447 -* Add notes to failing polars test by @vyasr in https://github.com/rapidsai/cudf/pull/21488 -* Support CuPy 14 by @Matt711 in https://github.com/rapidsai/cudf/pull/21504 -* Fix misaligned memory access in groupby by @PointKernel in https://github.com/rapidsai/cudf/pull/21513 -* Deterministic build and probe table generation for join benchmarks by @shrshi in https://github.com/rapidsai/cudf/pull/21410 -* Fix pds-h benchmark with `--suffix` by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21520 -* Fix OOB read in chunked Parquet reads with `skip_rows`/`num_rows` by @vuule in https://github.com/rapidsai/cudf/pull/21407 -* rapidsmpf compatibility in unspill_partitions by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21539 -* Handle edge case dectection in pds-h validation by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21544 -* Fix scan/ewm logic to pass stream to column_device_view::create by @davidwendt in https://github.com/rapidsai/cudf/pull/21557 -* Account for the rank index in streaming ``Sink`` by @rjzamora in https://github.com/rapidsai/cudf/pull/21555 -* Fix decimal precision loss in cast: float->decimal by @Matt711 in https://github.com/rapidsai/cudf/pull/21450 -* Fix JIT filtering of join indices to properly handle nulls in nullable columns by @shrshi in https://github.com/rapidsai/cudf/pull/21570 -* Fix span index-out-of-range error/assert in parquet::decode_page_data_generic kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/21575 -* Add stream parameter to indexalator make_input_optional_iterator by @davidwendt in https://github.com/rapidsai/cudf/pull/21545 -* Adapt to latest rapidsmpf API changes by @wence- in https://github.com/rapidsai/cudf/pull/21601 -* Usage of `template` keyword required for some nvcc and gcc combinations by @robertmaynard in https://github.com/rapidsai/cudf/pull/21577 -* Fix caching the result of is_sorted by @jberg5 in https://github.com/rapidsai/cudf/pull/21603 -* Small bugfix: json sink logic was being evaluated incorrectly by @jberg5 in https://github.com/rapidsai/cudf/pull/21605 -* fix: correct typos in comments by @thecaptain789 in https://github.com/rapidsai/cudf/pull/21380 -* Fix chunked_pack::next() returning overflowed int for >2GB data by @ttnghia in https://github.com/rapidsai/cudf/pull/21643 -* Adjust precision in pds-h query 11 by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21651 -* Fix segfault in `HYBRID_SCAN_MULTITHREAD_NVBENCH` by @ttnghia in https://github.com/rapidsai/cudf/pull/21661 -* Improve attribute lookup by 500x in `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21656 -* Fix `pl.Expr.split_exact` children result for null data by @mroeschke in https://github.com/rapidsai/cudf/pull/21660 -* Increase timeout for the devcontainer build workflow by @davidwendt in https://github.com/rapidsai/cudf/pull/21713 -* Support large partition counts in `hash_partition` by @ttnghia in https://github.com/rapidsai/cudf/pull/21674 -* Use larger datatypes when calculating mean and other groupby-aggs by @quasiben in https://github.com/rapidsai/cudf/pull/21719 -* Remove default stream use in PQ reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21629 -* Fix dynamic-join logic for minimum shuffle modulus by @rjzamora in https://github.com/rapidsai/cudf/pull/21712 -* Support fixed point types in Parquet predicate pushdown by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21681 -* Fix inherited partitioning metadata in cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/21748 -* Fix validation failures in PDS-DS Q23 and Q61 by @Matt711 in https://github.com/rapidsai/cudf/pull/21754 -* Fix validation failures in Q69 and Q93 by @Matt711 in https://github.com/rapidsai/cudf/pull/21758 -* Fix thread pool worker threads using wrong CUDA device in multi-GPU environments by @vuule in https://github.com/rapidsai/cudf/pull/21745 -* Fix quantile on decimals by @jberg5 in https://github.com/rapidsai/cudf/pull/21707 -* Add missing return after state.skip() in libcudf benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/21756 -* [BUG] Fix parsing of ld.param PTX instruction in UDFs by @lamarrr in https://github.com/rapidsai/cudf/pull/21597 -* Enable `astype(str)` in `cudf` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21699 -* Fix racecheck in join mark_retrieve_kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/21786 -* Fix approx distinct count nan handling by @PointKernel in https://github.com/rapidsai/cudf/pull/21806 -* Fix mypy pinning on rmm by @bdice in https://github.com/rapidsai/cudf/pull/21925 -* Fix JIT transform kernel crash on non-default CUDA streams by @bdice in https://github.com/rapidsai/cudf/pull/21921 +* Add missing synchronisation in decode_delta_byte_array_kernel by @wence- in https://github.com/NVIDIA/cudf/pull/21051 +* Fixes index-based filtering of left join indices by @shrshi in https://github.com/NVIDIA/cudf/pull/20990 +* Fix issues in CSV reader benchmarks by @vuule in https://github.com/NVIDIA/cudf/pull/21025 +* Fix .first/last cudf_polars expr with empty columns by @mroeschke in https://github.com/NVIDIA/cudf/pull/21089 +* Switch date type in pandas PDSH queries by @Matt711 in https://github.com/NVIDIA/cudf/pull/21101 +* Preserve partitioning information in `Cache` nodes by @rjzamora in https://github.com/NVIDIA/cudf/pull/21098 +* Use Dask "context" to populate rapidsmpf context for proper statistics by @quasiben in https://github.com/NVIDIA/cudf/pull/21126 +* Change output dtype to `int32` for datetime component methods by @Matt711 in https://github.com/NVIDIA/cudf/pull/21099 +* Update column names in pdsh query 18 by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21119 +* Preserve NA values in decimal ArrowExtensionArray comparisons by @Matt711 in https://github.com/NVIDIA/cudf/pull/21168 +* Handle empty format string in dt.strftime for GPU engine in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21143 +* Handle Greek final sigma in `str.lower()` by @Matt711 in https://github.com/NVIDIA/cudf/pull/21154 +* Preserve type metadata in unary ops by @Matt711 in https://github.com/NVIDIA/cudf/pull/21170 +* Fix the host-device access to device_span data in cumulative_centroid_weight_grouped by @vuule in https://github.com/NVIDIA/cudf/pull/21163 +* Make JNI partition output consistent with cuDF by @zpuller in https://github.com/NVIDIA/cudf/pull/21029 +* Ensure empty windows return NaN for rolling aggregations by @Matt711 in https://github.com/NVIDIA/cudf/pull/20677 +* Fix `at`/`iat` indexers to reject list-like keys by @Matt711 in https://github.com/NVIDIA/cudf/pull/21213 +* Vendor Pandas' to_xarray in cudf.pandas by @Matt711 in https://github.com/NVIDIA/cudf/pull/21175 +* Raise error for string slicing on non-monotonic DatetimeIndex by @Matt711 in https://github.com/NVIDIA/cudf/pull/21218 +* Add `__array_priority__` to proxy scalar types to match pandas by @Matt711 in https://github.com/NVIDIA/cudf/pull/21235 +* Change find_first_set_bit_kernel to use BlockReduce instead of cg::reduce by @davidwendt in https://github.com/NVIDIA/cudf/pull/21234 +* Add noinline declaration to secondary type-dispatching row-operators in Debug build by @davidwendt in https://github.com/NVIDIA/cudf/pull/21197 +* Fix casting to StringDtype for ArrowExtensionArray duration and decimal types by @Matt711 in https://github.com/NVIDIA/cudf/pull/21230 +* Fix `pl.Expr.concat_str` with boolean columns in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/21145 +* Only serialize column slice by @pentschev in https://github.com/NVIDIA/cudf/pull/21301 +* Fix return type of str.ip2int to uint32 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21324 +* Fix integer overflow in `cudf::hash_partition` by @wence- in https://github.com/NVIDIA/cudf/pull/21298 +* Remove pandas compatibility gating for dtypes in `datetime` and `timedelta` columns by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21252 +* Fix how cudf-polars tracing finds dataframe inputs by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20647 +* Use ColumnBase.create in from_arrow, follow-ups in pyarrow 19 decimal support by @mroeschke in https://github.com/NVIDIA/cudf/pull/21317 +* Fix validation for datetimes coming from arrow by @vyasr in https://github.com/NVIDIA/cudf/pull/21369 +* Fix n_workers detection in tpc-h benchmark by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21368 +* Fix list literal translation by @Matt711 in https://github.com/NVIDIA/cudf/pull/20801 +* Fix exponentially weighted mean with leading NULLs by @Matt711 in https://github.com/NVIDIA/cudf/pull/20556 +* Fix various issues in Python microbenchmarks by @vyasr in https://github.com/NVIDIA/cudf/pull/21340 +* Allow np.dtype(str) to coerce to object type for cudf column construction by @mroeschke in https://github.com/NVIDIA/cudf/pull/21403 +* Remove 4-byte padding requirement from the Brotli decompressor by @vuule in https://github.com/NVIDIA/cudf/pull/21392 +* Replace `thrust::get` with `cuda::std::get` by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21416 +* Tests all APIs docstrings in doctests by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21404 +* [BUG] Fixed a bug in the 10min notebook by @drivanov in https://github.com/NVIDIA/cudf/pull/21400 +* Fix empty string generation in get_json_object by @vyasr in https://github.com/NVIDIA/cudf/pull/21418 +* Save duckdb pdsh results to output file by @Matt711 in https://github.com/NVIDIA/cudf/pull/21391 +* Pin cupy<14 for wheel-tests-cudf and cudf-pandas unit tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/21469 +* Fix generic_scalar_device_view::value() to honor decimal scale by @davidwendt in https://github.com/NVIDIA/cudf/pull/21447 +* Add notes to failing polars test by @vyasr in https://github.com/NVIDIA/cudf/pull/21488 +* Support CuPy 14 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21504 +* Fix misaligned memory access in groupby by @PointKernel in https://github.com/NVIDIA/cudf/pull/21513 +* Deterministic build and probe table generation for join benchmarks by @shrshi in https://github.com/NVIDIA/cudf/pull/21410 +* Fix pds-h benchmark with `--suffix` by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21520 +* Fix OOB read in chunked Parquet reads with `skip_rows`/`num_rows` by @vuule in https://github.com/NVIDIA/cudf/pull/21407 +* rapidsmpf compatibility in unspill_partitions by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21539 +* Handle edge case dectection in pds-h validation by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21544 +* Fix scan/ewm logic to pass stream to column_device_view::create by @davidwendt in https://github.com/NVIDIA/cudf/pull/21557 +* Account for the rank index in streaming ``Sink`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/21555 +* Fix decimal precision loss in cast: float->decimal by @Matt711 in https://github.com/NVIDIA/cudf/pull/21450 +* Fix JIT filtering of join indices to properly handle nulls in nullable columns by @shrshi in https://github.com/NVIDIA/cudf/pull/21570 +* Fix span index-out-of-range error/assert in parquet::decode_page_data_generic kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/21575 +* Add stream parameter to indexalator make_input_optional_iterator by @davidwendt in https://github.com/NVIDIA/cudf/pull/21545 +* Adapt to latest rapidsmpf API changes by @wence- in https://github.com/NVIDIA/cudf/pull/21601 +* Usage of `template` keyword required for some nvcc and gcc combinations by @robertmaynard in https://github.com/NVIDIA/cudf/pull/21577 +* Fix caching the result of is_sorted by @jberg5 in https://github.com/NVIDIA/cudf/pull/21603 +* Small bugfix: json sink logic was being evaluated incorrectly by @jberg5 in https://github.com/NVIDIA/cudf/pull/21605 +* fix: correct typos in comments by @thecaptain789 in https://github.com/NVIDIA/cudf/pull/21380 +* Fix chunked_pack::next() returning overflowed int for >2GB data by @ttnghia in https://github.com/NVIDIA/cudf/pull/21643 +* Adjust precision in pds-h query 11 by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21651 +* Fix segfault in `HYBRID_SCAN_MULTITHREAD_NVBENCH` by @ttnghia in https://github.com/NVIDIA/cudf/pull/21661 +* Improve attribute lookup by 500x in `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21656 +* Fix `pl.Expr.split_exact` children result for null data by @mroeschke in https://github.com/NVIDIA/cudf/pull/21660 +* Increase timeout for the devcontainer build workflow by @davidwendt in https://github.com/NVIDIA/cudf/pull/21713 +* Support large partition counts in `hash_partition` by @ttnghia in https://github.com/NVIDIA/cudf/pull/21674 +* Use larger datatypes when calculating mean and other groupby-aggs by @quasiben in https://github.com/NVIDIA/cudf/pull/21719 +* Remove default stream use in PQ reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21629 +* Fix dynamic-join logic for minimum shuffle modulus by @rjzamora in https://github.com/NVIDIA/cudf/pull/21712 +* Support fixed point types in Parquet predicate pushdown by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21681 +* Fix inherited partitioning metadata in cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/21748 +* Fix validation failures in PDS-DS Q23 and Q61 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21754 +* Fix validation failures in Q69 and Q93 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21758 +* Fix thread pool worker threads using wrong CUDA device in multi-GPU environments by @vuule in https://github.com/NVIDIA/cudf/pull/21745 +* Fix quantile on decimals by @jberg5 in https://github.com/NVIDIA/cudf/pull/21707 +* Add missing return after state.skip() in libcudf benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/21756 +* [BUG] Fix parsing of ld.param PTX instruction in UDFs by @lamarrr in https://github.com/NVIDIA/cudf/pull/21597 +* Enable `astype(str)` in `cudf` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21699 +* Fix racecheck in join mark_retrieve_kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/21786 +* Fix approx distinct count nan handling by @PointKernel in https://github.com/NVIDIA/cudf/pull/21806 +* Fix mypy pinning on rmm by @bdice in https://github.com/NVIDIA/cudf/pull/21925 +* Fix JIT transform kernel crash on non-default CUDA streams by @bdice in https://github.com/NVIDIA/cudf/pull/21921 ### 📖 Documentation -* ``cudf`` -> ``pd`` typo in ``README`` by @jrbourbeau in https://github.com/rapidsai/cudf/pull/21134 -* Fix docs for cudf::index_of_first_set_bit API by @davidwendt in https://github.com/rapidsai/cudf/pull/21313 -* [DOC] Document Java ColumnView reductions [skip ci] by @rishic3 in https://github.com/rapidsai/cudf/pull/21205 -* Remove outdated issue links by @rishic3 in https://github.com/rapidsai/cudf/pull/21259 -* Move Parquet IO utils to a new doxygen group by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21412 -* Remove link to deprecated JIT-unspilling documentation by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21710 +* ``cudf`` -> ``pd`` typo in ``README`` by @jrbourbeau in https://github.com/NVIDIA/cudf/pull/21134 +* Fix docs for cudf::index_of_first_set_bit API by @davidwendt in https://github.com/NVIDIA/cudf/pull/21313 +* [DOC] Document Java ColumnView reductions [skip ci] by @rishic3 in https://github.com/NVIDIA/cudf/pull/21205 +* Remove outdated issue links by @rishic3 in https://github.com/NVIDIA/cudf/pull/21259 +* Move Parquet IO utils to a new doxygen group by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21412 +* Remove link to deprecated JIT-unspilling documentation by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21710 ### 🚀 New Features -* Add remaining pandas PDSH queries by @Matt711 in https://github.com/rapidsai/cudf/pull/21108 -* Add the fast type to the `NamedAgg` wrapper in cudf.pandas by @Matt711 in https://github.com/rapidsai/cudf/pull/21135 -* ENH: Expose split_part to Python API via pylibcudf by @Umang-projects in https://github.com/rapidsai/cudf/pull/21068 -* Experimental deletion vector parquet APIs to accept information via const ref to a struct by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21044 -* Add cudf::index_of_first_set_bit API by @davidwendt in https://github.com/rapidsai/cudf/pull/21207 -* Expose `cudf::filter` to python by @Matt711 in https://github.com/rapidsai/cudf/pull/21300 -* Add a Java wrapper for deletion vector APIs by @jihoonson in https://github.com/rapidsai/cudf/pull/21231 -* Add ``allgather_reduce`` utility for the rapidsmpf runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/21339 -* Add Hybrid scan APIs to chunked materialize all columns by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21377 -* Util to get packed table buffer size by @nirandaperera in https://github.com/rapidsai/cudf/pull/21322 -* Add PDS-DS Q11 by @Matt711 in https://github.com/rapidsai/cudf/pull/19658 -* Move hybrid scan IO utils to libcudf by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21360 -* Expose column_nans_to_nulls API to Python and update cudf to use it by @vyasr in https://github.com/rapidsai/cudf/pull/21438 -* Add PDS-DS Q12-16 by @Matt711 in https://github.com/rapidsai/cudf/pull/20389 -* [FEA] Enable Pre-compiled Headers for faster JIT by @lamarrr in https://github.com/rapidsai/cudf/pull/21457 -* Support named capture groups in cudf extract by @davidwendt in https://github.com/rapidsai/cudf/pull/21484 -* Use cuda::std::optional for parquet LogicalType to enable device compatibility by @PointKernel in https://github.com/rapidsai/cudf/pull/21470 -* [FEA] update cudf::transform and cudf::filter callsites by @lamarrr in https://github.com/rapidsai/cudf/pull/21537 -* Example to demonstrate inter-parquet-file pipelining using hybrid scan APIs by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20722 -* Add PDS-DS Q30-36 by @Matt711 in https://github.com/rapidsai/cudf/pull/20393 -* [FEA] Fix deprecation warnings in cudf::transform and cudf::filter by @lamarrr in https://github.com/rapidsai/cudf/pull/21551 -* Retrieve cudf::data_type schema from Parquet metadata by @Matt711 in https://github.com/rapidsai/cudf/pull/21458 -* Add divmod support, respect array priority in ufuncs, and properly handle negative infs and zeros by @vyasr in https://github.com/rapidsai/cudf/pull/21523 -* Add PDS-DS Q63-72 by @Matt711 in https://github.com/rapidsai/cudf/pull/20400 -* Dynamic Distinct and GroupBy support by @rjzamora in https://github.com/rapidsai/cudf/pull/21433 -* Support ArrowStringView from device in cudf::from_arrow_device by @davidwendt in https://github.com/rapidsai/cudf/pull/21385 -* Add PDS-DS Q24-29 by @Matt711 in https://github.com/rapidsai/cudf/pull/20392 -* Add set_decimal_width option to parquet reader by @Matt711 in https://github.com/rapidsai/cudf/pull/21472 -* Add PDS-DS Q17-23 by @Matt711 in https://github.com/rapidsai/cudf/pull/20390 -* Add PDS-DS Q73-85 by @Matt711 in https://github.com/rapidsai/cudf/pull/20402 -* Add PDS-DS Q86-99 by @Matt711 in https://github.com/rapidsai/cudf/pull/20404 -* Add PDS-DS Q37-43 by @Matt711 in https://github.com/rapidsai/cudf/pull/20394 -* [FEA] Introduce `udf_source_type` enum to distinguish between CUDA and PTX UDFs. by @lamarrr in https://github.com/rapidsai/cudf/pull/21567 -* Add PDS-DS Q53-62 by @Matt711 in https://github.com/rapidsai/cudf/pull/20399 -* Add PDS-DS Q44-52 by @Matt711 in https://github.com/rapidsai/cudf/pull/20395 -* Benchmark parquet footer read and column select time by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21422 -* Return file footer buffer at Parquet writer close by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21378 -* Add dynamic join actor by @rjzamora in https://github.com/rapidsai/cudf/pull/21578 -* Add --io-mode for page cache control (cold/lukewarm/hot) by @Matt711 in https://github.com/rapidsai/cudf/pull/21653 -* Expand and improve AST expression support in Parquet readers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21515 -* Support count aggregation (valid/all) for cudf::scan by @davidwendt in https://github.com/rapidsai/cudf/pull/21565 -* [FEA] Add JIT cache management functions by @lamarrr in https://github.com/rapidsai/cudf/pull/21553 -* Add mark_join to support semi/anti join with left table reuse by @shrshi in https://github.com/rapidsai/cudf/pull/21440 -* Restore multithreaded optimization in the CSV reader by @vuule in https://github.com/rapidsai/cudf/pull/21307 +* Add remaining pandas PDSH queries by @Matt711 in https://github.com/NVIDIA/cudf/pull/21108 +* Add the fast type to the `NamedAgg` wrapper in cudf.pandas by @Matt711 in https://github.com/NVIDIA/cudf/pull/21135 +* ENH: Expose split_part to Python API via pylibcudf by @Umang-projects in https://github.com/NVIDIA/cudf/pull/21068 +* Experimental deletion vector parquet APIs to accept information via const ref to a struct by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21044 +* Add cudf::index_of_first_set_bit API by @davidwendt in https://github.com/NVIDIA/cudf/pull/21207 +* Expose `cudf::filter` to python by @Matt711 in https://github.com/NVIDIA/cudf/pull/21300 +* Add a Java wrapper for deletion vector APIs by @jihoonson in https://github.com/NVIDIA/cudf/pull/21231 +* Add ``allgather_reduce`` utility for the rapidsmpf runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/21339 +* Add Hybrid scan APIs to chunked materialize all columns by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21377 +* Util to get packed table buffer size by @nirandaperera in https://github.com/NVIDIA/cudf/pull/21322 +* Add PDS-DS Q11 by @Matt711 in https://github.com/NVIDIA/cudf/pull/19658 +* Move hybrid scan IO utils to libcudf by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21360 +* Expose column_nans_to_nulls API to Python and update cudf to use it by @vyasr in https://github.com/NVIDIA/cudf/pull/21438 +* Add PDS-DS Q12-16 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20389 +* [FEA] Enable Pre-compiled Headers for faster JIT by @lamarrr in https://github.com/NVIDIA/cudf/pull/21457 +* Support named capture groups in cudf extract by @davidwendt in https://github.com/NVIDIA/cudf/pull/21484 +* Use cuda::std::optional for parquet LogicalType to enable device compatibility by @PointKernel in https://github.com/NVIDIA/cudf/pull/21470 +* [FEA] update cudf::transform and cudf::filter callsites by @lamarrr in https://github.com/NVIDIA/cudf/pull/21537 +* Example to demonstrate inter-parquet-file pipelining using hybrid scan APIs by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20722 +* Add PDS-DS Q30-36 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20393 +* [FEA] Fix deprecation warnings in cudf::transform and cudf::filter by @lamarrr in https://github.com/NVIDIA/cudf/pull/21551 +* Retrieve cudf::data_type schema from Parquet metadata by @Matt711 in https://github.com/NVIDIA/cudf/pull/21458 +* Add divmod support, respect array priority in ufuncs, and properly handle negative infs and zeros by @vyasr in https://github.com/NVIDIA/cudf/pull/21523 +* Add PDS-DS Q63-72 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20400 +* Dynamic Distinct and GroupBy support by @rjzamora in https://github.com/NVIDIA/cudf/pull/21433 +* Support ArrowStringView from device in cudf::from_arrow_device by @davidwendt in https://github.com/NVIDIA/cudf/pull/21385 +* Add PDS-DS Q24-29 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20392 +* Add set_decimal_width option to parquet reader by @Matt711 in https://github.com/NVIDIA/cudf/pull/21472 +* Add PDS-DS Q17-23 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20390 +* Add PDS-DS Q73-85 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20402 +* Add PDS-DS Q86-99 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20404 +* Add PDS-DS Q37-43 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20394 +* [FEA] Introduce `udf_source_type` enum to distinguish between CUDA and PTX UDFs. by @lamarrr in https://github.com/NVIDIA/cudf/pull/21567 +* Add PDS-DS Q53-62 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20399 +* Add PDS-DS Q44-52 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20395 +* Benchmark parquet footer read and column select time by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21422 +* Return file footer buffer at Parquet writer close by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21378 +* Add dynamic join actor by @rjzamora in https://github.com/NVIDIA/cudf/pull/21578 +* Add --io-mode for page cache control (cold/lukewarm/hot) by @Matt711 in https://github.com/NVIDIA/cudf/pull/21653 +* Expand and improve AST expression support in Parquet readers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21515 +* Support count aggregation (valid/all) for cudf::scan by @davidwendt in https://github.com/NVIDIA/cudf/pull/21565 +* [FEA] Add JIT cache management functions by @lamarrr in https://github.com/NVIDIA/cudf/pull/21553 +* Add mark_join to support semi/anti join with left table reuse by @shrshi in https://github.com/NVIDIA/cudf/pull/21440 +* Restore multithreaded optimization in the CSV reader by @vuule in https://github.com/NVIDIA/cudf/pull/21307 ### 🛠️ Improvements -* Add no_racecheck filter to snap, unsnap, and inflate kernel names by @davidwendt in https://github.com/rapidsai/cudf/pull/21043 -* Main merge release/26.02 by @mmccarty in https://github.com/rapidsai/cudf/pull/21071 -* Disable unneeded checking in QUANTILES_TEST during racecheck runs by @davidwendt in https://github.com/rapidsai/cudf/pull/21011 -* Optionally write results from pdsh benchmarks by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21052 -* Main merge release/26.02 by @vyasr in https://github.com/rapidsai/cudf/pull/21073 -* Use CRTP (Curiously Recurring Template Pattern) to implement `clone()` method for aggregation classes by @ttnghia in https://github.com/rapidsai/cudf/pull/21066 -* Add a hybrid scan API to build an all-true row mask by @aryansri05 in https://github.com/rapidsai/cudf/pull/20862 -* [polars-rapidsmpf] Shutdown RapidsMPF's Context by @madsbk in https://github.com/rapidsai/cudf/pull/21083 -* Allow for scheduler file and existing dask cluster when using pdsh by @quasiben in https://github.com/rapidsai/cudf/pull/21024 -* Remove children from ColumnBase by @vyasr in https://github.com/rapidsai/cudf/pull/21077 -* Add new pinned vector factory functions by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21033 -* Add std::span operator to cudf span classes by @davidwendt in https://github.com/rapidsai/cudf/pull/21053 -* [polars-rapidsmpf] enable spill to pinned memory by @madsbk in https://github.com/rapidsai/cudf/pull/21067 -* Expand ``Metadata`` for better multi-GPU planning information by @rjzamora in https://github.com/rapidsai/cudf/pull/20908 -* Removal IntervalDtype/StructDtype inheritance by @vyasr in https://github.com/rapidsai/cudf/pull/21114 -* install 'binutils' in conda Python test env by @jameslamb in https://github.com/rapidsai/cudf/pull/21165 -* Add missing stream synchronizations to various tests by @vyasr in https://github.com/rapidsai/cudf/pull/21122 -* Skip __setitem__ PeriodDtype tests by @Matt711 in https://github.com/rapidsai/cudf/pull/21164 -* Refactor groupby aggregation, removing `simple_aggregations_collector` and `aggregation_finalizer` classes by @ttnghia in https://github.com/rapidsai/cudf/pull/21064 -* Use verify-hardcoded-version pre-commit hook by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/21152 -* Use cub FlaggedIf for copy-if with stencil parameter by @davidwendt in https://github.com/rapidsai/cudf/pull/21156 -* Fix `pl.Expr.fill_null` in cudf_polars with strategy "zero"/"one" for boolean columns by @mroeschke in https://github.com/rapidsai/cudf/pull/21146 -* Use `Span` thoughout pylibcudf by @Matt711 in https://github.com/rapidsai/cudf/pull/21087 -* Ignore hardcoded versions in deprecated code by @bdice in https://github.com/rapidsai/cudf/pull/21192 -* Replace thrust::count_if with calls to cudf::detail::count_if by @davidwendt in https://github.com/rapidsai/cudf/pull/21162 -* Add cudf::detail::reduce_by_key_async utility by @davidwendt in https://github.com/rapidsai/cudf/pull/21184 -* Fix for DeviceMin and DeviceMax default device operators by @davidwendt in https://github.com/rapidsai/cudf/pull/21118 -* Forward merge release/26.02 into main by @Matt711 in https://github.com/rapidsai/cudf/pull/21191 -* Move pylibcudf tests to correct directory by @vyasr in https://github.com/rapidsai/cudf/pull/21188 -* Use the same channel for TableChunk and Metadata messages by @rjzamora in https://github.com/rapidsai/cudf/pull/21182 -* Add new ColumnBase.create API by @vyasr in https://github.com/rapidsai/cudf/pull/21187 -* Use ColumnBase.create more by @vyasr in https://github.com/rapidsai/cudf/pull/21216 -* tighten wheel size limits, expand CI-skipping logic, other small build changes by @jameslamb in https://github.com/rapidsai/cudf/pull/21220 -* Remove numba-cuda upper bound by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/21117 -* Drop Python 3.10 support by @gforsyth in https://github.com/rapidsai/cudf/pull/21225 -* Added an extra-info field to pds-h output by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21219 -* Allow IR nodes to use multiple collective IDs by @rjzamora in https://github.com/rapidsai/cudf/pull/21203 -* Remove public scan/reduce ColumnBase APIs by @vyasr in https://github.com/rapidsai/cudf/pull/21217 -* Remove pandas compatibility gating in `cudf.factorize` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21238 -* Remove pandas compatibility gating for dtypes in `join` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21239 -* Remove pandas compatibility gating in frame and groupby methods by @galipremsagar in https://github.com/rapidsai/cudf/pull/21254 -* Add ``DynamicPlanningOptions`` configuration by @rjzamora in https://github.com/rapidsai/cudf/pull/21250 -* Remove pandas compatibility gating of dtypes in `Index` methods by @galipremsagar in https://github.com/rapidsai/cudf/pull/21255 -* Update verify-copyright configuration by @bdice in https://github.com/rapidsai/cudf/pull/21196 -* Remove pandas compatibility gating for dtypes in timdelta column by @galipremsagar in https://github.com/rapidsai/cudf/pull/21258 -* Move rapids-logger statement inside if in test_cpp.sh by @davidwendt in https://github.com/rapidsai/cudf/pull/21266 -* Reduces CategoricalColumn reliance on codes and improve property cache clearing by @vyasr in https://github.com/rapidsai/cudf/pull/21264 -* Remove pandas compatibility gating for dtypes in `struct`, `list` and `interval` columns by @galipremsagar in https://github.com/rapidsai/cudf/pull/21240 -* Use ColumnBase.create more in StringColumn by @mroeschke in https://github.com/rapidsai/cudf/pull/21265 -* Replace thrust::copy_if calls with cudf::detail::copy_if by @davidwendt in https://github.com/rapidsai/cudf/pull/21268 -* Update rolling logic to allow unsorted dictionary keys by @davidwendt in https://github.com/rapidsai/cudf/pull/21037 -* Remove deprecated pylibcudf arrow APIs by @Matt711 in https://github.com/rapidsai/cudf/pull/21284 -* Don't gate ColumnBase.to_pandas behind mode.pandas_compatible by @mroeschke in https://github.com/rapidsai/cudf/pull/21256 -* Remove pandas compatibility gating for dtypes in DataFrame APIs by @galipremsagar in https://github.com/rapidsai/cudf/pull/21261 -* Remove pandas compatibility gating for dtypes in decimal column by @galipremsagar in https://github.com/rapidsai/cudf/pull/21243 -* Remove pandas compatibility gating for dtypes in rolling window by @galipremsagar in https://github.com/rapidsai/cudf/pull/21260 -* remove pip.conf migration code in CI scripts, update CI-skipping rules by @jameslamb in https://github.com/rapidsai/cudf/pull/21303 -* Don't hard-code branch names by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/21270 -* Change CUB If and FlaggedIf to use cudf::detail::copy_if by @davidwendt in https://github.com/rapidsai/cudf/pull/21274 -* Migrate to rapidsmpf ``ChannelMetadata`` by @rjzamora in https://github.com/rapidsai/cudf/pull/21275 -* Remove deprecation warnings from benchmarks/ndsh/utilities.cpp by @davidwendt in https://github.com/rapidsai/cudf/pull/21308 -* Improve pylibcudf usage in cudf by @vyasr in https://github.com/rapidsai/cudf/pull/21285 -* Continue refactoring column construction by @vyasr in https://github.com/rapidsai/cudf/pull/21262 -* Is already partitioned by @rjzamora in https://github.com/rapidsai/cudf/pull/21310 -* Consolidate cudf-classic dtype validators to their own file and make them stricter by @mroeschke in https://github.com/rapidsai/cudf/pull/21281 -* Preprocess parquet repetition and definition levels by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/21139 -* Remove pandas compatibility dtypes gating in series methods by @galipremsagar in https://github.com/rapidsai/cudf/pull/21257 -* Implement `na` for `match` and `contains` in `str` API by @galipremsagar in https://github.com/rapidsai/cudf/pull/21236 -* Remove redundant lowering and dispatching infrastructure for the rapidsmpf runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/21337 -* Full structlog-based tracing by @rjzamora in https://github.com/rapidsai/cudf/pull/21316 -* Remove some unneeded libcudf debug-specific code paths by @davidwendt in https://github.com/rapidsai/cudf/pull/21335 -* Move has_nulls template parameter to runtime in rolling window by @davidwendt in https://github.com/rapidsai/cudf/pull/21312 -* Enable KvikIO datasource in hybrid scan examples by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21318 -* Reserve more collective IDs for "dynamic" IR nodes by @rjzamora in https://github.com/rapidsai/cudf/pull/21343 -* Deprecate `values_host` in favor of `to_numpy` by @Matt711 in https://github.com/rapidsai/cudf/pull/21346 -* refactor: build wheels and conda packages using Python limited API by @gforsyth in https://github.com/rapidsai/cudf/pull/21329 -* Ensure ColumnBase.create asserts that plc_column type is equivalent to passed dtype by @mroeschke in https://github.com/rapidsai/cudf/pull/21325 -* Move generate_nested_types.hpp definitions to .cpp file by @davidwendt in https://github.com/rapidsai/cudf/pull/21344 -* Remove dead code checking scheduler in fallback logic by @Matt711 in https://github.com/rapidsai/cudf/pull/21351 -* Skip failing cuDF Python tests on RTX PRO 6000 by @bdice in https://github.com/rapidsai/cudf/pull/21358 -* Remove mode.pandas_compatible from StringColumn.all_characters_of_type by @mroeschke in https://github.com/rapidsai/cudf/pull/21356 -* Remove export from cudf::detail::sort and cudf::detail::merge APIs by @davidwendt in https://github.com/rapidsai/cudf/pull/21345 -* Cleanup duplicated code in hybrid scan by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21353 -* Make custreamz and dask-cudf independent of cuda-version by @vyasr in https://github.com/rapidsai/cudf/pull/21348 -* Use ColumnBase.create more in cudf classic APIs by @mroeschke in https://github.com/rapidsai/cudf/pull/21306 -* Avoid single-partition assumptions when dynamic-planning is enabled by @rjzamora in https://github.com/rapidsai/cudf/pull/21370 -* Add a fallback-mode cli arg to the pdsh benchmarks by @Matt711 in https://github.com/rapidsai/cudf/pull/21365 -* Add benchmarks for dictionary set_keys and sort by @davidwendt in https://github.com/rapidsai/cudf/pull/21366 -* fix(deps): bump `pyarrow` floor in conda recipes by @gforsyth in https://github.com/rapidsai/cudf/pull/21383 -* More create usage and tightening up dtype specification by @vyasr in https://github.com/rapidsai/cudf/pull/21359 -* Optimize cudf usage of pylibcudf by @vyasr in https://github.com/rapidsai/cudf/pull/21362 -* Include a run_id in pds-h benchmark run by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21375 -* Capture query plan in pds-h output by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21315 -* Remove cudf compat gating at more places by @galipremsagar in https://github.com/rapidsai/cudf/pull/21350 -* Use ColumnBase.create more in frame.py by @mroeschke in https://github.com/rapidsai/cudf/pull/21394 -* Remove `._with_dtype_metadata` in cudf classic groupby by @mroeschke in https://github.com/rapidsai/cudf/pull/21389 -* Allocate single compressed buffer per source in parquet reader by @vuule in https://github.com/rapidsai/cudf/pull/21323 -* Preserve ArrowDtype in `Series.struct.fields` by @mroeschke in https://github.com/rapidsai/cudf/pull/21379 -* Use ColumnBase.create instead of from_pylibcudf in indexed_frame.py by @mroeschke in https://github.com/rapidsai/cudf/pull/21401 -* Use GHA id-token for `sccache-dist` auth token by @trxcllnt in https://github.com/rapidsai/cudf/pull/21406 -* Remove more mode.pandas_compatible from StringColumn by @mroeschke in https://github.com/rapidsai/cudf/pull/21393 -* Rewrite CategoricalColumn operations in terms of pylibcudf and avoid `_with_type_metadata` by @vyasr in https://github.com/rapidsai/cudf/pull/21402 -* Remove from_pylibcudf/_with_dtype_metadata in IO readers/rolling by @mroeschke in https://github.com/rapidsai/cudf/pull/21409 -* Remove Frame._copy_type_metadata method by @vyasr in https://github.com/rapidsai/cudf/pull/21405 -* Apply PIMPL idiom to sort-merge join by @shrshi in https://github.com/rapidsai/cudf/pull/21349 -* Avoid `_with_dtype_metadata` in dask_cudf by @mroeschke in https://github.com/rapidsai/cudf/pull/21411 -* Remove hash_vocab_utils by @vyasr in https://github.com/rapidsai/cudf/pull/21419 -* Refactor find_and_replace in Numerical and String columns by @vyasr in https://github.com/rapidsai/cudf/pull/21413 -* Use ColumnBase.create over .from_pylibcudf in MultiIndex, groupby, misc. by @mroeschke in https://github.com/rapidsai/cudf/pull/21423 -* Skip RTX PRO 6000 tests in test workflow by @vyasr in https://github.com/rapidsai/cudf/pull/21431 -* Split up algorithm.cuh into reduce.cuh and copy_if.cuh by @davidwendt in https://github.com/rapidsai/cudf/pull/21387 -* Put typename disambiguator before nested name specifier by @seanbaxter in https://github.com/rapidsai/cudf/pull/21408 -* Use ColumnBase.create over .from_pylibcudf in dataframe/index/indexed_frame.py by @mroeschke in https://github.com/rapidsai/cudf/pull/21439 -* Remove mode.pandas_compatable from ColumnBase.cast by @mroeschke in https://github.com/rapidsai/cudf/pull/21441 -* Remove mode.pandas_compatible from ColumnBase._validate_args by @mroeschke in https://github.com/rapidsai/cudf/pull/21426 -* Make rapidsmpf py_executor thread pool size configurable by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21446 -* Replace `thrust::constant_iterator` with `cuda::constant_iterator` by @miscco in https://github.com/rapidsai/cudf/pull/21445 -* Remove _with_type_metadata by @vyasr in https://github.com/rapidsai/cudf/pull/21421 -* Drop `thrust::reverse_iterator` in favor of `cuda::std::reverse_iterator` by @miscco in https://github.com/rapidsai/cudf/pull/21455 -* Move detail::sorted_order decl from sort_column_impl.cuh to sort.hpp by @davidwendt in https://github.com/rapidsai/cudf/pull/21436 -* Remove remaining uses of ColumnBase.from_pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/21448 -* Add public bitmask_and with host_span parameters by @davidwendt in https://github.com/rapidsai/cudf/pull/21437 -* RapidsMPF renamed `Node` to `Actor` by @madsbk in https://github.com/rapidsai/cudf/pull/21476 -* Reduce nvbench parameter space for high-multiplicity and sort-merge join benchmarks by @shrshi in https://github.com/rapidsai/cudf/pull/21076 -* Reduce PR workflow jobs for cpp-only changes by @davidwendt in https://github.com/rapidsai/cudf/pull/21381 -* Optimize PDSH queries for `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/21462 -* Pin polars >=1.30 and <1.39 by @Matt711 in https://github.com/rapidsai/cudf/pull/21304 -* Use param substitutions from dsqgen by @Matt711 in https://github.com/rapidsai/cudf/pull/21465 -* Make cudf-polars config generic over its executor by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21459 -* Allow running PDS-DS with qualification parameters by @Matt711 in https://github.com/rapidsai/cudf/pull/21478 -* Use new local_partitions method of shuffler by @wence- in https://github.com/rapidsai/cudf/pull/21461 -* Unify normalization, buffer wrapping, and validation by @vyasr in https://github.com/rapidsai/cudf/pull/21451 -* Update/remove version checks for outdated versions in Python libraries by @mroeschke in https://github.com/rapidsai/cudf/pull/21483 -* check-nightly-ci: update to new version by @jameslamb in https://github.com/rapidsai/cudf/pull/21464 -* Remove pyarrow string_view workarounds in pylibcudf/cudf by @mroeschke in https://github.com/rapidsai/cudf/pull/21490 -* Remove dead decimal alignment code by @Matt711 in https://github.com/rapidsai/cudf/pull/21474 -* Move Column normalization to only required code paths by @vyasr in https://github.com/rapidsai/cudf/pull/21487 -* Rename columns to column_names in parquet reader options by @Matt711 in https://github.com/rapidsai/cudf/pull/21473 -* Fix more deprecated calls in JNI source by @davidwendt in https://github.com/rapidsai/cudf/pull/21482 -* Remove inefficiencies in gather by @galipremsagar in https://github.com/rapidsai/cudf/pull/21429 -* pds-h validation improvements by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21388 -* check-nightly-ci: remove testing config by @jameslamb in https://github.com/rapidsai/cudf/pull/21494 -* Update RapidsMPF's `Context()` since it no longer takes a statistics argument. by @madsbk in https://github.com/rapidsai/cudf/pull/21503 -* Drop uses of thrust/functional by @miscco in https://github.com/rapidsai/cudf/pull/21477 -* Refactor some common utilities out of find_and_replace by @vyasr in https://github.com/rapidsai/cudf/pull/21500 -* Sort specialization logic for dictionary column by @davidwendt in https://github.com/rapidsai/cudf/pull/21420 -* Add cudf::detail::copy_if_async utility by @davidwendt in https://github.com/rapidsai/cudf/pull/21495 -* Update range values for gather/scatter benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/21479 -* Replace some cudf::detail calls with public API equivalents by @davidwendt in https://github.com/rapidsai/cudf/pull/21507 -* Reenable pandas tests disabled due to prior CoW incompatibilities by @vyasr in https://github.com/rapidsai/cudf/pull/21510 -* Add utility to extract partitioning moduli from the ChannelMetadata by @rjzamora in https://github.com/rapidsai/cudf/pull/21491 -* Refresh cudf.pandas xpass list by @vyasr in https://github.com/rapidsai/cudf/pull/21521 -* List and struct support for make_empty_column by @quasiben in https://github.com/rapidsai/cudf/pull/21493 -* Create ColumnBase.from_range to reduce as_column usage by @mroeschke in https://github.com/rapidsai/cudf/pull/21516 -* Add properties in serialized `explain` for Scan and Shuffle by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21497 -* Implement sink nodes for rapidsmpf runtime by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20965 -* Reduce CI by skipping tests unaffected by changes based on package dependency chain by @Matt711 in https://github.com/rapidsai/cudf/pull/21506 -* Remove unneeded specialized device_vector/device_span gtest by @davidwendt in https://github.com/rapidsai/cudf/pull/21508 -* Rework reductions/scan_tests.cpp to help with adding aggregation types by @davidwendt in https://github.com/rapidsai/cudf/pull/21511 -* Fix pdsh benchmark help text by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21524 -* Update casts in pds-h for --validate-directory by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21534 -* Change thrust::make_discard_iterator to cuda::make_discard_iterator by @davidwendt in https://github.com/rapidsai/cudf/pull/21526 -* Avoid cudf.dtype dtype parsing internally by @mroeschke in https://github.com/rapidsai/cudf/pull/21522 -* Use ``--rmm-release-threshold`` for both single- and multi-gpu benchmarks by @rjzamora in https://github.com/rapidsai/cudf/pull/21535 -* Test cudf-polars with the "rapidsmpf" runtime and "distributed" cluster by @rjzamora in https://github.com/rapidsai/cudf/pull/21028 -* chore(deps): bump numba ceiling to <0.65.0 by @gforsyth in https://github.com/rapidsai/cudf/pull/21543 -* Support JIT filtering of join output indices by @shrshi in https://github.com/rapidsai/cudf/pull/21202 -* Use stricter dtype object checks in cudf classic by @mroeschke in https://github.com/rapidsai/cudf/pull/21527 -* Use new modules for RMM CudaStreamFlags by @bdice in https://github.com/rapidsai/cudf/pull/21559 -* chore(deps): bump numba upper bound in conda recipe by @gforsyth in https://github.com/rapidsai/cudf/pull/21561 -* Validate PDS-DS Queries 1-16 by @Matt711 in https://github.com/rapidsai/cudf/pull/21556 -* Update conftest-patch for minor numpy changes by @vyasr in https://github.com/rapidsai/cudf/pull/21568 -* Create wrapper for pylibcudf function calls by @vyasr in https://github.com/rapidsai/cudf/pull/21501 -* Set track-stream-ordered-races=no in compute-sanitizer racecheck for LARGE_STRINGS_TEST by @davidwendt in https://github.com/rapidsai/cudf/pull/21562 -* Skip scan decimal introspection logic when there are no decimals in the schema by @Matt711 in https://github.com/rapidsai/cudf/pull/21571 -* Add per-file, unprivileged page cache dropping to cuDF benchmarks by @kingcrimsontianyu in https://github.com/rapidsai/cudf/pull/21397 -* Use non-blocking streams in libcudf's internal stream pool by @vuule in https://github.com/rapidsai/cudf/pull/21449 -* Make cudf-polars pds-h script more robust by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21538 -* Remove `Index` materialization in all query results by @galipremsagar in https://github.com/rapidsai/cudf/pull/21542 +* Add no_racecheck filter to snap, unsnap, and inflate kernel names by @davidwendt in https://github.com/NVIDIA/cudf/pull/21043 +* Main merge release/26.02 by @mmccarty in https://github.com/NVIDIA/cudf/pull/21071 +* Disable unneeded checking in QUANTILES_TEST during racecheck runs by @davidwendt in https://github.com/NVIDIA/cudf/pull/21011 +* Optionally write results from pdsh benchmarks by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21052 +* Main merge release/26.02 by @vyasr in https://github.com/NVIDIA/cudf/pull/21073 +* Use CRTP (Curiously Recurring Template Pattern) to implement `clone()` method for aggregation classes by @ttnghia in https://github.com/NVIDIA/cudf/pull/21066 +* Add a hybrid scan API to build an all-true row mask by @aryansri05 in https://github.com/NVIDIA/cudf/pull/20862 +* [polars-rapidsmpf] Shutdown RapidsMPF's Context by @madsbk in https://github.com/NVIDIA/cudf/pull/21083 +* Allow for scheduler file and existing dask cluster when using pdsh by @quasiben in https://github.com/NVIDIA/cudf/pull/21024 +* Remove children from ColumnBase by @vyasr in https://github.com/NVIDIA/cudf/pull/21077 +* Add new pinned vector factory functions by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21033 +* Add std::span operator to cudf span classes by @davidwendt in https://github.com/NVIDIA/cudf/pull/21053 +* [polars-rapidsmpf] enable spill to pinned memory by @madsbk in https://github.com/NVIDIA/cudf/pull/21067 +* Expand ``Metadata`` for better multi-GPU planning information by @rjzamora in https://github.com/NVIDIA/cudf/pull/20908 +* Removal IntervalDtype/StructDtype inheritance by @vyasr in https://github.com/NVIDIA/cudf/pull/21114 +* install 'binutils' in conda Python test env by @jameslamb in https://github.com/NVIDIA/cudf/pull/21165 +* Add missing stream synchronizations to various tests by @vyasr in https://github.com/NVIDIA/cudf/pull/21122 +* Skip __setitem__ PeriodDtype tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/21164 +* Refactor groupby aggregation, removing `simple_aggregations_collector` and `aggregation_finalizer` classes by @ttnghia in https://github.com/NVIDIA/cudf/pull/21064 +* Use verify-hardcoded-version pre-commit hook by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/21152 +* Use cub FlaggedIf for copy-if with stencil parameter by @davidwendt in https://github.com/NVIDIA/cudf/pull/21156 +* Fix `pl.Expr.fill_null` in cudf_polars with strategy "zero"/"one" for boolean columns by @mroeschke in https://github.com/NVIDIA/cudf/pull/21146 +* Use `Span` thoughout pylibcudf by @Matt711 in https://github.com/NVIDIA/cudf/pull/21087 +* Ignore hardcoded versions in deprecated code by @bdice in https://github.com/NVIDIA/cudf/pull/21192 +* Replace thrust::count_if with calls to cudf::detail::count_if by @davidwendt in https://github.com/NVIDIA/cudf/pull/21162 +* Add cudf::detail::reduce_by_key_async utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/21184 +* Fix for DeviceMin and DeviceMax default device operators by @davidwendt in https://github.com/NVIDIA/cudf/pull/21118 +* Forward merge release/26.02 into main by @Matt711 in https://github.com/NVIDIA/cudf/pull/21191 +* Move pylibcudf tests to correct directory by @vyasr in https://github.com/NVIDIA/cudf/pull/21188 +* Use the same channel for TableChunk and Metadata messages by @rjzamora in https://github.com/NVIDIA/cudf/pull/21182 +* Add new ColumnBase.create API by @vyasr in https://github.com/NVIDIA/cudf/pull/21187 +* Use ColumnBase.create more by @vyasr in https://github.com/NVIDIA/cudf/pull/21216 +* tighten wheel size limits, expand CI-skipping logic, other small build changes by @jameslamb in https://github.com/NVIDIA/cudf/pull/21220 +* Remove numba-cuda upper bound by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/21117 +* Drop Python 3.10 support by @gforsyth in https://github.com/NVIDIA/cudf/pull/21225 +* Added an extra-info field to pds-h output by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21219 +* Allow IR nodes to use multiple collective IDs by @rjzamora in https://github.com/NVIDIA/cudf/pull/21203 +* Remove public scan/reduce ColumnBase APIs by @vyasr in https://github.com/NVIDIA/cudf/pull/21217 +* Remove pandas compatibility gating in `cudf.factorize` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21238 +* Remove pandas compatibility gating for dtypes in `join` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21239 +* Remove pandas compatibility gating in frame and groupby methods by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21254 +* Add ``DynamicPlanningOptions`` configuration by @rjzamora in https://github.com/NVIDIA/cudf/pull/21250 +* Remove pandas compatibility gating of dtypes in `Index` methods by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21255 +* Update verify-copyright configuration by @bdice in https://github.com/NVIDIA/cudf/pull/21196 +* Remove pandas compatibility gating for dtypes in timdelta column by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21258 +* Move rapids-logger statement inside if in test_cpp.sh by @davidwendt in https://github.com/NVIDIA/cudf/pull/21266 +* Reduces CategoricalColumn reliance on codes and improve property cache clearing by @vyasr in https://github.com/NVIDIA/cudf/pull/21264 +* Remove pandas compatibility gating for dtypes in `struct`, `list` and `interval` columns by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21240 +* Use ColumnBase.create more in StringColumn by @mroeschke in https://github.com/NVIDIA/cudf/pull/21265 +* Replace thrust::copy_if calls with cudf::detail::copy_if by @davidwendt in https://github.com/NVIDIA/cudf/pull/21268 +* Update rolling logic to allow unsorted dictionary keys by @davidwendt in https://github.com/NVIDIA/cudf/pull/21037 +* Remove deprecated pylibcudf arrow APIs by @Matt711 in https://github.com/NVIDIA/cudf/pull/21284 +* Don't gate ColumnBase.to_pandas behind mode.pandas_compatible by @mroeschke in https://github.com/NVIDIA/cudf/pull/21256 +* Remove pandas compatibility gating for dtypes in DataFrame APIs by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21261 +* Remove pandas compatibility gating for dtypes in decimal column by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21243 +* Remove pandas compatibility gating for dtypes in rolling window by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21260 +* remove pip.conf migration code in CI scripts, update CI-skipping rules by @jameslamb in https://github.com/NVIDIA/cudf/pull/21303 +* Don't hard-code branch names by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/21270 +* Change CUB If and FlaggedIf to use cudf::detail::copy_if by @davidwendt in https://github.com/NVIDIA/cudf/pull/21274 +* Migrate to rapidsmpf ``ChannelMetadata`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/21275 +* Remove deprecation warnings from benchmarks/ndsh/utilities.cpp by @davidwendt in https://github.com/NVIDIA/cudf/pull/21308 +* Improve pylibcudf usage in cudf by @vyasr in https://github.com/NVIDIA/cudf/pull/21285 +* Continue refactoring column construction by @vyasr in https://github.com/NVIDIA/cudf/pull/21262 +* Is already partitioned by @rjzamora in https://github.com/NVIDIA/cudf/pull/21310 +* Consolidate cudf-classic dtype validators to their own file and make them stricter by @mroeschke in https://github.com/NVIDIA/cudf/pull/21281 +* Preprocess parquet repetition and definition levels by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/21139 +* Remove pandas compatibility dtypes gating in series methods by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21257 +* Implement `na` for `match` and `contains` in `str` API by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21236 +* Remove redundant lowering and dispatching infrastructure for the rapidsmpf runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/21337 +* Full structlog-based tracing by @rjzamora in https://github.com/NVIDIA/cudf/pull/21316 +* Remove some unneeded libcudf debug-specific code paths by @davidwendt in https://github.com/NVIDIA/cudf/pull/21335 +* Move has_nulls template parameter to runtime in rolling window by @davidwendt in https://github.com/NVIDIA/cudf/pull/21312 +* Enable KvikIO datasource in hybrid scan examples by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21318 +* Reserve more collective IDs for "dynamic" IR nodes by @rjzamora in https://github.com/NVIDIA/cudf/pull/21343 +* Deprecate `values_host` in favor of `to_numpy` by @Matt711 in https://github.com/NVIDIA/cudf/pull/21346 +* refactor: build wheels and conda packages using Python limited API by @gforsyth in https://github.com/NVIDIA/cudf/pull/21329 +* Ensure ColumnBase.create asserts that plc_column type is equivalent to passed dtype by @mroeschke in https://github.com/NVIDIA/cudf/pull/21325 +* Move generate_nested_types.hpp definitions to .cpp file by @davidwendt in https://github.com/NVIDIA/cudf/pull/21344 +* Remove dead code checking scheduler in fallback logic by @Matt711 in https://github.com/NVIDIA/cudf/pull/21351 +* Skip failing cuDF Python tests on RTX PRO 6000 by @bdice in https://github.com/NVIDIA/cudf/pull/21358 +* Remove mode.pandas_compatible from StringColumn.all_characters_of_type by @mroeschke in https://github.com/NVIDIA/cudf/pull/21356 +* Remove export from cudf::detail::sort and cudf::detail::merge APIs by @davidwendt in https://github.com/NVIDIA/cudf/pull/21345 +* Cleanup duplicated code in hybrid scan by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21353 +* Make custreamz and dask-cudf independent of cuda-version by @vyasr in https://github.com/NVIDIA/cudf/pull/21348 +* Use ColumnBase.create more in cudf classic APIs by @mroeschke in https://github.com/NVIDIA/cudf/pull/21306 +* Avoid single-partition assumptions when dynamic-planning is enabled by @rjzamora in https://github.com/NVIDIA/cudf/pull/21370 +* Add a fallback-mode cli arg to the pdsh benchmarks by @Matt711 in https://github.com/NVIDIA/cudf/pull/21365 +* Add benchmarks for dictionary set_keys and sort by @davidwendt in https://github.com/NVIDIA/cudf/pull/21366 +* fix(deps): bump `pyarrow` floor in conda recipes by @gforsyth in https://github.com/NVIDIA/cudf/pull/21383 +* More create usage and tightening up dtype specification by @vyasr in https://github.com/NVIDIA/cudf/pull/21359 +* Optimize cudf usage of pylibcudf by @vyasr in https://github.com/NVIDIA/cudf/pull/21362 +* Include a run_id in pds-h benchmark run by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21375 +* Capture query plan in pds-h output by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21315 +* Remove cudf compat gating at more places by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21350 +* Use ColumnBase.create more in frame.py by @mroeschke in https://github.com/NVIDIA/cudf/pull/21394 +* Remove `._with_dtype_metadata` in cudf classic groupby by @mroeschke in https://github.com/NVIDIA/cudf/pull/21389 +* Allocate single compressed buffer per source in parquet reader by @vuule in https://github.com/NVIDIA/cudf/pull/21323 +* Preserve ArrowDtype in `Series.struct.fields` by @mroeschke in https://github.com/NVIDIA/cudf/pull/21379 +* Use ColumnBase.create instead of from_pylibcudf in indexed_frame.py by @mroeschke in https://github.com/NVIDIA/cudf/pull/21401 +* Use GHA id-token for `sccache-dist` auth token by @trxcllnt in https://github.com/NVIDIA/cudf/pull/21406 +* Remove more mode.pandas_compatible from StringColumn by @mroeschke in https://github.com/NVIDIA/cudf/pull/21393 +* Rewrite CategoricalColumn operations in terms of pylibcudf and avoid `_with_type_metadata` by @vyasr in https://github.com/NVIDIA/cudf/pull/21402 +* Remove from_pylibcudf/_with_dtype_metadata in IO readers/rolling by @mroeschke in https://github.com/NVIDIA/cudf/pull/21409 +* Remove Frame._copy_type_metadata method by @vyasr in https://github.com/NVIDIA/cudf/pull/21405 +* Apply PIMPL idiom to sort-merge join by @shrshi in https://github.com/NVIDIA/cudf/pull/21349 +* Avoid `_with_dtype_metadata` in dask_cudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/21411 +* Remove hash_vocab_utils by @vyasr in https://github.com/NVIDIA/cudf/pull/21419 +* Refactor find_and_replace in Numerical and String columns by @vyasr in https://github.com/NVIDIA/cudf/pull/21413 +* Use ColumnBase.create over .from_pylibcudf in MultiIndex, groupby, misc. by @mroeschke in https://github.com/NVIDIA/cudf/pull/21423 +* Skip RTX PRO 6000 tests in test workflow by @vyasr in https://github.com/NVIDIA/cudf/pull/21431 +* Split up algorithm.cuh into reduce.cuh and copy_if.cuh by @davidwendt in https://github.com/NVIDIA/cudf/pull/21387 +* Put typename disambiguator before nested name specifier by @seanbaxter in https://github.com/NVIDIA/cudf/pull/21408 +* Use ColumnBase.create over .from_pylibcudf in dataframe/index/indexed_frame.py by @mroeschke in https://github.com/NVIDIA/cudf/pull/21439 +* Remove mode.pandas_compatable from ColumnBase.cast by @mroeschke in https://github.com/NVIDIA/cudf/pull/21441 +* Remove mode.pandas_compatible from ColumnBase._validate_args by @mroeschke in https://github.com/NVIDIA/cudf/pull/21426 +* Make rapidsmpf py_executor thread pool size configurable by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21446 +* Replace `thrust::constant_iterator` with `cuda::constant_iterator` by @miscco in https://github.com/NVIDIA/cudf/pull/21445 +* Remove _with_type_metadata by @vyasr in https://github.com/NVIDIA/cudf/pull/21421 +* Drop `thrust::reverse_iterator` in favor of `cuda::std::reverse_iterator` by @miscco in https://github.com/NVIDIA/cudf/pull/21455 +* Move detail::sorted_order decl from sort_column_impl.cuh to sort.hpp by @davidwendt in https://github.com/NVIDIA/cudf/pull/21436 +* Remove remaining uses of ColumnBase.from_pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/21448 +* Add public bitmask_and with host_span parameters by @davidwendt in https://github.com/NVIDIA/cudf/pull/21437 +* RapidsMPF renamed `Node` to `Actor` by @madsbk in https://github.com/NVIDIA/cudf/pull/21476 +* Reduce nvbench parameter space for high-multiplicity and sort-merge join benchmarks by @shrshi in https://github.com/NVIDIA/cudf/pull/21076 +* Reduce PR workflow jobs for cpp-only changes by @davidwendt in https://github.com/NVIDIA/cudf/pull/21381 +* Optimize PDSH queries for `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21462 +* Pin polars >=1.30 and <1.39 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21304 +* Use param substitutions from dsqgen by @Matt711 in https://github.com/NVIDIA/cudf/pull/21465 +* Make cudf-polars config generic over its executor by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21459 +* Allow running PDS-DS with qualification parameters by @Matt711 in https://github.com/NVIDIA/cudf/pull/21478 +* Use new local_partitions method of shuffler by @wence- in https://github.com/NVIDIA/cudf/pull/21461 +* Unify normalization, buffer wrapping, and validation by @vyasr in https://github.com/NVIDIA/cudf/pull/21451 +* Update/remove version checks for outdated versions in Python libraries by @mroeschke in https://github.com/NVIDIA/cudf/pull/21483 +* check-nightly-ci: update to new version by @jameslamb in https://github.com/NVIDIA/cudf/pull/21464 +* Remove pyarrow string_view workarounds in pylibcudf/cudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/21490 +* Remove dead decimal alignment code by @Matt711 in https://github.com/NVIDIA/cudf/pull/21474 +* Move Column normalization to only required code paths by @vyasr in https://github.com/NVIDIA/cudf/pull/21487 +* Rename columns to column_names in parquet reader options by @Matt711 in https://github.com/NVIDIA/cudf/pull/21473 +* Fix more deprecated calls in JNI source by @davidwendt in https://github.com/NVIDIA/cudf/pull/21482 +* Remove inefficiencies in gather by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21429 +* pds-h validation improvements by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21388 +* check-nightly-ci: remove testing config by @jameslamb in https://github.com/NVIDIA/cudf/pull/21494 +* Update RapidsMPF's `Context()` since it no longer takes a statistics argument. by @madsbk in https://github.com/NVIDIA/cudf/pull/21503 +* Drop uses of thrust/functional by @miscco in https://github.com/NVIDIA/cudf/pull/21477 +* Refactor some common utilities out of find_and_replace by @vyasr in https://github.com/NVIDIA/cudf/pull/21500 +* Sort specialization logic for dictionary column by @davidwendt in https://github.com/NVIDIA/cudf/pull/21420 +* Add cudf::detail::copy_if_async utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/21495 +* Update range values for gather/scatter benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/21479 +* Replace some cudf::detail calls with public API equivalents by @davidwendt in https://github.com/NVIDIA/cudf/pull/21507 +* Reenable pandas tests disabled due to prior CoW incompatibilities by @vyasr in https://github.com/NVIDIA/cudf/pull/21510 +* Add utility to extract partitioning moduli from the ChannelMetadata by @rjzamora in https://github.com/NVIDIA/cudf/pull/21491 +* Refresh cudf.pandas xpass list by @vyasr in https://github.com/NVIDIA/cudf/pull/21521 +* List and struct support for make_empty_column by @quasiben in https://github.com/NVIDIA/cudf/pull/21493 +* Create ColumnBase.from_range to reduce as_column usage by @mroeschke in https://github.com/NVIDIA/cudf/pull/21516 +* Add properties in serialized `explain` for Scan and Shuffle by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21497 +* Implement sink nodes for rapidsmpf runtime by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20965 +* Reduce CI by skipping tests unaffected by changes based on package dependency chain by @Matt711 in https://github.com/NVIDIA/cudf/pull/21506 +* Remove unneeded specialized device_vector/device_span gtest by @davidwendt in https://github.com/NVIDIA/cudf/pull/21508 +* Rework reductions/scan_tests.cpp to help with adding aggregation types by @davidwendt in https://github.com/NVIDIA/cudf/pull/21511 +* Fix pdsh benchmark help text by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21524 +* Update casts in pds-h for --validate-directory by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21534 +* Change thrust::make_discard_iterator to cuda::make_discard_iterator by @davidwendt in https://github.com/NVIDIA/cudf/pull/21526 +* Avoid cudf.dtype dtype parsing internally by @mroeschke in https://github.com/NVIDIA/cudf/pull/21522 +* Use ``--rmm-release-threshold`` for both single- and multi-gpu benchmarks by @rjzamora in https://github.com/NVIDIA/cudf/pull/21535 +* Test cudf-polars with the "rapidsmpf" runtime and "distributed" cluster by @rjzamora in https://github.com/NVIDIA/cudf/pull/21028 +* chore(deps): bump numba ceiling to <0.65.0 by @gforsyth in https://github.com/NVIDIA/cudf/pull/21543 +* Support JIT filtering of join output indices by @shrshi in https://github.com/NVIDIA/cudf/pull/21202 +* Use stricter dtype object checks in cudf classic by @mroeschke in https://github.com/NVIDIA/cudf/pull/21527 +* Use new modules for RMM CudaStreamFlags by @bdice in https://github.com/NVIDIA/cudf/pull/21559 +* chore(deps): bump numba upper bound in conda recipe by @gforsyth in https://github.com/NVIDIA/cudf/pull/21561 +* Validate PDS-DS Queries 1-16 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21556 +* Update conftest-patch for minor numpy changes by @vyasr in https://github.com/NVIDIA/cudf/pull/21568 +* Create wrapper for pylibcudf function calls by @vyasr in https://github.com/NVIDIA/cudf/pull/21501 +* Set track-stream-ordered-races=no in compute-sanitizer racecheck for LARGE_STRINGS_TEST by @davidwendt in https://github.com/NVIDIA/cudf/pull/21562 +* Skip scan decimal introspection logic when there are no decimals in the schema by @Matt711 in https://github.com/NVIDIA/cudf/pull/21571 +* Add per-file, unprivileged page cache dropping to cuDF benchmarks by @kingcrimsontianyu in https://github.com/NVIDIA/cudf/pull/21397 +* Use non-blocking streams in libcudf's internal stream pool by @vuule in https://github.com/NVIDIA/cudf/pull/21449 +* Make cudf-polars pds-h script more robust by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21538 +* Remove `Index` materialization in all query results by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21542 * refactor(limited api): add explicit `wheel.py-api` to `pyproject.toml` - by @gforsyth in https://github.com/rapidsai/cudf/pull/21560 -* Fix make_empty_histogram_like to call make_empty_column by @davidwendt in https://github.com/rapidsai/cudf/pull/21558 -* Make IntervalColumn more agnostic to pandas extension types by @mroeschke in https://github.com/rapidsai/cudf/pull/21589 -* Use UTC time more in DatetimeTZColumn by @mroeschke in https://github.com/rapidsai/cudf/pull/21566 -* Add validation options to cudf.pandas PDSH runner by @mroeschke in https://github.com/rapidsai/cudf/pull/21569 -* Avoid ArrowDtype conversion in ColumnBase.create by @mroeschke in https://github.com/rapidsai/cudf/pull/21528 -* Save duckdb "gold" results to a directory without validating by @Matt711 in https://github.com/rapidsai/cudf/pull/21599 -* Add null_count() back to cudf::column_device_view by @davidwendt in https://github.com/rapidsai/cudf/pull/21430 -* Make Struct/ListColumn respect copy(deep=True) by @mroeschke in https://github.com/rapidsai/cudf/pull/21583 -* Use `plc.interop.ColumnMetadata` for more `ColumnBase.to_arrow` conversions by @mroeschke in https://github.com/rapidsai/cudf/pull/21586 -* Avoid using `pyarrow.Scalar` as intermediary in cuDF by @mroeschke in https://github.com/rapidsai/cudf/pull/21591 -* Add hybrid scan multithreaded benchmark by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21363 -* Pass timeout-minutes to devcontainer CI job by @mroeschke in https://github.com/rapidsai/cudf/pull/21610 -* Update Cython lower bound pin to 3.2.2 by @vyasr in https://github.com/rapidsai/cudf/pull/21612 -* feat: add support for Python 3.14 by @gforsyth in https://github.com/rapidsai/cudf/pull/21540 -* Pin upper bound of numba-cuda to <0.28.0 by @mroeschke in https://github.com/rapidsai/cudf/pull/21628 -* Handle float edge cases in pds-h validation by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21585 -* Add pylibcudf interface for cudf::make_empty_lists_column by @davidwendt in https://github.com/rapidsai/cudf/pull/21564 -* Validate PDS-DS Queries 63-72 by @Matt711 in https://github.com/rapidsai/cudf/pull/21588 -* Pass `plc.interop.ColumnMetadata` to arrow conversion in `ColumnBase.element_indexing` by @mroeschke in https://github.com/rapidsai/cudf/pull/21587 -* Adjust retried exceptions in fetch-versions by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20954 -* Use PylibcudfFunction in more places by @vyasr in https://github.com/rapidsai/cudf/pull/21611 -* Add cudf::is_dictionary_key() traits utilities by @davidwendt in https://github.com/rapidsai/cudf/pull/21574 -* chore: followup fixes for Python 3.14 and test filtering by @gforsyth in https://github.com/rapidsai/cudf/pull/21622 -* Preserve partitioning metadata after simple ``Select`` operations by @rjzamora in https://github.com/rapidsai/cudf/pull/21547 -* Validate PDS-DS Queries 17-23 by @Matt711 in https://github.com/rapidsai/cudf/pull/21594 -* Make Struct/ListColumn more agnostic to pandas extension types by @mroeschke in https://github.com/rapidsai/cudf/pull/21550 -* Align JIT filter join naming with libcudf convention by @shrshi in https://github.com/rapidsai/cudf/pull/21621 -* Validate PDS-DS Queries 24-29 by @Matt711 in https://github.com/rapidsai/cudf/pull/21592 -* Add shift and shift_fill support by @quasiben in https://github.com/rapidsai/cudf/pull/21662 -* include pytorch conda packages in CUDA 13 test env by @jameslamb in https://github.com/rapidsai/cudf/pull/21663 -* Update defaults for the "rapidsmpf" runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/21668 -* Remove validation from normalization by @vyasr in https://github.com/rapidsai/cudf/pull/21664 -* Minor follow-up cleanups from merged PRs by @vyasr in https://github.com/rapidsai/cudf/pull/21658 -* Introduce packed_metadata_view utility for introspecting `cudf::pack`s metadata wire format by @wence- in https://github.com/rapidsai/cudf/pull/21670 -* Reduce parameters for JSON_READER_NVBENCH to improve setup runtime by @davidwendt in https://github.com/rapidsai/cudf/pull/21666 -* Split up PARQUET_READER_NVBENCH benchmark by @davidwendt in https://github.com/rapidsai/cudf/pull/21667 -* Improvement for libcudf benchmark utility create_structs_data by @davidwendt in https://github.com/rapidsai/cudf/pull/21657 -* Validate PDS-DS Q30-36 by @Matt711 in https://github.com/rapidsai/cudf/pull/21576 -* Enable dynamic-planning by default with the "rapidsmpf" runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/21675 -* Use sccache and sccache-dist in spark-rapids-jni workflow by @trxcllnt in https://github.com/rapidsai/cudf/pull/21682 -* Validate PDS-DS Queries 37-43 by @Matt711 in https://github.com/rapidsai/cudf/pull/21615 -* Unpin `numba-cuda` upper bound and adjust groupby.apply fallback conditons by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/21637 -* Validate PDS-DS Queries 44-48 by @Matt711 in https://github.com/rapidsai/cudf/pull/21634 -* Validate PDS-DS Queries 49-52 by @Matt711 in https://github.com/rapidsai/cudf/pull/21636 -* Validate PDS-DS Queries 53-57 by @Matt711 in https://github.com/rapidsai/cudf/pull/21633 -* Use PylibcudfFunction for simple table functions by @vyasr in https://github.com/rapidsai/cudf/pull/21678 -* Validate PDS-DS Queries 58-62 by @Matt711 in https://github.com/rapidsai/cudf/pull/21640 -* Validate PDS-DS Queries 78-82 by @Matt711 in https://github.com/rapidsai/cudf/pull/21648 -* Replace max_precision() utility with STL function by @davidwendt in https://github.com/rapidsai/cudf/pull/21691 -* Change detail calls of mask functions to public equivalent by @davidwendt in https://github.com/rapidsai/cudf/pull/21563 -* Use cudaMemcpyBatchAsync by @bdice in https://github.com/rapidsai/cudf/pull/20800 -* Remove `nvidia-libnvcomp` from `dependencies.yaml` by @trxcllnt in https://github.com/rapidsai/cudf/pull/21679 -* Validate PDS-DS Queries 73-77 by @Matt711 in https://github.com/rapidsai/cudf/pull/21647 -* Validate PDS-DS Queries 86-92 by @Matt711 in https://github.com/rapidsai/cudf/pull/21617 -* Embed PDS-DS query parameters as Python module instead of JSON file by @Matt711 in https://github.com/rapidsai/cudf/pull/21715 -* Validate PDS-DS Queries 93-99 by @Matt711 in https://github.com/rapidsai/cudf/pull/21618 -* Remove unused dictionary::detail::get_insert_index utility by @davidwendt in https://github.com/rapidsai/cudf/pull/21711 -* Use `CUDF_EXPECTS` when the condition has no side effects by @vuule in https://github.com/rapidsai/cudf/pull/21722 -* Validate PDS-DS Queries 83-85 by @Matt711 in https://github.com/rapidsai/cudf/pull/21649 -* Remove libcudf gtest detail_gather_test.cu by @davidwendt in https://github.com/rapidsai/cudf/pull/21735 -* Move true_if device function from detail/.hpp to src/.cuh header by @davidwendt in https://github.com/rapidsai/cudf/pull/21717 -* RapidsMPF SPMD mode in PDSH benchmarks by @madsbk in https://github.com/rapidsai/cudf/pull/21652 -* Remove benchmark of benchmark utility create_structs_data by @davidwendt in https://github.com/rapidsai/cudf/pull/21724 -* Use ThreadPoolExecutor contextmanager to ensure graceful shutdown in cudf_polars with rapidsmpf by @mroeschke in https://github.com/rapidsai/cudf/pull/21753 -* Remove unneeded internal make_strings_column factory by @davidwendt in https://github.com/rapidsai/cudf/pull/21747 -* Use CUB topK as fast-path for cudf::top_k by @davidwendt in https://github.com/rapidsai/cudf/pull/21582 -* fix verify-hardcoded-versions issues by @jameslamb in https://github.com/rapidsai/cudf/pull/21770 -* Cap numba-cuda upper bound at <0.29.0 by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/21785 -* enforce a floor on libnvjitlink, build wheels with CUDA 13.0.x, test wheels against mix of CTK versions by @jameslamb in https://github.com/rapidsai/cudf/pull/21671 + by @gforsyth in https://github.com/NVIDIA/cudf/pull/21560 +* Fix make_empty_histogram_like to call make_empty_column by @davidwendt in https://github.com/NVIDIA/cudf/pull/21558 +* Make IntervalColumn more agnostic to pandas extension types by @mroeschke in https://github.com/NVIDIA/cudf/pull/21589 +* Use UTC time more in DatetimeTZColumn by @mroeschke in https://github.com/NVIDIA/cudf/pull/21566 +* Add validation options to cudf.pandas PDSH runner by @mroeschke in https://github.com/NVIDIA/cudf/pull/21569 +* Avoid ArrowDtype conversion in ColumnBase.create by @mroeschke in https://github.com/NVIDIA/cudf/pull/21528 +* Save duckdb "gold" results to a directory without validating by @Matt711 in https://github.com/NVIDIA/cudf/pull/21599 +* Add null_count() back to cudf::column_device_view by @davidwendt in https://github.com/NVIDIA/cudf/pull/21430 +* Make Struct/ListColumn respect copy(deep=True) by @mroeschke in https://github.com/NVIDIA/cudf/pull/21583 +* Use `plc.interop.ColumnMetadata` for more `ColumnBase.to_arrow` conversions by @mroeschke in https://github.com/NVIDIA/cudf/pull/21586 +* Avoid using `pyarrow.Scalar` as intermediary in cuDF by @mroeschke in https://github.com/NVIDIA/cudf/pull/21591 +* Add hybrid scan multithreaded benchmark by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21363 +* Pass timeout-minutes to devcontainer CI job by @mroeschke in https://github.com/NVIDIA/cudf/pull/21610 +* Update Cython lower bound pin to 3.2.2 by @vyasr in https://github.com/NVIDIA/cudf/pull/21612 +* feat: add support for Python 3.14 by @gforsyth in https://github.com/NVIDIA/cudf/pull/21540 +* Pin upper bound of numba-cuda to <0.28.0 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21628 +* Handle float edge cases in pds-h validation by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21585 +* Add pylibcudf interface for cudf::make_empty_lists_column by @davidwendt in https://github.com/NVIDIA/cudf/pull/21564 +* Validate PDS-DS Queries 63-72 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21588 +* Pass `plc.interop.ColumnMetadata` to arrow conversion in `ColumnBase.element_indexing` by @mroeschke in https://github.com/NVIDIA/cudf/pull/21587 +* Adjust retried exceptions in fetch-versions by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20954 +* Use PylibcudfFunction in more places by @vyasr in https://github.com/NVIDIA/cudf/pull/21611 +* Add cudf::is_dictionary_key() traits utilities by @davidwendt in https://github.com/NVIDIA/cudf/pull/21574 +* chore: followup fixes for Python 3.14 and test filtering by @gforsyth in https://github.com/NVIDIA/cudf/pull/21622 +* Preserve partitioning metadata after simple ``Select`` operations by @rjzamora in https://github.com/NVIDIA/cudf/pull/21547 +* Validate PDS-DS Queries 17-23 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21594 +* Make Struct/ListColumn more agnostic to pandas extension types by @mroeschke in https://github.com/NVIDIA/cudf/pull/21550 +* Align JIT filter join naming with libcudf convention by @shrshi in https://github.com/NVIDIA/cudf/pull/21621 +* Validate PDS-DS Queries 24-29 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21592 +* Add shift and shift_fill support by @quasiben in https://github.com/NVIDIA/cudf/pull/21662 +* include pytorch conda packages in CUDA 13 test env by @jameslamb in https://github.com/NVIDIA/cudf/pull/21663 +* Update defaults for the "rapidsmpf" runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/21668 +* Remove validation from normalization by @vyasr in https://github.com/NVIDIA/cudf/pull/21664 +* Minor follow-up cleanups from merged PRs by @vyasr in https://github.com/NVIDIA/cudf/pull/21658 +* Introduce packed_metadata_view utility for introspecting `cudf::pack`s metadata wire format by @wence- in https://github.com/NVIDIA/cudf/pull/21670 +* Reduce parameters for JSON_READER_NVBENCH to improve setup runtime by @davidwendt in https://github.com/NVIDIA/cudf/pull/21666 +* Split up PARQUET_READER_NVBENCH benchmark by @davidwendt in https://github.com/NVIDIA/cudf/pull/21667 +* Improvement for libcudf benchmark utility create_structs_data by @davidwendt in https://github.com/NVIDIA/cudf/pull/21657 +* Validate PDS-DS Q30-36 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21576 +* Enable dynamic-planning by default with the "rapidsmpf" runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/21675 +* Use sccache and sccache-dist in spark-rapids-jni workflow by @trxcllnt in https://github.com/NVIDIA/cudf/pull/21682 +* Validate PDS-DS Queries 37-43 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21615 +* Unpin `numba-cuda` upper bound and adjust groupby.apply fallback conditons by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/21637 +* Validate PDS-DS Queries 44-48 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21634 +* Validate PDS-DS Queries 49-52 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21636 +* Validate PDS-DS Queries 53-57 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21633 +* Use PylibcudfFunction for simple table functions by @vyasr in https://github.com/NVIDIA/cudf/pull/21678 +* Validate PDS-DS Queries 58-62 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21640 +* Validate PDS-DS Queries 78-82 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21648 +* Replace max_precision() utility with STL function by @davidwendt in https://github.com/NVIDIA/cudf/pull/21691 +* Change detail calls of mask functions to public equivalent by @davidwendt in https://github.com/NVIDIA/cudf/pull/21563 +* Use cudaMemcpyBatchAsync by @bdice in https://github.com/NVIDIA/cudf/pull/20800 +* Remove `nvidia-libnvcomp` from `dependencies.yaml` by @trxcllnt in https://github.com/NVIDIA/cudf/pull/21679 +* Validate PDS-DS Queries 73-77 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21647 +* Validate PDS-DS Queries 86-92 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21617 +* Embed PDS-DS query parameters as Python module instead of JSON file by @Matt711 in https://github.com/NVIDIA/cudf/pull/21715 +* Validate PDS-DS Queries 93-99 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21618 +* Remove unused dictionary::detail::get_insert_index utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/21711 +* Use `CUDF_EXPECTS` when the condition has no side effects by @vuule in https://github.com/NVIDIA/cudf/pull/21722 +* Validate PDS-DS Queries 83-85 by @Matt711 in https://github.com/NVIDIA/cudf/pull/21649 +* Remove libcudf gtest detail_gather_test.cu by @davidwendt in https://github.com/NVIDIA/cudf/pull/21735 +* Move true_if device function from detail/.hpp to src/.cuh header by @davidwendt in https://github.com/NVIDIA/cudf/pull/21717 +* RapidsMPF SPMD mode in PDSH benchmarks by @madsbk in https://github.com/NVIDIA/cudf/pull/21652 +* Remove benchmark of benchmark utility create_structs_data by @davidwendt in https://github.com/NVIDIA/cudf/pull/21724 +* Use ThreadPoolExecutor contextmanager to ensure graceful shutdown in cudf_polars with rapidsmpf by @mroeschke in https://github.com/NVIDIA/cudf/pull/21753 +* Remove unneeded internal make_strings_column factory by @davidwendt in https://github.com/NVIDIA/cudf/pull/21747 +* Use CUB topK as fast-path for cudf::top_k by @davidwendt in https://github.com/NVIDIA/cudf/pull/21582 +* fix verify-hardcoded-versions issues by @jameslamb in https://github.com/NVIDIA/cudf/pull/21770 +* Cap numba-cuda upper bound at <0.29.0 by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/21785 +* enforce a floor on libnvjitlink, build wheels with CUDA 13.0.x, test wheels against mix of CTK versions by @jameslamb in https://github.com/NVIDIA/cudf/pull/21671 ## New Contributors -* @jrbourbeau made their first contribution in https://github.com/rapidsai/cudf/pull/21134 -* @Umang-projects made their first contribution in https://github.com/rapidsai/cudf/pull/21068 -* @drivanov made their first contribution in https://github.com/rapidsai/cudf/pull/21400 -* @seanbaxter made their first contribution in https://github.com/rapidsai/cudf/pull/21408 -* @thecaptain789 made their first contribution in https://github.com/rapidsai/cudf/pull/21380 +* @jrbourbeau made their first contribution in https://github.com/NVIDIA/cudf/pull/21134 +* @Umang-projects made their first contribution in https://github.com/NVIDIA/cudf/pull/21068 +* @drivanov made their first contribution in https://github.com/NVIDIA/cudf/pull/21400 +* @seanbaxter made their first contribution in https://github.com/NVIDIA/cudf/pull/21408 +* @thecaptain789 made their first contribution in https://github.com/NVIDIA/cudf/pull/21380 -**Full Changelog**: https://github.com/rapidsai/cudf/compare/v26.04.00a...release/26.04 +**Full Changelog**: https://github.com/NVIDIA/cudf/compare/v26.04.00a...release/26.04 # cudf 26.02.00 (4 Feb 2026) ### 🚨 Breaking Changes -* Avoid counting nulls and creating null mask in groupby aggregation `MERGE_M2` by @ttnghia in https://github.com/rapidsai/cudf/pull/20716 -* Remove cudf::get_current_device_resource by @bdice in https://github.com/rapidsai/cudf/pull/20688 -* Avoid creating null mask in groupby aggregation `M2` by @ttnghia in https://github.com/rapidsai/cudf/pull/20726 -* Remove deprecated left semi- and anti- join APIs by @shrshi in https://github.com/rapidsai/cudf/pull/20668 -* Inline and simplify some column methods by @vyasr in https://github.com/rapidsai/cudf/pull/20819 -* Enable copy-on-write in cudf.pandas by @vyasr in https://github.com/rapidsai/cudf/pull/20401 -* [FEA] Improve Null-Aware Operator Support in AST-Codegen by @lamarrr in https://github.com/rapidsai/cudf/pull/20206 -* Remove legacy hash-combine logic and unify hashing with row hasher by @PointKernel in https://github.com/rapidsai/cudf/pull/20796 -* Remove deprecated .from_pandas constructors by @mroeschke in https://github.com/rapidsai/cudf/pull/20925 -* Remove deprecated Series.data by @mroeschke in https://github.com/rapidsai/cudf/pull/20914 -* Remove all base attributes from ColumnBase by @vyasr in https://github.com/rapidsai/cudf/pull/20961 -* Fix handling of unquoted strings in the CSV reader by @vuule in https://github.com/rapidsai/cudf/pull/20996 +* Avoid counting nulls and creating null mask in groupby aggregation `MERGE_M2` by @ttnghia in https://github.com/NVIDIA/cudf/pull/20716 +* Remove cudf::get_current_device_resource by @bdice in https://github.com/NVIDIA/cudf/pull/20688 +* Avoid creating null mask in groupby aggregation `M2` by @ttnghia in https://github.com/NVIDIA/cudf/pull/20726 +* Remove deprecated left semi- and anti- join APIs by @shrshi in https://github.com/NVIDIA/cudf/pull/20668 +* Inline and simplify some column methods by @vyasr in https://github.com/NVIDIA/cudf/pull/20819 +* Enable copy-on-write in cudf.pandas by @vyasr in https://github.com/NVIDIA/cudf/pull/20401 +* [FEA] Improve Null-Aware Operator Support in AST-Codegen by @lamarrr in https://github.com/NVIDIA/cudf/pull/20206 +* Remove legacy hash-combine logic and unify hashing with row hasher by @PointKernel in https://github.com/NVIDIA/cudf/pull/20796 +* Remove deprecated .from_pandas constructors by @mroeschke in https://github.com/NVIDIA/cudf/pull/20925 +* Remove deprecated Series.data by @mroeschke in https://github.com/NVIDIA/cudf/pull/20914 +* Remove all base attributes from ColumnBase by @vyasr in https://github.com/NVIDIA/cudf/pull/20961 +* Fix handling of unquoted strings in the CSV reader by @vuule in https://github.com/NVIDIA/cudf/pull/20996 ### 🐛 Bug Fixes -* Avoid duplicate streaming nodes for the rapidsmpf runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/20586 -* Handle scalar arguments in ternary expression by @Matt711 in https://github.com/rapidsai/cudf/pull/20600 -* fix(noarch): use noarch build script in noarch build by @gforsyth in https://github.com/rapidsai/cudf/pull/20654 -* fix(conda): matrix out noarch builds by cuda-major version by @gforsyth in https://github.com/rapidsai/cudf/pull/20678 -* Include RMM in type checking environment and update type annotations for optional `stream` by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20636 -* Add no-op path for `ArrowExtensionArray.astype` by @Matt711 in https://github.com/rapidsai/cudf/pull/20580 -* Skip pytorch integration tests if CUDA is not available by @Matt711 in https://github.com/rapidsai/cudf/pull/20729 -* Always delay CUDA Array Interface pointer access by @vyasr in https://github.com/rapidsai/cudf/pull/20719 -* Fix various copy-on-write bugs by @vyasr in https://github.com/rapidsai/cudf/pull/20744 -* Fix leaks in cuDF java tests by @abellina in https://github.com/rapidsai/cudf/pull/20767 -* Fix plc.Scalar.from_py(datetime.datetime) incorrectly localizing naive datetimes by @mroeschke in https://github.com/rapidsai/cudf/pull/20769 -* Don't remove double casts in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/20773 -* Fixes struct column handling in sort-merge joins by @shrshi in https://github.com/rapidsai/cudf/pull/20664 -* Fix for `synccheck` compute-sanitizer errors across Parquet gtest by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20775 -* Pin `numpy<2.4.0a0` in mypy pre-commit environment by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20781 -* Raise when trying to run queries on different devices in same process by @wence- in https://github.com/rapidsai/cudf/pull/20617 -* Ensure `min_periods=0` is passed through rolling aggregations by @Matt711 in https://github.com/rapidsai/cudf/pull/20653 -* Fix racecheck errors in the ORC reader by @vuule in https://github.com/rapidsai/cudf/pull/20792 -* Fix the crash of multi-threaded parquet reader benchmark by @kingcrimsontianyu in https://github.com/rapidsai/cudf/pull/20783 -* Fix racecheck reported by DATA_CHUNK_SOURCE_TEST in inflate_kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20804 -* Fix racecheck in the gpu_debrotli_kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20806 -* Ensure literal groupby aggregations are broadcasted to key length in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/20776 -* Pin `aiobotocore<3` to fix CI failures by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20844 -* Fix racecheck in parquet decode_page_data_generic kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20850 -* Avoid generating empty `TableChunks` in streaming scan nodes by @rjzamora in https://github.com/rapidsai/cudf/pull/20815 -* Fix dask imports in ``CudfFusedParquetIOHost`` by @rjzamora in https://github.com/rapidsai/cudf/pull/20845 -* Fix UB due to OOM Exception in ParquetReaderTest.ManyLargeLists by @lamarrr in https://github.com/rapidsai/cudf/pull/20841 -* Fix racecheck/synccheck in JSON parse_fn_string_parallel kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20856 -* Fix racecheck in ORC decode_column_data_kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20853 -* Disable flatbuffers tests in CMake configuration by @bdice in https://github.com/rapidsai/cudf/pull/20848 -* Upper bound on aiosqlite in polars-upstream job by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20866 -* Fix boolean casting consistency with Pandas (#20746) by @aryansri05 in https://github.com/rapidsai/cudf/pull/20747 -* Add retries to requests made to PyPI's JSON API by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20865 -* Fix `size_type` overflow in multiple APIs by @vuule in https://github.com/rapidsai/cudf/pull/20857 -* Fix racecheck in parquet compute_string_page_bounds_kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20868 -* Fix dictionary::encode to honor indices-type parameter by @davidwendt in https://github.com/rapidsai/cudf/pull/20842 -* Add missing headers to row_ir.hpp, row_ir.cpp by @bdice in https://github.com/rapidsai/cudf/pull/20834 -* Fix `parquet_options` in pdsh benchmark by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20893 -* Add stream synchronize to tdigest generate_group_cluster_info by @davidwendt in https://github.com/rapidsai/cudf/pull/20846 -* Only install RMM in mypy env on linux by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20878 -* Make nvcomp export unconditional by @vyasr in https://github.com/rapidsai/cudf/pull/20828 -* Ensure we have nvjitlink from the CUDA version used at build time or newer and upgrade numba-cuda lower bound by @bdice in https://github.com/rapidsai/cudf/pull/20873 -* Fix size_type overflow in the ORC writer by @vuule in https://github.com/rapidsai/cudf/pull/20889 -* Constrain pyparsing version by @vyasr in https://github.com/rapidsai/cudf/pull/20935 -* Revert #20902 by @vyasr in https://github.com/rapidsai/cudf/pull/20955 -* Add force-blocking-launches to run_compute_sanitizer_test script by @davidwendt in https://github.com/rapidsai/cudf/pull/20962 -* Fix racecheck error in parquet delta_byte_array_decoder::string_scan by @davidwendt in https://github.com/rapidsai/cudf/pull/20967 -* Fix racechecks reported in parquet gpuEncodePages kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20975 -* Don't encode s3 paths for kvikio_remote_io in read_json by @mroeschke in https://github.com/rapidsai/cudf/pull/20976 -* Allow sort merge join to go above int32 output row limits by @revans2 in https://github.com/rapidsai/cudf/pull/20960 -* Correct stream ordered deallocation in `Join` by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20981 -* Reintroduce `Buffer.nbytes` property by @pentschev in https://github.com/rapidsai/cudf/pull/21027 -* Fix SHA hash OOB on strings that are exact multiples of message chunk size by @rishic3 in https://github.com/rapidsai/cudf/pull/21004 -* Temporarily disable IWYU for nightly tests by @davidwendt in https://github.com/rapidsai/cudf/pull/21045 -* Fix cudf-polars multi-partition distributed sort by @TomAugspurger in https://github.com/rapidsai/cudf/pull/21047 -* Backport #21051 by @wence- in https://github.com/rapidsai/cudf/pull/21086 -* Pin pandas for `pylibcudf` testing by @galipremsagar in https://github.com/rapidsai/cudf/pull/21124 -* Hide pinned pool instantiation to avoid symbol conflicts with nvcomp by @vyasr in https://github.com/rapidsai/cudf/pull/21161 -* Specialize field type checking for bool in Parquet thrift list decoder by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21144 -* Fix reading of CSV files with double quotes in unquoted strings by @vuule in https://github.com/rapidsai/cudf/pull/21151 -* Revert the multithreaded optimization in the CSV reader by @vuule in https://github.com/rapidsai/cudf/pull/21198 -* Pin sqlglot in third-party integration tests by @Matt711 in https://github.com/rapidsai/cudf/pull/21271 -* Exclude sqlglot version 28.7 from CI by @Matt711 in https://github.com/rapidsai/cudf/pull/21293 +* Avoid duplicate streaming nodes for the rapidsmpf runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/20586 +* Handle scalar arguments in ternary expression by @Matt711 in https://github.com/NVIDIA/cudf/pull/20600 +* fix(noarch): use noarch build script in noarch build by @gforsyth in https://github.com/NVIDIA/cudf/pull/20654 +* fix(conda): matrix out noarch builds by cuda-major version by @gforsyth in https://github.com/NVIDIA/cudf/pull/20678 +* Include RMM in type checking environment and update type annotations for optional `stream` by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20636 +* Add no-op path for `ArrowExtensionArray.astype` by @Matt711 in https://github.com/NVIDIA/cudf/pull/20580 +* Skip pytorch integration tests if CUDA is not available by @Matt711 in https://github.com/NVIDIA/cudf/pull/20729 +* Always delay CUDA Array Interface pointer access by @vyasr in https://github.com/NVIDIA/cudf/pull/20719 +* Fix various copy-on-write bugs by @vyasr in https://github.com/NVIDIA/cudf/pull/20744 +* Fix leaks in cuDF java tests by @abellina in https://github.com/NVIDIA/cudf/pull/20767 +* Fix plc.Scalar.from_py(datetime.datetime) incorrectly localizing naive datetimes by @mroeschke in https://github.com/NVIDIA/cudf/pull/20769 +* Don't remove double casts in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/20773 +* Fixes struct column handling in sort-merge joins by @shrshi in https://github.com/NVIDIA/cudf/pull/20664 +* Fix for `synccheck` compute-sanitizer errors across Parquet gtest by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20775 +* Pin `numpy<2.4.0a0` in mypy pre-commit environment by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20781 +* Raise when trying to run queries on different devices in same process by @wence- in https://github.com/NVIDIA/cudf/pull/20617 +* Ensure `min_periods=0` is passed through rolling aggregations by @Matt711 in https://github.com/NVIDIA/cudf/pull/20653 +* Fix racecheck errors in the ORC reader by @vuule in https://github.com/NVIDIA/cudf/pull/20792 +* Fix the crash of multi-threaded parquet reader benchmark by @kingcrimsontianyu in https://github.com/NVIDIA/cudf/pull/20783 +* Fix racecheck reported by DATA_CHUNK_SOURCE_TEST in inflate_kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20804 +* Fix racecheck in the gpu_debrotli_kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20806 +* Ensure literal groupby aggregations are broadcasted to key length in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/20776 +* Pin `aiobotocore<3` to fix CI failures by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20844 +* Fix racecheck in parquet decode_page_data_generic kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20850 +* Avoid generating empty `TableChunks` in streaming scan nodes by @rjzamora in https://github.com/NVIDIA/cudf/pull/20815 +* Fix dask imports in ``CudfFusedParquetIOHost`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/20845 +* Fix UB due to OOM Exception in ParquetReaderTest.ManyLargeLists by @lamarrr in https://github.com/NVIDIA/cudf/pull/20841 +* Fix racecheck/synccheck in JSON parse_fn_string_parallel kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20856 +* Fix racecheck in ORC decode_column_data_kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20853 +* Disable flatbuffers tests in CMake configuration by @bdice in https://github.com/NVIDIA/cudf/pull/20848 +* Upper bound on aiosqlite in polars-upstream job by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20866 +* Fix boolean casting consistency with Pandas (#20746) by @aryansri05 in https://github.com/NVIDIA/cudf/pull/20747 +* Add retries to requests made to PyPI's JSON API by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20865 +* Fix `size_type` overflow in multiple APIs by @vuule in https://github.com/NVIDIA/cudf/pull/20857 +* Fix racecheck in parquet compute_string_page_bounds_kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20868 +* Fix dictionary::encode to honor indices-type parameter by @davidwendt in https://github.com/NVIDIA/cudf/pull/20842 +* Add missing headers to row_ir.hpp, row_ir.cpp by @bdice in https://github.com/NVIDIA/cudf/pull/20834 +* Fix `parquet_options` in pdsh benchmark by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20893 +* Add stream synchronize to tdigest generate_group_cluster_info by @davidwendt in https://github.com/NVIDIA/cudf/pull/20846 +* Only install RMM in mypy env on linux by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20878 +* Make nvcomp export unconditional by @vyasr in https://github.com/NVIDIA/cudf/pull/20828 +* Ensure we have nvjitlink from the CUDA version used at build time or newer and upgrade numba-cuda lower bound by @bdice in https://github.com/NVIDIA/cudf/pull/20873 +* Fix size_type overflow in the ORC writer by @vuule in https://github.com/NVIDIA/cudf/pull/20889 +* Constrain pyparsing version by @vyasr in https://github.com/NVIDIA/cudf/pull/20935 +* Revert #20902 by @vyasr in https://github.com/NVIDIA/cudf/pull/20955 +* Add force-blocking-launches to run_compute_sanitizer_test script by @davidwendt in https://github.com/NVIDIA/cudf/pull/20962 +* Fix racecheck error in parquet delta_byte_array_decoder::string_scan by @davidwendt in https://github.com/NVIDIA/cudf/pull/20967 +* Fix racechecks reported in parquet gpuEncodePages kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20975 +* Don't encode s3 paths for kvikio_remote_io in read_json by @mroeschke in https://github.com/NVIDIA/cudf/pull/20976 +* Allow sort merge join to go above int32 output row limits by @revans2 in https://github.com/NVIDIA/cudf/pull/20960 +* Correct stream ordered deallocation in `Join` by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20981 +* Reintroduce `Buffer.nbytes` property by @pentschev in https://github.com/NVIDIA/cudf/pull/21027 +* Fix SHA hash OOB on strings that are exact multiples of message chunk size by @rishic3 in https://github.com/NVIDIA/cudf/pull/21004 +* Temporarily disable IWYU for nightly tests by @davidwendt in https://github.com/NVIDIA/cudf/pull/21045 +* Fix cudf-polars multi-partition distributed sort by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/21047 +* Backport #21051 by @wence- in https://github.com/NVIDIA/cudf/pull/21086 +* Pin pandas for `pylibcudf` testing by @galipremsagar in https://github.com/NVIDIA/cudf/pull/21124 +* Hide pinned pool instantiation to avoid symbol conflicts with nvcomp by @vyasr in https://github.com/NVIDIA/cudf/pull/21161 +* Specialize field type checking for bool in Parquet thrift list decoder by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21144 +* Fix reading of CSV files with double quotes in unquoted strings by @vuule in https://github.com/NVIDIA/cudf/pull/21151 +* Revert the multithreaded optimization in the CSV reader by @vuule in https://github.com/NVIDIA/cudf/pull/21198 +* Pin sqlglot in third-party integration tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/21271 +* Exclude sqlglot version 28.7 from CI by @Matt711 in https://github.com/NVIDIA/cudf/pull/21293 ### 📖 Documentation -* Add note to developer guide about null values being undefined by @bdice in https://github.com/rapidsai/cudf/pull/20645 -* [DOC] Add cudf-polars to the example build command by @Matt711 in https://github.com/rapidsai/cudf/pull/20763 -* Clarify internal API header placement guidelines for details headers by @PointKernel in https://github.com/rapidsai/cudf/pull/20985 -* Clarify deprecation message for cudf::round by @nirandaperera in https://github.com/rapidsai/cudf/pull/20809 -* Require nvcc 12.9 in contributing guide by @bdice in https://github.com/rapidsai/cudf/pull/21186 +* Add note to developer guide about null values being undefined by @bdice in https://github.com/NVIDIA/cudf/pull/20645 +* [DOC] Add cudf-polars to the example build command by @Matt711 in https://github.com/NVIDIA/cudf/pull/20763 +* Clarify internal API header placement guidelines for details headers by @PointKernel in https://github.com/NVIDIA/cudf/pull/20985 +* Clarify deprecation message for cudf::round by @nirandaperera in https://github.com/NVIDIA/cudf/pull/20809 +* Require nvcc 12.9 in contributing guide by @bdice in https://github.com/NVIDIA/cudf/pull/21186 ### 🚀 New Features -* Expose `cudf::compute_column_jit` to python by @Matt711 in https://github.com/rapidsai/cudf/pull/20697 -* Add configuration option for max-io-threads by @quasiben in https://github.com/rapidsai/cudf/pull/20606 -* Return stats from `lower_ir_graph` by @rjzamora in https://github.com/rapidsai/cudf/pull/20528 -* Promote join_kind from detail namespace to public by @PointKernel in https://github.com/rapidsai/cudf/pull/20703 -* Make DataFrameScan and DataFrameSourceInfo pickle-able by @rjzamora in https://github.com/rapidsai/cudf/pull/20732 -* Add compute-sanitizer dispatch action by @bdice in https://github.com/rapidsai/cudf/pull/20542 -* Add RapidsMPF AllGather manager to cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/20731 -* Use metadata channel for the "rapidsmpf" runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/20738 -* Enable distributed execution with the "rapidsmpf" runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/20662 -* Filter row groups using byte range in the new experimental parquet reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20733 -* Make row hasher 64-bit hashing compatible by @PointKernel in https://github.com/rapidsai/cudf/pull/20777 -* Expose parquet JIT filter option to python by @Matt711 in https://github.com/rapidsai/cudf/pull/20790 -* Add filter_join_indices by @PointKernel in https://github.com/rapidsai/cudf/pull/20385 -* Add support for topk aggregation in libcudf groupby by @davidwendt in https://github.com/rapidsai/cudf/pull/20632 -* Allow parquet readers to use existing `datasource`s and `metadata`s by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20693 -* Reader and writer for a simple CudfTable format by @vuule in https://github.com/rapidsai/cudf/pull/20811 -* Add support for dictionary types in the row hasher by @PointKernel in https://github.com/rapidsai/cudf/pull/20989 -* Support left joins using sort-merge algorithm by @shrshi in https://github.com/rapidsai/cudf/pull/20787 -* Implement `batch_null_count` to count nulls for multiple null masks by a single kernel call, and application in groupby aggregations by @ttnghia in https://github.com/rapidsai/cudf/pull/20872 -* Support multiple roaring bitmap deletion vectors in parquet readers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20840 -* Add approx_distinct_count by @PointKernel in https://github.com/rapidsai/cudf/pull/20735 -* Pin Polars>=1.30,<1.36 by @Matt711 in https://github.com/rapidsai/cudf/pull/20791 -* Support `is_compressed` V2 flag in the Parquet writer by @vuule in https://github.com/rapidsai/cudf/pull/21050 -* Example to demonstrate intra-parquet-file pipelining using hybrid scan APIs by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20918 +* Expose `cudf::compute_column_jit` to python by @Matt711 in https://github.com/NVIDIA/cudf/pull/20697 +* Add configuration option for max-io-threads by @quasiben in https://github.com/NVIDIA/cudf/pull/20606 +* Return stats from `lower_ir_graph` by @rjzamora in https://github.com/NVIDIA/cudf/pull/20528 +* Promote join_kind from detail namespace to public by @PointKernel in https://github.com/NVIDIA/cudf/pull/20703 +* Make DataFrameScan and DataFrameSourceInfo pickle-able by @rjzamora in https://github.com/NVIDIA/cudf/pull/20732 +* Add compute-sanitizer dispatch action by @bdice in https://github.com/NVIDIA/cudf/pull/20542 +* Add RapidsMPF AllGather manager to cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/20731 +* Use metadata channel for the "rapidsmpf" runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/20738 +* Enable distributed execution with the "rapidsmpf" runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/20662 +* Filter row groups using byte range in the new experimental parquet reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20733 +* Make row hasher 64-bit hashing compatible by @PointKernel in https://github.com/NVIDIA/cudf/pull/20777 +* Expose parquet JIT filter option to python by @Matt711 in https://github.com/NVIDIA/cudf/pull/20790 +* Add filter_join_indices by @PointKernel in https://github.com/NVIDIA/cudf/pull/20385 +* Add support for topk aggregation in libcudf groupby by @davidwendt in https://github.com/NVIDIA/cudf/pull/20632 +* Allow parquet readers to use existing `datasource`s and `metadata`s by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20693 +* Reader and writer for a simple CudfTable format by @vuule in https://github.com/NVIDIA/cudf/pull/20811 +* Add support for dictionary types in the row hasher by @PointKernel in https://github.com/NVIDIA/cudf/pull/20989 +* Support left joins using sort-merge algorithm by @shrshi in https://github.com/NVIDIA/cudf/pull/20787 +* Implement `batch_null_count` to count nulls for multiple null masks by a single kernel call, and application in groupby aggregations by @ttnghia in https://github.com/NVIDIA/cudf/pull/20872 +* Support multiple roaring bitmap deletion vectors in parquet readers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20840 +* Add approx_distinct_count by @PointKernel in https://github.com/NVIDIA/cudf/pull/20735 +* Pin Polars>=1.30,<1.36 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20791 +* Support `is_compressed` V2 flag in the Parquet writer by @vuule in https://github.com/NVIDIA/cudf/pull/21050 +* Example to demonstrate intra-parquet-file pipelining using hybrid scan APIs by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20918 ### 🛠️ Improvements -* feat(conda): build noarch python packages separately by @gforsyth in https://github.com/rapidsai/cudf/pull/20613 -* Fix rapidsmpf dependency updates by @bdice in https://github.com/rapidsai/cudf/pull/20624 -* Print duckDB query plan and change Q17 join type by @Matt711 in https://github.com/rapidsai/cudf/pull/20615 -* Update RapidsMPF imports by @madsbk in https://github.com/rapidsai/cudf/pull/20665 -* Forward-merge release/25.12 into main by @bdice in https://github.com/rapidsai/cudf/pull/20676 -* Remove cudfjar install target by @vyasr in https://github.com/rapidsai/cudf/pull/20670 -* Use `RAPIDS_BRANCH` in cmake-format invocations that need rapids-cmake configs by @bdice in https://github.com/rapidsai/cudf/pull/20415 -* Merge release/25.12 into main by @vyasr in https://github.com/rapidsai/cudf/pull/20706 -* Use strict priority in CI conda tests by @bdice in https://github.com/rapidsai/cudf/pull/20690 -* Minor improvements to pylibcudf recipe by @bdice in https://github.com/rapidsai/cudf/pull/20684 -* Remove unnecessary nanoarrow fetch by @vyasr in https://github.com/rapidsai/cudf/pull/20669 -* Revert pytest pin by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20643 -* Use real row-group sample to estimate partition size by @rjzamora in https://github.com/rapidsai/cudf/pull/20567 -* Move rapidsmpf-specific testing in cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/20695 -* Include thrust::pair headers by @bdice in https://github.com/rapidsai/cudf/pull/20708 -* Remove sccache calls in noarch builds by @vyasr in https://github.com/rapidsai/cudf/pull/20710 -* Replace rmm::mr::get_current_device_resource() with cudf::get_current_device_resource_ref() by @davidwendt in https://github.com/rapidsai/cudf/pull/20694 -* Improved implementation for get_mask_offset_word utility by @davidwendt in https://github.com/rapidsai/cudf/pull/20622 -* Remove unneeded cudaMemcpy() calls by @davidwendt in https://github.com/rapidsai/cudf/pull/20618 -* Simplify broadcast-join algorithm in cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/20724 -* Add spilling support to staged fanout chunks by @rjzamora in https://github.com/rapidsai/cudf/pull/20642 -* Use rapidsmpf ShufflerAsync by @rjzamora in https://github.com/rapidsai/cudf/pull/20701 -* Move thrust::tuple usages to cuda::std::tuple by @davidwendt in https://github.com/rapidsai/cudf/pull/20717 -* Add job-specific timeouts to GHA test jobs by @bdice in https://github.com/rapidsai/cudf/pull/20730 -* Compatibility updates for CCCL 3.2 by @bdice in https://github.com/rapidsai/cudf/pull/20725 -* Move googlebench benchmarks to nvbench by @davidwendt in https://github.com/rapidsai/cudf/pull/20698 -* Enable blocking mechanism to avoid proxy object transfers in `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/19805 -* Remove googlebench dependency for libcudf by @davidwendt in https://github.com/rapidsai/cudf/pull/20739 -* Upgrade nanoarrow by @vyasr in https://github.com/rapidsai/cudf/pull/20711 -* Improve local pandas testing experience by @vyasr in https://github.com/rapidsai/cudf/pull/20753 -* Use .plc_column instead of .to_pylibcudf in IO methods by @mroeschke in https://github.com/rapidsai/cudf/pull/20742 -* Use .plc_column instead of .to_pylibcudf in indexing_utils, public objects by @mroeschke in https://github.com/rapidsai/cudf/pull/20758 -* Add back previously failing json test with stream by @vyasr in https://github.com/rapidsai/cudf/pull/19865 -* Add libcudf dictionary encode benchmark by @davidwendt in https://github.com/rapidsai/cudf/pull/20696 -* Remove unneeded aggregation kind_to_type utility and macro by @davidwendt in https://github.com/rapidsai/cudf/pull/20682 -* Test copy-on-write in CI by @vyasr in https://github.com/rapidsai/cudf/pull/20745 -* Stop using Dtype annotation more internally in cudf classic by @mroeschke in https://github.com/rapidsai/cudf/pull/20760 -* Parquet: Only fill in null values for string lengths and list offsets by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20671 -* Enable mypy's disallow_untyped_defs = true in cudf.core.column.* by @mroeschke in https://github.com/rapidsai/cudf/pull/20759 -* Improve groupby test utils to include the original location of failure by @ttnghia in https://github.com/rapidsai/cudf/pull/20718 -* use CUDA 13 for third-party integration tests by @jameslamb in https://github.com/rapidsai/cudf/pull/20748 -* Use strict priority in CI conda tests by @bdice in https://github.com/rapidsai/cudf/pull/20772 -* Upgrade to nvcomp 5.1.0.21 by @bdice in https://github.com/rapidsai/cudf/pull/20770 -* Use RapidsMPF's `reserve_device_memory_and_spill()` by @madsbk in https://github.com/rapidsai/cudf/pull/20778 -* avoid passing `start` as keyword argument to `np.arange` by @jorenham in https://github.com/rapidsai/cudf/pull/20788 -* Use env var to disable long tests when run with racecheck by @davidwendt in https://github.com/rapidsai/cudf/pull/20755 -* Improve performance for small string gather by @tgujar in https://github.com/rapidsai/cudf/pull/20656 -* Deprecate sort-merge join functional APIs by @shrshi in https://github.com/rapidsai/cudf/pull/20785 -* Partially revert broadcast-join change by @rjzamora in https://github.com/rapidsai/cudf/pull/20779 -* Type checking compatibility for numpy 2.4.0rc1 and other fixes by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20795 -* Support pl.Expr.cast(strict=False) in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/20784 -* chore(noarch): standardize noarch artifact naming by @gforsyth in https://github.com/rapidsai/cudf/pull/20794 -* Remove alpha specs from non-RAPIDS dependencies by @bdice in https://github.com/rapidsai/cudf/pull/20797 -* Enable merge barriers by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/20813 -* Update to numba-cuda `>=0.22.1,<0.23.0` by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/20750 -* Enable using multithreaded `setup_page_index` in hybrid scan reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20721 -* Remove size and offsets from Column by @vyasr in https://github.com/rapidsai/cudf/pull/20824 -* Add devcontainer fallback for C++ test location by @bdice in https://github.com/rapidsai/cudf/pull/20838 -* Add cudf-polars option to control rapidsmpf Shuffle insertion method by @TomAugspurger in https://github.com/rapidsai/cudf/pull/19634 -* Make null_count delegate to plc_column by @vyasr in https://github.com/rapidsai/cudf/pull/20854 -* Replace thrust reductions in Parquet reader with CUB + pinned memory based implementations by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20821 -* Reduce stream synchronization in `(mutable_)column_device_view::create()` and `(mutable_)table_device_view::create()` by @ttnghia in https://github.com/rapidsai/cudf/pull/20852 -* Clean up hash-based groupby aggregation, reducing overhead and memory usage by @ttnghia in https://github.com/rapidsai/cudf/pull/20658 -* Support decomposing Len expressions in cudf_polars streaming executor by @mroeschke in https://github.com/rapidsai/cudf/pull/20786 -* Add parameter to disable native `read_parquet` node by @rjzamora in https://github.com/rapidsai/cudf/pull/20858 -* Support arbitrary span-like data storage in pylibcudf Column by @vyasr in https://github.com/rapidsai/cudf/pull/20869 -* Merge ExposureTrackedBuffer into Buffer to simplify class hierarchy by @vyasr in https://github.com/rapidsai/cudf/pull/20874 -* Replace thrust logical functions with CUB + pinned memory based implementations in Parquet reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20822 -* Sync stream in host_memory.cpp by @bdice in https://github.com/rapidsai/cudf/pull/20687 -* Remove extra syncthreads() call from ORC DecodeRowPositions device function by @davidwendt in https://github.com/rapidsai/cudf/pull/20867 -* Temporarily increase max_days_without_success for nightly CI check by @bdice in https://github.com/rapidsai/cudf/pull/20880 -* Add zstd kernels to compute-sanitizer filter parameter by @davidwendt in https://github.com/rapidsai/cudf/pull/20875 -* Replace `thrust::reduce_by_key` with CUB + pinned memory based wrapper by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20860 -* cuml 26.2.0 compatibility by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20883 -* Implement pandas 3.0, backward compatible changes by @mroeschke in https://github.com/rapidsai/cudf/pull/20803 -* Improve column selection in the new experimental parquet reader by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20604 -* Fix some gtests to not assume dictionary keys order by @davidwendt in https://github.com/rapidsai/cudf/pull/20827 -* Parquet decode: Skip up to first_row for non-lists by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20835 -* Disable DeeplyNestedArithmeticLogicalExpression jit gtest for driver < 12.9 by @davidwendt in https://github.com/rapidsai/cudf/pull/20894 -* Make base_data and base_mask passthroughs by @vyasr in https://github.com/rapidsai/cudf/pull/20896 -* Changes needed for CCCL 3.2 compatibility by @bdice in https://github.com/rapidsai/cudf/pull/20810 -* Modify the default pinned pool to allow growth when the pool is exhausted by @vuule in https://github.com/rapidsai/cudf/pull/20839 -* Empty commit to trigger a build by @bdice in https://github.com/rapidsai/cudf/pull/20922 -* Fix clang-tidy errors by @vyasr in https://github.com/rapidsai/cudf/pull/20929 -* Replace thrust `count_if` and `copy_if` with CUB + pinned memory based wrappers by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20861 -* Parquet: Reuse string offset preprocessing when allocating output memory by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20902 -* Clean up includes for rmm::mr::polymorphic_allocator by @bdice in https://github.com/rapidsai/cudf/pull/20371 -* Convert to plc_column wherever possible by @vyasr in https://github.com/rapidsai/cudf/pull/20940 -* Push more arrow conversion logic down to pylibcudf by @vyasr in https://github.com/rapidsai/cudf/pull/20919 -* Simplify categorical column by @vyasr in https://github.com/rapidsai/cudf/pull/20942 -* Remove get_ptr from buffer owner classes by @vyasr in https://github.com/rapidsai/cudf/pull/20949 -* Fix null counts in mutating pylibcudf operations by @vyasr in https://github.com/rapidsai/cudf/pull/20950 -* Add context manager to control access mode by @vyasr in https://github.com/rapidsai/cudf/pull/20952 -* Convert column children computation from lazy to eager by @vyasr in https://github.com/rapidsai/cudf/pull/20953 -* Use SPDX license identifiers in pyproject.toml, bump build dependency floors by @jameslamb in https://github.com/rapidsai/cudf/pull/20959 -* Compatibility for cuML deprecation warnings by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20884 -* Use larger node for cpp-linters job in nightly tests by @vyasr in https://github.com/rapidsai/cudf/pull/20963 -* Fix min/max reduction logic for dictionary columns by @davidwendt in https://github.com/rapidsai/cudf/pull/20847 -* Remove null masks for intermediate results when computing compound hash-based groupby aggregations by @ttnghia in https://github.com/rapidsai/cudf/pull/20736 -* Fix warnings in dask-cudf test suite by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20951 -* Add CUDA 13.1 support by @bdice in https://github.com/rapidsai/cudf/pull/20870 -* Enable spill lock acquisition via context by @vyasr in https://github.com/rapidsai/cudf/pull/20964 -* Restore string preprocess PR and fix memcheck by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20969 -* Enable sccache-dist for cpp-linters by @vyasr in https://github.com/rapidsai/cudf/pull/20968 -* [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/rapidsai/cudf/pull/20971 -* Clean up mixed join common utilities by @PointKernel in https://github.com/rapidsai/cudf/pull/20836 -* Disable TRANSPOSE_TEST checking logic for CI racecheck runs by @davidwendt in https://github.com/rapidsai/cudf/pull/20970 -* Use nosync execution policy everywhere by @bdice in https://github.com/rapidsai/cudf/pull/20807 -* Remove `cuda.core.experimental` warnings filters by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/20933 -* Implement more flexible runtime to compile-time dispatching by @vyasr in https://github.com/rapidsai/cudf/pull/20927 -* Use per-column context in place of acquire_spill_lock by @vyasr in https://github.com/rapidsai/cudf/pull/20977 -* Fix cudf::clamp() for dictionary column types by @davidwendt in https://github.com/rapidsai/cudf/pull/20898 -* Patch installed pandas for cudf.pandas, pandas unit test run with CoW fix by @mroeschke in https://github.com/rapidsai/cudf/pull/20973 -* build and test against CUDA 13.1.0 by @jameslamb in https://github.com/rapidsai/cudf/pull/20972 -* Add ``opaque_reservation`` utility by @rjzamora in https://github.com/rapidsai/cudf/pull/20885 -* Remove exposure on column construction and unwrap buffers on pylibcudf conversion by @vyasr in https://github.com/rapidsai/cudf/pull/20980 -* Apply nosync execution policy in tests, benchmarks, Python, Java, and add docs by @bdice in https://github.com/rapidsai/cudf/pull/20978 -* Use `D` instead of `d` for time units by @galipremsagar in https://github.com/rapidsai/cudf/pull/20910 -* Add missing standard library headers to groupby/hash and jit files by @bdice in https://github.com/rapidsai/cudf/pull/20982 -* Add in key remapping for improved sort merge join performance by @revans2 in https://github.com/rapidsai/cudf/pull/20826 -* Use pinned memory in PQ reader to avoid pageable copies by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20820 -* Add Hybrid scan APIs for single-step table materialization by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20906 -* Add utility for deferring allocations on a stream by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20987 -* Remove CUDF_EXPORT from cudf::detail::contains by @davidwendt in https://github.com/rapidsai/cudf/pull/20991 -* Restrict objects that construct cuDF Python Buffer by @mroeschke in https://github.com/rapidsai/cudf/pull/20983 -* Fix min/max groupby logic for dictionary columns by @davidwendt in https://github.com/rapidsai/cudf/pull/20887 -* Centralize cudf Column creation as much as possible by @vyasr in https://github.com/rapidsai/cudf/pull/20999 -* Empty commit to trigger a build by @jameslamb in https://github.com/rapidsai/cudf/pull/21014 -* Rearrange variables to reduce padding by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/21016 -* Clean up buffer and access context implementations by @vyasr in https://github.com/rapidsai/cudf/pull/21013 -* Add missing thrust/tuple.h include for thrust::tie by @bdice in https://github.com/rapidsai/cudf/pull/21009 -* Replace remaining small pageable copies in PQ reader with pinned by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21006 -* Add dictionary specialization to row comparators by @davidwendt in https://github.com/rapidsai/cudf/pull/20830 -* Add no_sanitizer filter to compute-sanitizer script by @davidwendt in https://github.com/rapidsai/cudf/pull/20992 -* Make test_json_writer compatible with pandas 3 by @mroeschke in https://github.com/rapidsai/cudf/pull/21015 -* Use main shared-workflows branch by @jameslamb in https://github.com/rapidsai/cudf/pull/21038 -* Improve usage of polymorphism in columns by @vyasr in https://github.com/rapidsai/cudf/pull/21030 -* Increase memcheck timeout in nightly test script by @davidwendt in https://github.com/rapidsai/cudf/pull/21040 -* wheel builds: react to changes in pip's handling of build constraints by @mmccarty in https://github.com/rapidsai/cudf/pull/21048 -* Stop using non-pylibcudf children by @vyasr in https://github.com/rapidsai/cudf/pull/21057 -* Backport #21033: Add new pinned vector factory functions by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/21106 -* Use a multi-level host thread pool to avoid deadlocks by @vuule in https://github.com/rapidsai/cudf/pull/21075 -* fix(build): build package on merge to `release/*` branch by @gforsyth in https://github.com/rapidsai/cudf/pull/21181 -* Fallback to numba-cuda with no extra CUDA packages if 'cuda_suffixed' isn't true by @trxcllnt in https://github.com/rapidsai/cudf/pull/21185 +* feat(conda): build noarch python packages separately by @gforsyth in https://github.com/NVIDIA/cudf/pull/20613 +* Fix rapidsmpf dependency updates by @bdice in https://github.com/NVIDIA/cudf/pull/20624 +* Print duckDB query plan and change Q17 join type by @Matt711 in https://github.com/NVIDIA/cudf/pull/20615 +* Update RapidsMPF imports by @madsbk in https://github.com/NVIDIA/cudf/pull/20665 +* Forward-merge release/25.12 into main by @bdice in https://github.com/NVIDIA/cudf/pull/20676 +* Remove cudfjar install target by @vyasr in https://github.com/NVIDIA/cudf/pull/20670 +* Use `RAPIDS_BRANCH` in cmake-format invocations that need rapids-cmake configs by @bdice in https://github.com/NVIDIA/cudf/pull/20415 +* Merge release/25.12 into main by @vyasr in https://github.com/NVIDIA/cudf/pull/20706 +* Use strict priority in CI conda tests by @bdice in https://github.com/NVIDIA/cudf/pull/20690 +* Minor improvements to pylibcudf recipe by @bdice in https://github.com/NVIDIA/cudf/pull/20684 +* Remove unnecessary nanoarrow fetch by @vyasr in https://github.com/NVIDIA/cudf/pull/20669 +* Revert pytest pin by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20643 +* Use real row-group sample to estimate partition size by @rjzamora in https://github.com/NVIDIA/cudf/pull/20567 +* Move rapidsmpf-specific testing in cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/20695 +* Include thrust::pair headers by @bdice in https://github.com/NVIDIA/cudf/pull/20708 +* Remove sccache calls in noarch builds by @vyasr in https://github.com/NVIDIA/cudf/pull/20710 +* Replace rmm::mr::get_current_device_resource() with cudf::get_current_device_resource_ref() by @davidwendt in https://github.com/NVIDIA/cudf/pull/20694 +* Improved implementation for get_mask_offset_word utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/20622 +* Remove unneeded cudaMemcpy() calls by @davidwendt in https://github.com/NVIDIA/cudf/pull/20618 +* Simplify broadcast-join algorithm in cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/20724 +* Add spilling support to staged fanout chunks by @rjzamora in https://github.com/NVIDIA/cudf/pull/20642 +* Use rapidsmpf ShufflerAsync by @rjzamora in https://github.com/NVIDIA/cudf/pull/20701 +* Move thrust::tuple usages to cuda::std::tuple by @davidwendt in https://github.com/NVIDIA/cudf/pull/20717 +* Add job-specific timeouts to GHA test jobs by @bdice in https://github.com/NVIDIA/cudf/pull/20730 +* Compatibility updates for CCCL 3.2 by @bdice in https://github.com/NVIDIA/cudf/pull/20725 +* Move googlebench benchmarks to nvbench by @davidwendt in https://github.com/NVIDIA/cudf/pull/20698 +* Enable blocking mechanism to avoid proxy object transfers in `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/19805 +* Remove googlebench dependency for libcudf by @davidwendt in https://github.com/NVIDIA/cudf/pull/20739 +* Upgrade nanoarrow by @vyasr in https://github.com/NVIDIA/cudf/pull/20711 +* Improve local pandas testing experience by @vyasr in https://github.com/NVIDIA/cudf/pull/20753 +* Use .plc_column instead of .to_pylibcudf in IO methods by @mroeschke in https://github.com/NVIDIA/cudf/pull/20742 +* Use .plc_column instead of .to_pylibcudf in indexing_utils, public objects by @mroeschke in https://github.com/NVIDIA/cudf/pull/20758 +* Add back previously failing json test with stream by @vyasr in https://github.com/NVIDIA/cudf/pull/19865 +* Add libcudf dictionary encode benchmark by @davidwendt in https://github.com/NVIDIA/cudf/pull/20696 +* Remove unneeded aggregation kind_to_type utility and macro by @davidwendt in https://github.com/NVIDIA/cudf/pull/20682 +* Test copy-on-write in CI by @vyasr in https://github.com/NVIDIA/cudf/pull/20745 +* Stop using Dtype annotation more internally in cudf classic by @mroeschke in https://github.com/NVIDIA/cudf/pull/20760 +* Parquet: Only fill in null values for string lengths and list offsets by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20671 +* Enable mypy's disallow_untyped_defs = true in cudf.core.column.* by @mroeschke in https://github.com/NVIDIA/cudf/pull/20759 +* Improve groupby test utils to include the original location of failure by @ttnghia in https://github.com/NVIDIA/cudf/pull/20718 +* use CUDA 13 for third-party integration tests by @jameslamb in https://github.com/NVIDIA/cudf/pull/20748 +* Use strict priority in CI conda tests by @bdice in https://github.com/NVIDIA/cudf/pull/20772 +* Upgrade to nvcomp 5.1.0.21 by @bdice in https://github.com/NVIDIA/cudf/pull/20770 +* Use RapidsMPF's `reserve_device_memory_and_spill()` by @madsbk in https://github.com/NVIDIA/cudf/pull/20778 +* avoid passing `start` as keyword argument to `np.arange` by @jorenham in https://github.com/NVIDIA/cudf/pull/20788 +* Use env var to disable long tests when run with racecheck by @davidwendt in https://github.com/NVIDIA/cudf/pull/20755 +* Improve performance for small string gather by @tgujar in https://github.com/NVIDIA/cudf/pull/20656 +* Deprecate sort-merge join functional APIs by @shrshi in https://github.com/NVIDIA/cudf/pull/20785 +* Partially revert broadcast-join change by @rjzamora in https://github.com/NVIDIA/cudf/pull/20779 +* Type checking compatibility for numpy 2.4.0rc1 and other fixes by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20795 +* Support pl.Expr.cast(strict=False) in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/20784 +* chore(noarch): standardize noarch artifact naming by @gforsyth in https://github.com/NVIDIA/cudf/pull/20794 +* Remove alpha specs from non-RAPIDS dependencies by @bdice in https://github.com/NVIDIA/cudf/pull/20797 +* Enable merge barriers by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/20813 +* Update to numba-cuda `>=0.22.1,<0.23.0` by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/20750 +* Enable using multithreaded `setup_page_index` in hybrid scan reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20721 +* Remove size and offsets from Column by @vyasr in https://github.com/NVIDIA/cudf/pull/20824 +* Add devcontainer fallback for C++ test location by @bdice in https://github.com/NVIDIA/cudf/pull/20838 +* Add cudf-polars option to control rapidsmpf Shuffle insertion method by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/19634 +* Make null_count delegate to plc_column by @vyasr in https://github.com/NVIDIA/cudf/pull/20854 +* Replace thrust reductions in Parquet reader with CUB + pinned memory based implementations by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20821 +* Reduce stream synchronization in `(mutable_)column_device_view::create()` and `(mutable_)table_device_view::create()` by @ttnghia in https://github.com/NVIDIA/cudf/pull/20852 +* Clean up hash-based groupby aggregation, reducing overhead and memory usage by @ttnghia in https://github.com/NVIDIA/cudf/pull/20658 +* Support decomposing Len expressions in cudf_polars streaming executor by @mroeschke in https://github.com/NVIDIA/cudf/pull/20786 +* Add parameter to disable native `read_parquet` node by @rjzamora in https://github.com/NVIDIA/cudf/pull/20858 +* Support arbitrary span-like data storage in pylibcudf Column by @vyasr in https://github.com/NVIDIA/cudf/pull/20869 +* Merge ExposureTrackedBuffer into Buffer to simplify class hierarchy by @vyasr in https://github.com/NVIDIA/cudf/pull/20874 +* Replace thrust logical functions with CUB + pinned memory based implementations in Parquet reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20822 +* Sync stream in host_memory.cpp by @bdice in https://github.com/NVIDIA/cudf/pull/20687 +* Remove extra syncthreads() call from ORC DecodeRowPositions device function by @davidwendt in https://github.com/NVIDIA/cudf/pull/20867 +* Temporarily increase max_days_without_success for nightly CI check by @bdice in https://github.com/NVIDIA/cudf/pull/20880 +* Add zstd kernels to compute-sanitizer filter parameter by @davidwendt in https://github.com/NVIDIA/cudf/pull/20875 +* Replace `thrust::reduce_by_key` with CUB + pinned memory based wrapper by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20860 +* cuml 26.2.0 compatibility by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20883 +* Implement pandas 3.0, backward compatible changes by @mroeschke in https://github.com/NVIDIA/cudf/pull/20803 +* Improve column selection in the new experimental parquet reader by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20604 +* Fix some gtests to not assume dictionary keys order by @davidwendt in https://github.com/NVIDIA/cudf/pull/20827 +* Parquet decode: Skip up to first_row for non-lists by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20835 +* Disable DeeplyNestedArithmeticLogicalExpression jit gtest for driver < 12.9 by @davidwendt in https://github.com/NVIDIA/cudf/pull/20894 +* Make base_data and base_mask passthroughs by @vyasr in https://github.com/NVIDIA/cudf/pull/20896 +* Changes needed for CCCL 3.2 compatibility by @bdice in https://github.com/NVIDIA/cudf/pull/20810 +* Modify the default pinned pool to allow growth when the pool is exhausted by @vuule in https://github.com/NVIDIA/cudf/pull/20839 +* Empty commit to trigger a build by @bdice in https://github.com/NVIDIA/cudf/pull/20922 +* Fix clang-tidy errors by @vyasr in https://github.com/NVIDIA/cudf/pull/20929 +* Replace thrust `count_if` and `copy_if` with CUB + pinned memory based wrappers by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20861 +* Parquet: Reuse string offset preprocessing when allocating output memory by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20902 +* Clean up includes for rmm::mr::polymorphic_allocator by @bdice in https://github.com/NVIDIA/cudf/pull/20371 +* Convert to plc_column wherever possible by @vyasr in https://github.com/NVIDIA/cudf/pull/20940 +* Push more arrow conversion logic down to pylibcudf by @vyasr in https://github.com/NVIDIA/cudf/pull/20919 +* Simplify categorical column by @vyasr in https://github.com/NVIDIA/cudf/pull/20942 +* Remove get_ptr from buffer owner classes by @vyasr in https://github.com/NVIDIA/cudf/pull/20949 +* Fix null counts in mutating pylibcudf operations by @vyasr in https://github.com/NVIDIA/cudf/pull/20950 +* Add context manager to control access mode by @vyasr in https://github.com/NVIDIA/cudf/pull/20952 +* Convert column children computation from lazy to eager by @vyasr in https://github.com/NVIDIA/cudf/pull/20953 +* Use SPDX license identifiers in pyproject.toml, bump build dependency floors by @jameslamb in https://github.com/NVIDIA/cudf/pull/20959 +* Compatibility for cuML deprecation warnings by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20884 +* Use larger node for cpp-linters job in nightly tests by @vyasr in https://github.com/NVIDIA/cudf/pull/20963 +* Fix min/max reduction logic for dictionary columns by @davidwendt in https://github.com/NVIDIA/cudf/pull/20847 +* Remove null masks for intermediate results when computing compound hash-based groupby aggregations by @ttnghia in https://github.com/NVIDIA/cudf/pull/20736 +* Fix warnings in dask-cudf test suite by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20951 +* Add CUDA 13.1 support by @bdice in https://github.com/NVIDIA/cudf/pull/20870 +* Enable spill lock acquisition via context by @vyasr in https://github.com/NVIDIA/cudf/pull/20964 +* Restore string preprocess PR and fix memcheck by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20969 +* Enable sccache-dist for cpp-linters by @vyasr in https://github.com/NVIDIA/cudf/pull/20968 +* [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NVIDIA/cudf/pull/20971 +* Clean up mixed join common utilities by @PointKernel in https://github.com/NVIDIA/cudf/pull/20836 +* Disable TRANSPOSE_TEST checking logic for CI racecheck runs by @davidwendt in https://github.com/NVIDIA/cudf/pull/20970 +* Use nosync execution policy everywhere by @bdice in https://github.com/NVIDIA/cudf/pull/20807 +* Remove `cuda.core.experimental` warnings filters by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/20933 +* Implement more flexible runtime to compile-time dispatching by @vyasr in https://github.com/NVIDIA/cudf/pull/20927 +* Use per-column context in place of acquire_spill_lock by @vyasr in https://github.com/NVIDIA/cudf/pull/20977 +* Fix cudf::clamp() for dictionary column types by @davidwendt in https://github.com/NVIDIA/cudf/pull/20898 +* Patch installed pandas for cudf.pandas, pandas unit test run with CoW fix by @mroeschke in https://github.com/NVIDIA/cudf/pull/20973 +* build and test against CUDA 13.1.0 by @jameslamb in https://github.com/NVIDIA/cudf/pull/20972 +* Add ``opaque_reservation`` utility by @rjzamora in https://github.com/NVIDIA/cudf/pull/20885 +* Remove exposure on column construction and unwrap buffers on pylibcudf conversion by @vyasr in https://github.com/NVIDIA/cudf/pull/20980 +* Apply nosync execution policy in tests, benchmarks, Python, Java, and add docs by @bdice in https://github.com/NVIDIA/cudf/pull/20978 +* Use `D` instead of `d` for time units by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20910 +* Add missing standard library headers to groupby/hash and jit files by @bdice in https://github.com/NVIDIA/cudf/pull/20982 +* Add in key remapping for improved sort merge join performance by @revans2 in https://github.com/NVIDIA/cudf/pull/20826 +* Use pinned memory in PQ reader to avoid pageable copies by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20820 +* Add Hybrid scan APIs for single-step table materialization by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20906 +* Add utility for deferring allocations on a stream by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20987 +* Remove CUDF_EXPORT from cudf::detail::contains by @davidwendt in https://github.com/NVIDIA/cudf/pull/20991 +* Restrict objects that construct cuDF Python Buffer by @mroeschke in https://github.com/NVIDIA/cudf/pull/20983 +* Fix min/max groupby logic for dictionary columns by @davidwendt in https://github.com/NVIDIA/cudf/pull/20887 +* Centralize cudf Column creation as much as possible by @vyasr in https://github.com/NVIDIA/cudf/pull/20999 +* Empty commit to trigger a build by @jameslamb in https://github.com/NVIDIA/cudf/pull/21014 +* Rearrange variables to reduce padding by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/21016 +* Clean up buffer and access context implementations by @vyasr in https://github.com/NVIDIA/cudf/pull/21013 +* Add missing thrust/tuple.h include for thrust::tie by @bdice in https://github.com/NVIDIA/cudf/pull/21009 +* Replace remaining small pageable copies in PQ reader with pinned by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21006 +* Add dictionary specialization to row comparators by @davidwendt in https://github.com/NVIDIA/cudf/pull/20830 +* Add no_sanitizer filter to compute-sanitizer script by @davidwendt in https://github.com/NVIDIA/cudf/pull/20992 +* Make test_json_writer compatible with pandas 3 by @mroeschke in https://github.com/NVIDIA/cudf/pull/21015 +* Use main shared-workflows branch by @jameslamb in https://github.com/NVIDIA/cudf/pull/21038 +* Improve usage of polymorphism in columns by @vyasr in https://github.com/NVIDIA/cudf/pull/21030 +* Increase memcheck timeout in nightly test script by @davidwendt in https://github.com/NVIDIA/cudf/pull/21040 +* wheel builds: react to changes in pip's handling of build constraints by @mmccarty in https://github.com/NVIDIA/cudf/pull/21048 +* Stop using non-pylibcudf children by @vyasr in https://github.com/NVIDIA/cudf/pull/21057 +* Backport #21033: Add new pinned vector factory functions by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/21106 +* Use a multi-level host thread pool to avoid deadlocks by @vuule in https://github.com/NVIDIA/cudf/pull/21075 +* fix(build): build package on merge to `release/*` branch by @gforsyth in https://github.com/NVIDIA/cudf/pull/21181 +* Fallback to numba-cuda with no extra CUDA packages if 'cuda_suffixed' isn't true by @trxcllnt in https://github.com/NVIDIA/cudf/pull/21185 ## New Contributors -* @jorenham made their first contribution in https://github.com/rapidsai/cudf/pull/20788 -* @nirandaperera made their first contribution in https://github.com/rapidsai/cudf/pull/20809 -* @rishic3 made their first contribution in https://github.com/rapidsai/cudf/pull/21004 +* @jorenham made their first contribution in https://github.com/NVIDIA/cudf/pull/20788 +* @nirandaperera made their first contribution in https://github.com/NVIDIA/cudf/pull/20809 +* @rishic3 made their first contribution in https://github.com/NVIDIA/cudf/pull/21004 -**Full Changelog**: https://github.com/rapidsai/cudf/compare/v26.02.00a...release/26.02 +**Full Changelog**: https://github.com/NVIDIA/cudf/compare/v26.02.00a...release/26.02 # cudf 25.12.00 (10 Dec 2025) ### 🚨 Breaking Changes -* Rewrite JNI functions to use `JNI_TRY`/`JNI_CATCH` by @ttnghia in https://github.com/rapidsai/cudf/pull/19053 -* Remove compatibility with nvCOMP versions before 5.0 by @vuule in https://github.com/rapidsai/cudf/pull/20140 -* Remove DataFrame.apply_chunks, Groupby.apply_grouped by @mroeschke in https://github.com/rapidsai/cudf/pull/20194 -* Change .str.starts/endswith with tuple argument to match any pattern instead of pairwise matching by @mroeschke in https://github.com/rapidsai/cudf/pull/20249 -* [cudf-polars] CUDA stream by @madsbk in https://github.com/rapidsai/cudf/pull/20154 -* Chunked read parquet, prepend index column, and apply deletion vector by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20201 -* Zero-copy `hostdevice_vector` on integrated systems by @vuule in https://github.com/rapidsai/cudf/pull/20225 -* Use int64_t for the num_rows slot in parquet_reader_options by @wence- in https://github.com/rapidsai/cudf/pull/20256 -* Require CUDA 12.2+ by @jakirkham in https://github.com/rapidsai/cudf/pull/20416 -* Remove compatibility for CCCL < 3.1 by @bdice in https://github.com/rapidsai/cudf/pull/20468 -* Remove deprecated types and APIs by @vuule in https://github.com/rapidsai/cudf/pull/20422 -* Support signed integers and decimals in `SUM_WITH_OVERFLOW` groupby by @PointKernel in https://github.com/rapidsai/cudf/pull/19598 -* Change groupby-scan COUNT to 1-based results by @davidwendt in https://github.com/rapidsai/cudf/pull/20168 -* Change strings::like() pattern parameter from string_scalar to string_view by @davidwendt in https://github.com/rapidsai/cudf/pull/20428 -* No-op performance tracking wrappers by @galipremsagar in https://github.com/rapidsai/cudf/pull/20595 +* Rewrite JNI functions to use `JNI_TRY`/`JNI_CATCH` by @ttnghia in https://github.com/NVIDIA/cudf/pull/19053 +* Remove compatibility with nvCOMP versions before 5.0 by @vuule in https://github.com/NVIDIA/cudf/pull/20140 +* Remove DataFrame.apply_chunks, Groupby.apply_grouped by @mroeschke in https://github.com/NVIDIA/cudf/pull/20194 +* Change .str.starts/endswith with tuple argument to match any pattern instead of pairwise matching by @mroeschke in https://github.com/NVIDIA/cudf/pull/20249 +* [cudf-polars] CUDA stream by @madsbk in https://github.com/NVIDIA/cudf/pull/20154 +* Chunked read parquet, prepend index column, and apply deletion vector by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20201 +* Zero-copy `hostdevice_vector` on integrated systems by @vuule in https://github.com/NVIDIA/cudf/pull/20225 +* Use int64_t for the num_rows slot in parquet_reader_options by @wence- in https://github.com/NVIDIA/cudf/pull/20256 +* Require CUDA 12.2+ by @jakirkham in https://github.com/NVIDIA/cudf/pull/20416 +* Remove compatibility for CCCL < 3.1 by @bdice in https://github.com/NVIDIA/cudf/pull/20468 +* Remove deprecated types and APIs by @vuule in https://github.com/NVIDIA/cudf/pull/20422 +* Support signed integers and decimals in `SUM_WITH_OVERFLOW` groupby by @PointKernel in https://github.com/NVIDIA/cudf/pull/19598 +* Change groupby-scan COUNT to 1-based results by @davidwendt in https://github.com/NVIDIA/cudf/pull/20168 +* Change strings::like() pattern parameter from string_scalar to string_view by @davidwendt in https://github.com/NVIDIA/cudf/pull/20428 +* No-op performance tracking wrappers by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20595 ### 🐛 Bug Fixes -* Copy `attrs` at correct place in `DataFrame` constructor by @galipremsagar in https://github.com/rapidsai/cudf/pull/20074 -* Handle missing nightly runs in pandas tests job by @galipremsagar in https://github.com/rapidsai/cudf/pull/20081 -* Fix numpy ufunc for `DataFrame` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20070 -* Unproxy few unnecessary testing utilities in pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/20088 -* Fix libcudf groupby benchmarks to not include internal cache by @davidwendt in https://github.com/rapidsai/cudf/pull/20038 -* Fix cudf.date_range with non-iso start and end date strings by @mroeschke in https://github.com/rapidsai/cudf/pull/20116 -* Fix create_distinct_rows_column to create non-nullable columns by @davidwendt in https://github.com/rapidsai/cudf/pull/20082 -* Fix arrow timestamp frequency cases in `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20128 -* Cast inputs to true division from decimal to float by @Matt711 in https://github.com/rapidsai/cudf/pull/20077 -* Handle NVMLError_NotSupported in cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20179 -* Fix RMM JNI pinned_fallback_host_memory_resource for CCCL 3.1.0 by @bdice in https://github.com/rapidsai/cudf/pull/20160 -* Require passing memory resources to from_libcudf methods by @vyasr in https://github.com/rapidsai/cudf/pull/20171 -* Enable hash-groupby for decimal32/64 type and MEAN aggregation by @davidwendt in https://github.com/rapidsai/cudf/pull/20040 -* Align decimal dtypes in predicate before conditional join by @Matt711 in https://github.com/rapidsai/cudf/pull/20060 -* Change stream_checking_resource_adaptor::do_deallocate to noexcept by @vyasr in https://github.com/rapidsai/cudf/pull/20218 -* Deallocation should be noexcept by @bdice in https://github.com/rapidsai/cudf/pull/20219 -* Fix a race condition in the decode of delta encoded Parquet columns by @vuule in https://github.com/rapidsai/cudf/pull/20216 -* Fix the host-device tdigest offsets by using cuda::std::span by @PointKernel in https://github.com/rapidsai/cudf/pull/20220 -* Add `stream` and `mr` arguments to `Column.from_arrow` type stub by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20244 -* Pin `deltalake` in cudf-polars-polars-tests CI job by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20255 -* Pin ibis-framework<11.0.0 by @Matt711 in https://github.com/rapidsai/cudf/pull/20267 -* Add private attributes for `cudf.pandas` proxy objects by @galipremsagar in https://github.com/rapidsai/cudf/pull/20276 -* Add Proxy for `SparseAccessor` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20278 -* We need this to pacify mypy by @wence- in https://github.com/rapidsai/cudf/pull/20285 -* Purge non-empty nulls for the generated lists columns in data generation utility by @ttnghia in https://github.com/rapidsai/cudf/pull/20283 -* Fix missing table compatibility check in two_table_comparator constructor by @PointKernel in https://github.com/rapidsai/cudf/pull/20305 -* Fix the check for equal `num_cols` across empty parquet sources by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20320 -* Add `nans_to_nulls` to `Frame` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20314 -* Add support for list type in `get` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20332 -* Fix decimal dtype serialization in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/20300 -* Make the `GroupedRollingWindow`expression node reconstructable in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/20288 -* Ensure pylibcudf.Scalar.from_py uses CUDA streams by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20340 -* Skip failing cudf-polars test due to hash groupby bug by @Matt711 in https://github.com/rapidsai/cudf/pull/20356 -* Support order by keys for order-sensitive scalar aggregations in grouped windows by @Matt711 in https://github.com/rapidsai/cudf/pull/20350 -* Honor user-passed stream in slice_strings for scalar inputs by @mroeschke in https://github.com/rapidsai/cudf/pull/20349 -* Thread missing streams in column/table view creation to char size calculation by @vyasr in https://github.com/rapidsai/cudf/pull/20351 -* Fix missed-sync for `mapping_indices_kernel` in hash-based groupby aggregation by @ttnghia in https://github.com/rapidsai/cudf/pull/20370 -* Fix a few SPDX-related issues by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/20364 -* Fix a `dtype` bug in column constructor by @galipremsagar in https://github.com/rapidsai/cudf/pull/20384 -* Refactor `as_column` dtype parameter calls by @galipremsagar in https://github.com/rapidsai/cudf/pull/20379 -* Add CUDA stream to `cudf_polars.Column.deserialize` by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20396 -* Add missing CUDA stream to cudf-polars left-semi join by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20398 -* Fix various string APIs to work with extension types by @galipremsagar in https://github.com/rapidsai/cudf/pull/20368 -* Add parameter validation for `merge` and `MultiIndex.from_frame` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20382 -* Fix nvtext::normalize_characters special token case by @davidwendt in https://github.com/rapidsai/cudf/pull/20242 -* Fix pinned memory resource `shared_pointer` lifetime in tests. by @bdice in https://github.com/rapidsai/cudf/pull/20407 -* Support new `nvcompStatus_t` enum value by @vuule in https://github.com/rapidsai/cudf/pull/20376 -* Don't skip blank CSV lines rows after the header in cudf-polars scan_csv by @mroeschke in https://github.com/rapidsai/cudf/pull/20341 -* Fix OOB accesses in JSON_CornerCase_Empty test and get_row_array_parent_col_id function by @bdice in https://github.com/rapidsai/cudf/pull/20421 -* Change calls to cudaMemcpyToSymbol to cudaMemcpyToSymbolAsync by @davidwendt in https://github.com/rapidsai/cudf/pull/20374 -* Do not accelerate `pandas._config.config` by @Matt711 in https://github.com/rapidsai/cudf/pull/20413 -* Return timedelta instead of datetime type with std with datetime type with missing values by @mroeschke in https://github.com/rapidsai/cudf/pull/20439 -* Disallow non-bool skipna arguments to reduction methods by @mroeschke in https://github.com/rapidsai/cudf/pull/20436 -* Fix parquet scans for duckDB PDS-DS by @Matt711 in https://github.com/rapidsai/cudf/pull/20388 -* Support `__array_function__` on the proxy array type by @Matt711 in https://github.com/rapidsai/cudf/pull/20419 -* Make `memory_usage` and `__sizeof__` proxy attributes and always skip all memory usage tests by @Matt711 in https://github.com/rapidsai/cudf/pull/20425 -* Add input validation for `from_records` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20412 -* Use computed reduction result type for empty sum and product aggregations by @mroeschke in https://github.com/rapidsai/cudf/pull/20438 -* Correct level arg validation for Index.isin, unique by @mroeschke in https://github.com/rapidsai/cudf/pull/20449 -* Add private `_grouper` attribute to `DataFrameGroupBy` proxy type by @Matt711 in https://github.com/rapidsai/cudf/pull/20448 -* Raise ValueError when indexing with zero step slice by @mroeschke in https://github.com/rapidsai/cudf/pull/20453 -* Raise IndexError for float-like indexers in RangeIndex/MultiIndex.__getitem__ by @mroeschke in https://github.com/rapidsai/cudf/pull/20454 -* Disallow slice(bool, ...) in DataFrame.loc with MultiIndex by @mroeschke in https://github.com/rapidsai/cudf/pull/20457 -* Fix core dump in MemoryCleaner by @res-life in https://github.com/rapidsai/cudf/pull/19872 -* Disallow multiple ellipse values in loc/iloc indexing by @mroeschke in https://github.com/rapidsai/cudf/pull/20456 -* Fix `scan` operations for `string` columns by @galipremsagar in https://github.com/rapidsai/cudf/pull/20460 -* Fix UTF8 data generator in libcudf benchmarks utility by @davidwendt in https://github.com/rapidsai/cudf/pull/20465 -* Handle dealloc in stream-ordered cudf-polars ops by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20467 -* Raise on unsupported unstack cases by @Matt711 in https://github.com/rapidsai/cudf/pull/20463 -* Allow early exit for left semi-/anti- joins with empty build/probe tables by @shrshi in https://github.com/rapidsai/cudf/pull/20452 -* Fix OOB memory access in JSON reader ingest_raw utility by @davidwendt in https://github.com/rapidsai/cudf/pull/20451 -* Round up small-type groupby outputs to 4-byte boundary by @PointKernel in https://github.com/rapidsai/cudf/pull/20455 -* Fix GPU acceleration bug in decimal type-cast by @galipremsagar in https://github.com/rapidsai/cudf/pull/20471 -* Add missing CUDA stream in cudf_polars Distinct by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20477 -* Support `__arrow_array__` on proxy extension array by @Matt711 in https://github.com/rapidsai/cudf/pull/20478 -* Enable scan operation for `datetime64` and `timedelta64` types by @galipremsagar in https://github.com/rapidsai/cudf/pull/20464 -* Remove unneeded type check in cudf::strings::slice_strings by @davidwendt in https://github.com/rapidsai/cudf/pull/20437 -* Fix join match context tests by @PointKernel in https://github.com/rapidsai/cudf/pull/20472 -* Fix the statistics_mr in benchmark fixture by @PointKernel in https://github.com/rapidsai/cudf/pull/20496 -* Guard `__sizeof__` in pandas compatability mode by @Matt711 in https://github.com/rapidsai/cudf/pull/20495 -* Fix OOB memory access in Orc and Parquet stacks from fixed-width unaligned loads by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20458 -* Fix cudf.pandas Timestamp/Timedelta not subclassing stdlib datetime objects by @mroeschke in https://github.com/rapidsai/cudf/pull/20433 -* Revert benchmark input generation logic for list type by @davidwendt in https://github.com/rapidsai/cudf/pull/20498 -* Avoid using pylibcudf directly in rapidsmpf runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/20501 -* Suppress NVRTC arch warnings by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/20517 -* Fix ``ChannelManager`` and ``Lineariser`` by @rjzamora in https://github.com/rapidsai/cudf/pull/20516 -* Synchronize streams in ``LocalShuffle`` by @rjzamora in https://github.com/rapidsai/cudf/pull/20515 -* Make `argsort` have return type `np.intp` to match pandas by @Matt711 in https://github.com/rapidsai/cudf/pull/20487 -* Fix `polars.concat_str` with one column in cudf_polars by @mroeschke in https://github.com/rapidsai/cudf/pull/20535 -* Override `__sizeof__` for `cudf.Index` by @Matt711 in https://github.com/rapidsai/cudf/pull/20530 -* Fix `pl.scan_csv(...).slice(...).collect(engine="gpu")` with None endpoint by @mroeschke in https://github.com/rapidsai/cudf/pull/20519 -* Fix DataChunkSourceTest by syncing default stream by @davidwendt in https://github.com/rapidsai/cudf/pull/20492 -* Fix data size errors in some libcudf benchmarks by @davidwendt in https://github.com/rapidsai/cudf/pull/20512 -* Pin cython and pytest dependencies by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20571 -* Pin Cython pre-3.2.0 and PyTest pre-9 by @jakirkham in https://github.com/rapidsai/cudf/pull/20573 -* Handle `Empty` child IRs in `_decompose` by @Matt711 in https://github.com/rapidsai/cudf/pull/20409 -* Skip flaky pandas datetime test by @Matt711 in https://github.com/rapidsai/cudf/pull/20585 -* Fix max-pool-size-exceeded error in DATA_CHUNK_SOURCE_TEST by @davidwendt in https://github.com/rapidsai/cudf/pull/20534 -* Fix racecheck in nvtext wordpiece tokenizer kernel by @davidwendt in https://github.com/rapidsai/cudf/pull/20588 -* Fix the check to determine if all column chunk pages are dict encoded by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20524 -* Add stream synchronize to QUANTILES_TEST PercentileApprox gtests by @davidwendt in https://github.com/rapidsai/cudf/pull/20558 -* updated update-version.sh to handle release branch version changes by @rockhowse in https://github.com/rapidsai/cudf/pull/20598 -* Fix nvtext tokenizers handling invalid UTF8 data by @davidwendt in https://github.com/rapidsai/cudf/pull/20514 -* Fix overflow errors in distinct and filtered joins when hash table size exceeds int32 limits by @shrshi in https://github.com/rapidsai/cudf/pull/20594 -* [FEA] Optimize JIT Filter for Low-Selectivity by @lamarrr in https://github.com/rapidsai/cudf/pull/20222 -* Compute boolean function(NOT) on integers as a bitwise invert by @Matt711 in https://github.com/rapidsai/cudf/pull/20599 -* Cast output dtype of rolling aggregations to match pandas by @Matt711 in https://github.com/rapidsai/cudf/pull/20526 -* Add noop path for `Frame.astype` by @Matt711 in https://github.com/rapidsai/cudf/pull/20581 -* Fix `copy` semantics bugs thus reduce copies and memory usage by @galipremsagar in https://github.com/rapidsai/cudf/pull/20121 -* Ensure the sum after expression decomposition for mean has float output dtype by @Matt711 in https://github.com/rapidsai/cudf/pull/20596 -* Use `Decimal(0)` literal for all-null decimal groups in groupby-sum by @Matt711 in https://github.com/rapidsai/cudf/pull/20591 -* Do not drop `freq` when constructing `DatetimeIndex` from pandas by @brandon-b-miller in https://github.com/rapidsai/cudf/pull/18778 -* Fix --validation flag for cudf.pandas PDSH benchmarks by @mroeschke in https://github.com/rapidsai/cudf/pull/20540 -* Enable GPU acceleration for more binops by @galipremsagar in https://github.com/rapidsai/cudf/pull/20507 -* Fix `rmm` function calls due to removed deprecated APIs and macro by @ttnghia in https://github.com/rapidsai/cudf/pull/20661 -* Fix orc reader bool bug due to not being able to resume rle decode by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20666 -* Fix categorical comparisons in `cudf` to match `pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20674 -* Fix `any` and `all` to match pandas by @galipremsagar in https://github.com/rapidsai/cudf/pull/20679 -* Fix return types of string APIs in `cudf.pandas` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20683 -* Resolve pandas test failures by @galipremsagar in https://github.com/rapidsai/cudf/pull/20704 -* Fix DatetimeIndex pickling by @vyasr in https://github.com/rapidsai/cudf/pull/20709 -* `DatetimeIndex.serialize()` headers are msgpack serializable by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20714 +* Copy `attrs` at correct place in `DataFrame` constructor by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20074 +* Handle missing nightly runs in pandas tests job by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20081 +* Fix numpy ufunc for `DataFrame` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20070 +* Unproxy few unnecessary testing utilities in pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20088 +* Fix libcudf groupby benchmarks to not include internal cache by @davidwendt in https://github.com/NVIDIA/cudf/pull/20038 +* Fix cudf.date_range with non-iso start and end date strings by @mroeschke in https://github.com/NVIDIA/cudf/pull/20116 +* Fix create_distinct_rows_column to create non-nullable columns by @davidwendt in https://github.com/NVIDIA/cudf/pull/20082 +* Fix arrow timestamp frequency cases in `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20128 +* Cast inputs to true division from decimal to float by @Matt711 in https://github.com/NVIDIA/cudf/pull/20077 +* Handle NVMLError_NotSupported in cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20179 +* Fix RMM JNI pinned_fallback_host_memory_resource for CCCL 3.1.0 by @bdice in https://github.com/NVIDIA/cudf/pull/20160 +* Require passing memory resources to from_libcudf methods by @vyasr in https://github.com/NVIDIA/cudf/pull/20171 +* Enable hash-groupby for decimal32/64 type and MEAN aggregation by @davidwendt in https://github.com/NVIDIA/cudf/pull/20040 +* Align decimal dtypes in predicate before conditional join by @Matt711 in https://github.com/NVIDIA/cudf/pull/20060 +* Change stream_checking_resource_adaptor::do_deallocate to noexcept by @vyasr in https://github.com/NVIDIA/cudf/pull/20218 +* Deallocation should be noexcept by @bdice in https://github.com/NVIDIA/cudf/pull/20219 +* Fix a race condition in the decode of delta encoded Parquet columns by @vuule in https://github.com/NVIDIA/cudf/pull/20216 +* Fix the host-device tdigest offsets by using cuda::std::span by @PointKernel in https://github.com/NVIDIA/cudf/pull/20220 +* Add `stream` and `mr` arguments to `Column.from_arrow` type stub by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20244 +* Pin `deltalake` in cudf-polars-polars-tests CI job by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20255 +* Pin ibis-framework<11.0.0 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20267 +* Add private attributes for `cudf.pandas` proxy objects by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20276 +* Add Proxy for `SparseAccessor` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20278 +* We need this to pacify mypy by @wence- in https://github.com/NVIDIA/cudf/pull/20285 +* Purge non-empty nulls for the generated lists columns in data generation utility by @ttnghia in https://github.com/NVIDIA/cudf/pull/20283 +* Fix missing table compatibility check in two_table_comparator constructor by @PointKernel in https://github.com/NVIDIA/cudf/pull/20305 +* Fix the check for equal `num_cols` across empty parquet sources by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20320 +* Add `nans_to_nulls` to `Frame` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20314 +* Add support for list type in `get` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20332 +* Fix decimal dtype serialization in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/20300 +* Make the `GroupedRollingWindow`expression node reconstructable in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/20288 +* Ensure pylibcudf.Scalar.from_py uses CUDA streams by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20340 +* Skip failing cudf-polars test due to hash groupby bug by @Matt711 in https://github.com/NVIDIA/cudf/pull/20356 +* Support order by keys for order-sensitive scalar aggregations in grouped windows by @Matt711 in https://github.com/NVIDIA/cudf/pull/20350 +* Honor user-passed stream in slice_strings for scalar inputs by @mroeschke in https://github.com/NVIDIA/cudf/pull/20349 +* Thread missing streams in column/table view creation to char size calculation by @vyasr in https://github.com/NVIDIA/cudf/pull/20351 +* Fix missed-sync for `mapping_indices_kernel` in hash-based groupby aggregation by @ttnghia in https://github.com/NVIDIA/cudf/pull/20370 +* Fix a few SPDX-related issues by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/20364 +* Fix a `dtype` bug in column constructor by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20384 +* Refactor `as_column` dtype parameter calls by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20379 +* Add CUDA stream to `cudf_polars.Column.deserialize` by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20396 +* Add missing CUDA stream to cudf-polars left-semi join by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20398 +* Fix various string APIs to work with extension types by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20368 +* Add parameter validation for `merge` and `MultiIndex.from_frame` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20382 +* Fix nvtext::normalize_characters special token case by @davidwendt in https://github.com/NVIDIA/cudf/pull/20242 +* Fix pinned memory resource `shared_pointer` lifetime in tests. by @bdice in https://github.com/NVIDIA/cudf/pull/20407 +* Support new `nvcompStatus_t` enum value by @vuule in https://github.com/NVIDIA/cudf/pull/20376 +* Don't skip blank CSV lines rows after the header in cudf-polars scan_csv by @mroeschke in https://github.com/NVIDIA/cudf/pull/20341 +* Fix OOB accesses in JSON_CornerCase_Empty test and get_row_array_parent_col_id function by @bdice in https://github.com/NVIDIA/cudf/pull/20421 +* Change calls to cudaMemcpyToSymbol to cudaMemcpyToSymbolAsync by @davidwendt in https://github.com/NVIDIA/cudf/pull/20374 +* Do not accelerate `pandas._config.config` by @Matt711 in https://github.com/NVIDIA/cudf/pull/20413 +* Return timedelta instead of datetime type with std with datetime type with missing values by @mroeschke in https://github.com/NVIDIA/cudf/pull/20439 +* Disallow non-bool skipna arguments to reduction methods by @mroeschke in https://github.com/NVIDIA/cudf/pull/20436 +* Fix parquet scans for duckDB PDS-DS by @Matt711 in https://github.com/NVIDIA/cudf/pull/20388 +* Support `__array_function__` on the proxy array type by @Matt711 in https://github.com/NVIDIA/cudf/pull/20419 +* Make `memory_usage` and `__sizeof__` proxy attributes and always skip all memory usage tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/20425 +* Add input validation for `from_records` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20412 +* Use computed reduction result type for empty sum and product aggregations by @mroeschke in https://github.com/NVIDIA/cudf/pull/20438 +* Correct level arg validation for Index.isin, unique by @mroeschke in https://github.com/NVIDIA/cudf/pull/20449 +* Add private `_grouper` attribute to `DataFrameGroupBy` proxy type by @Matt711 in https://github.com/NVIDIA/cudf/pull/20448 +* Raise ValueError when indexing with zero step slice by @mroeschke in https://github.com/NVIDIA/cudf/pull/20453 +* Raise IndexError for float-like indexers in RangeIndex/MultiIndex.__getitem__ by @mroeschke in https://github.com/NVIDIA/cudf/pull/20454 +* Disallow slice(bool, ...) in DataFrame.loc with MultiIndex by @mroeschke in https://github.com/NVIDIA/cudf/pull/20457 +* Fix core dump in MemoryCleaner by @res-life in https://github.com/NVIDIA/cudf/pull/19872 +* Disallow multiple ellipse values in loc/iloc indexing by @mroeschke in https://github.com/NVIDIA/cudf/pull/20456 +* Fix `scan` operations for `string` columns by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20460 +* Fix UTF8 data generator in libcudf benchmarks utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/20465 +* Handle dealloc in stream-ordered cudf-polars ops by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20467 +* Raise on unsupported unstack cases by @Matt711 in https://github.com/NVIDIA/cudf/pull/20463 +* Allow early exit for left semi-/anti- joins with empty build/probe tables by @shrshi in https://github.com/NVIDIA/cudf/pull/20452 +* Fix OOB memory access in JSON reader ingest_raw utility by @davidwendt in https://github.com/NVIDIA/cudf/pull/20451 +* Round up small-type groupby outputs to 4-byte boundary by @PointKernel in https://github.com/NVIDIA/cudf/pull/20455 +* Fix GPU acceleration bug in decimal type-cast by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20471 +* Add missing CUDA stream in cudf_polars Distinct by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20477 +* Support `__arrow_array__` on proxy extension array by @Matt711 in https://github.com/NVIDIA/cudf/pull/20478 +* Enable scan operation for `datetime64` and `timedelta64` types by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20464 +* Remove unneeded type check in cudf::strings::slice_strings by @davidwendt in https://github.com/NVIDIA/cudf/pull/20437 +* Fix join match context tests by @PointKernel in https://github.com/NVIDIA/cudf/pull/20472 +* Fix the statistics_mr in benchmark fixture by @PointKernel in https://github.com/NVIDIA/cudf/pull/20496 +* Guard `__sizeof__` in pandas compatability mode by @Matt711 in https://github.com/NVIDIA/cudf/pull/20495 +* Fix OOB memory access in Orc and Parquet stacks from fixed-width unaligned loads by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20458 +* Fix cudf.pandas Timestamp/Timedelta not subclassing stdlib datetime objects by @mroeschke in https://github.com/NVIDIA/cudf/pull/20433 +* Revert benchmark input generation logic for list type by @davidwendt in https://github.com/NVIDIA/cudf/pull/20498 +* Avoid using pylibcudf directly in rapidsmpf runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/20501 +* Suppress NVRTC arch warnings by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/20517 +* Fix ``ChannelManager`` and ``Lineariser`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/20516 +* Synchronize streams in ``LocalShuffle`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/20515 +* Make `argsort` have return type `np.intp` to match pandas by @Matt711 in https://github.com/NVIDIA/cudf/pull/20487 +* Fix `polars.concat_str` with one column in cudf_polars by @mroeschke in https://github.com/NVIDIA/cudf/pull/20535 +* Override `__sizeof__` for `cudf.Index` by @Matt711 in https://github.com/NVIDIA/cudf/pull/20530 +* Fix `pl.scan_csv(...).slice(...).collect(engine="gpu")` with None endpoint by @mroeschke in https://github.com/NVIDIA/cudf/pull/20519 +* Fix DataChunkSourceTest by syncing default stream by @davidwendt in https://github.com/NVIDIA/cudf/pull/20492 +* Fix data size errors in some libcudf benchmarks by @davidwendt in https://github.com/NVIDIA/cudf/pull/20512 +* Pin cython and pytest dependencies by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20571 +* Pin Cython pre-3.2.0 and PyTest pre-9 by @jakirkham in https://github.com/NVIDIA/cudf/pull/20573 +* Handle `Empty` child IRs in `_decompose` by @Matt711 in https://github.com/NVIDIA/cudf/pull/20409 +* Skip flaky pandas datetime test by @Matt711 in https://github.com/NVIDIA/cudf/pull/20585 +* Fix max-pool-size-exceeded error in DATA_CHUNK_SOURCE_TEST by @davidwendt in https://github.com/NVIDIA/cudf/pull/20534 +* Fix racecheck in nvtext wordpiece tokenizer kernel by @davidwendt in https://github.com/NVIDIA/cudf/pull/20588 +* Fix the check to determine if all column chunk pages are dict encoded by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20524 +* Add stream synchronize to QUANTILES_TEST PercentileApprox gtests by @davidwendt in https://github.com/NVIDIA/cudf/pull/20558 +* updated update-version.sh to handle release branch version changes by @rockhowse in https://github.com/NVIDIA/cudf/pull/20598 +* Fix nvtext tokenizers handling invalid UTF8 data by @davidwendt in https://github.com/NVIDIA/cudf/pull/20514 +* Fix overflow errors in distinct and filtered joins when hash table size exceeds int32 limits by @shrshi in https://github.com/NVIDIA/cudf/pull/20594 +* [FEA] Optimize JIT Filter for Low-Selectivity by @lamarrr in https://github.com/NVIDIA/cudf/pull/20222 +* Compute boolean function(NOT) on integers as a bitwise invert by @Matt711 in https://github.com/NVIDIA/cudf/pull/20599 +* Cast output dtype of rolling aggregations to match pandas by @Matt711 in https://github.com/NVIDIA/cudf/pull/20526 +* Add noop path for `Frame.astype` by @Matt711 in https://github.com/NVIDIA/cudf/pull/20581 +* Fix `copy` semantics bugs thus reduce copies and memory usage by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20121 +* Ensure the sum after expression decomposition for mean has float output dtype by @Matt711 in https://github.com/NVIDIA/cudf/pull/20596 +* Use `Decimal(0)` literal for all-null decimal groups in groupby-sum by @Matt711 in https://github.com/NVIDIA/cudf/pull/20591 +* Do not drop `freq` when constructing `DatetimeIndex` from pandas by @brandon-b-miller in https://github.com/NVIDIA/cudf/pull/18778 +* Fix --validation flag for cudf.pandas PDSH benchmarks by @mroeschke in https://github.com/NVIDIA/cudf/pull/20540 +* Enable GPU acceleration for more binops by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20507 +* Fix `rmm` function calls due to removed deprecated APIs and macro by @ttnghia in https://github.com/NVIDIA/cudf/pull/20661 +* Fix orc reader bool bug due to not being able to resume rle decode by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20666 +* Fix categorical comparisons in `cudf` to match `pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20674 +* Fix `any` and `all` to match pandas by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20679 +* Fix return types of string APIs in `cudf.pandas` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20683 +* Resolve pandas test failures by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20704 +* Fix DatetimeIndex pickling by @vyasr in https://github.com/NVIDIA/cudf/pull/20709 +* `DatetimeIndex.serialize()` headers are msgpack serializable by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20714 ### 📖 Documentation -* Add note that --rmm-async only affects distributed scheduler. by @bdice in https://github.com/rapidsai/cudf/pull/20129 -* Add profiling guide by @bdice in https://github.com/rapidsai/cudf/pull/20292 -* Find RMM before CCCL by @wence- in https://github.com/rapidsai/cudf/pull/20336 -* Use current system architecture in conda environment creation command by @bdice in https://github.com/rapidsai/cudf/pull/20500 -* Use uname -m instead of arch command by @bdice in https://github.com/rapidsai/cudf/pull/20502 -* Use RAPIDS_BRANCH file for documentation links by @bdice in https://github.com/rapidsai/cudf/pull/20494 +* Add note that --rmm-async only affects distributed scheduler. by @bdice in https://github.com/NVIDIA/cudf/pull/20129 +* Add profiling guide by @bdice in https://github.com/NVIDIA/cudf/pull/20292 +* Find RMM before CCCL by @wence- in https://github.com/NVIDIA/cudf/pull/20336 +* Use current system architecture in conda environment creation command by @bdice in https://github.com/NVIDIA/cudf/pull/20500 +* Use uname -m instead of arch command by @bdice in https://github.com/NVIDIA/cudf/pull/20502 +* Use RAPIDS_BRANCH file for documentation links by @bdice in https://github.com/NVIDIA/cudf/pull/20494 ### 🚀 New Features -* Add memory resources to unary, transform, and filling modules by @vyasr in https://github.com/rapidsai/cudf/pull/20054 -* Add memory resources to binaryop, copying, and stream_compaction by @vyasr in https://github.com/rapidsai/cudf/pull/20059 -* Add memory resources to groupby, datetime, and lists modules by @vyasr in https://github.com/rapidsai/cudf/pull/20102 -* Add memory resources to search, reshape, and partitioning module by @vyasr in https://github.com/rapidsai/cudf/pull/20101 -* Add memory resources to rolling, sorting, and quantiles modules by @vyasr in https://github.com/rapidsai/cudf/pull/20099 -* [FEA] Implement JIT Filter for read_parquet by @lamarrr in https://github.com/rapidsai/cudf/pull/19831 -* Add memory resources to all nvtext APIs by @vyasr in https://github.com/rapidsai/cudf/pull/20119 -* Add memory resource to all strings modules by @vyasr in https://github.com/rapidsai/cudf/pull/20123 -* Add memory resources to reduce, column, column_factories, and contiguous_split by @vyasr in https://github.com/rapidsai/cudf/pull/20135 -* Add memory resources to I/O modules by @vyasr in https://github.com/rapidsai/cudf/pull/20136 -* Remove rounding from cudf java by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20110 -* Add memory resources to replace, json, and hashing by @vyasr in https://github.com/rapidsai/cudf/pull/20150 -* Add support for maintain_order param in joins by @Matt711 in https://github.com/rapidsai/cudf/pull/17698 -* Add an example to inspect parquet files and dump row group and page level metadata information by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20117 -* Support forward/backward filling null values in a grouped window context by @Matt711 in https://github.com/rapidsai/cudf/pull/19907 -* Allow multiple calls to `cudf::initialize` and `cudf::deinitialize` by @vuule in https://github.com/rapidsai/cudf/pull/20111 -* Add remaining memory resources by @vyasr in https://github.com/rapidsai/cudf/pull/20197 -* Add memory resources to scalars by @vyasr in https://github.com/rapidsai/cudf/pull/20196 -* Add pylibcudf is_valid_reduce_aggregation API by @davidwendt in https://github.com/rapidsai/cudf/pull/20145 -* Support decimal literals in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/20147 -* Support `cum_sum(...).over(...)` expressions in cudf-polars by @Matt711 in https://github.com/rapidsai/cudf/pull/19908 -* Passthrough unary ops through Parquet predicate pushdown by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20127 -* Implement `ARGMIN` and `ARGMAX` aggregations for reduction by @ttnghia in https://github.com/rapidsai/cudf/pull/20207 -* Skip decompression of pruned parquet pages by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20192 -* Add an example to demonstrate the use of next-gen parquet reader to read a parquet file with highly selective filters by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/19469 -* Evaluate `IS_NULL` at row group and page level in Parquet filtering by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20144 -* [Java] Add optional native deps loader by @zpuller in https://github.com/rapidsai/cudf/pull/20414 -* Add cudf-polars + rapidsmpf CI check by @rjzamora in https://github.com/rapidsai/cudf/pull/20355 -* Add Python bindings for the hybrid scan reader by @vyasr in https://github.com/rapidsai/cudf/pull/20381 -* RapidsMPF streaming-engine translation by @rjzamora in https://github.com/rapidsai/cudf/pull/20161 -* [JNI] Use a read/write lock pattern in Rmm.class by @abellina in https://github.com/rapidsai/cudf/pull/20521 -* [Java] Supports output projection indices for `contiguousSplitGroupsAndGenUniqKeys` by @res-life in https://github.com/rapidsai/cudf/pull/20391 -* Support `Series.at` and `Series.iat` for pandas compatability by @Matt711 in https://github.com/rapidsai/cudf/pull/20529 -* Add COUNT_VALID aggregation support to groupby-scan by @davidwendt in https://github.com/rapidsai/cudf/pull/20531 -* Use RapidsMPF `read_parquet` in "rapidsmpf" runtime by @rjzamora in https://github.com/rapidsai/cudf/pull/20497 -* Support decimal128 SUM aggregation in hash-based groupby by @PointKernel in https://github.com/rapidsai/cudf/pull/20509 -* Add stream testing in pylibcudf by @vyasr in https://github.com/rapidsai/cudf/pull/20625 +* Add memory resources to unary, transform, and filling modules by @vyasr in https://github.com/NVIDIA/cudf/pull/20054 +* Add memory resources to binaryop, copying, and stream_compaction by @vyasr in https://github.com/NVIDIA/cudf/pull/20059 +* Add memory resources to groupby, datetime, and lists modules by @vyasr in https://github.com/NVIDIA/cudf/pull/20102 +* Add memory resources to search, reshape, and partitioning module by @vyasr in https://github.com/NVIDIA/cudf/pull/20101 +* Add memory resources to rolling, sorting, and quantiles modules by @vyasr in https://github.com/NVIDIA/cudf/pull/20099 +* [FEA] Implement JIT Filter for read_parquet by @lamarrr in https://github.com/NVIDIA/cudf/pull/19831 +* Add memory resources to all nvtext APIs by @vyasr in https://github.com/NVIDIA/cudf/pull/20119 +* Add memory resource to all strings modules by @vyasr in https://github.com/NVIDIA/cudf/pull/20123 +* Add memory resources to reduce, column, column_factories, and contiguous_split by @vyasr in https://github.com/NVIDIA/cudf/pull/20135 +* Add memory resources to I/O modules by @vyasr in https://github.com/NVIDIA/cudf/pull/20136 +* Remove rounding from cudf java by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20110 +* Add memory resources to replace, json, and hashing by @vyasr in https://github.com/NVIDIA/cudf/pull/20150 +* Add support for maintain_order param in joins by @Matt711 in https://github.com/NVIDIA/cudf/pull/17698 +* Add an example to inspect parquet files and dump row group and page level metadata information by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20117 +* Support forward/backward filling null values in a grouped window context by @Matt711 in https://github.com/NVIDIA/cudf/pull/19907 +* Allow multiple calls to `cudf::initialize` and `cudf::deinitialize` by @vuule in https://github.com/NVIDIA/cudf/pull/20111 +* Add remaining memory resources by @vyasr in https://github.com/NVIDIA/cudf/pull/20197 +* Add memory resources to scalars by @vyasr in https://github.com/NVIDIA/cudf/pull/20196 +* Add pylibcudf is_valid_reduce_aggregation API by @davidwendt in https://github.com/NVIDIA/cudf/pull/20145 +* Support decimal literals in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/20147 +* Support `cum_sum(...).over(...)` expressions in cudf-polars by @Matt711 in https://github.com/NVIDIA/cudf/pull/19908 +* Passthrough unary ops through Parquet predicate pushdown by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20127 +* Implement `ARGMIN` and `ARGMAX` aggregations for reduction by @ttnghia in https://github.com/NVIDIA/cudf/pull/20207 +* Skip decompression of pruned parquet pages by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20192 +* Add an example to demonstrate the use of next-gen parquet reader to read a parquet file with highly selective filters by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/19469 +* Evaluate `IS_NULL` at row group and page level in Parquet filtering by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20144 +* [Java] Add optional native deps loader by @zpuller in https://github.com/NVIDIA/cudf/pull/20414 +* Add cudf-polars + rapidsmpf CI check by @rjzamora in https://github.com/NVIDIA/cudf/pull/20355 +* Add Python bindings for the hybrid scan reader by @vyasr in https://github.com/NVIDIA/cudf/pull/20381 +* RapidsMPF streaming-engine translation by @rjzamora in https://github.com/NVIDIA/cudf/pull/20161 +* [JNI] Use a read/write lock pattern in Rmm.class by @abellina in https://github.com/NVIDIA/cudf/pull/20521 +* [Java] Supports output projection indices for `contiguousSplitGroupsAndGenUniqKeys` by @res-life in https://github.com/NVIDIA/cudf/pull/20391 +* Support `Series.at` and `Series.iat` for pandas compatability by @Matt711 in https://github.com/NVIDIA/cudf/pull/20529 +* Add COUNT_VALID aggregation support to groupby-scan by @davidwendt in https://github.com/NVIDIA/cudf/pull/20531 +* Use RapidsMPF `read_parquet` in "rapidsmpf" runtime by @rjzamora in https://github.com/NVIDIA/cudf/pull/20497 +* Support decimal128 SUM aggregation in hash-based groupby by @PointKernel in https://github.com/NVIDIA/cudf/pull/20509 +* Add stream testing in pylibcudf by @vyasr in https://github.com/NVIDIA/cudf/pull/20625 ### 🛠️ Improvements -* Deprecate .from_pandas constructor by @mroeschke in https://github.com/rapidsai/cudf/pull/19996 -* Prune entries in Sphinx nitpick_ignore by @mroeschke in https://github.com/rapidsai/cudf/pull/20045 -* Avoid direct CategoricalColumn calls in dask_cudf by @mroeschke in https://github.com/rapidsai/cudf/pull/20080 -* Fix typing issues in pylibcudf by @vyasr in https://github.com/rapidsai/cudf/pull/20069 -* Avoid shadowing module names by @vyasr in https://github.com/rapidsai/cudf/pull/20071 -* Remove calling to `purge_nonempty_nulls` in `make_lists_column` by @ttnghia in https://github.com/rapidsai/cudf/pull/12873 -* Reduce verbosity of running the pandas test suite by @vyasr in https://github.com/rapidsai/cudf/pull/20107 -* Clean up detail device atomic logic using atomic_ref by @PointKernel in https://github.com/rapidsai/cudf/pull/19924 -* Use 8 processes for pandas tests, show top 10 test times by @bdice in https://github.com/rapidsai/cudf/pull/20109 -* Update nvbench by @bdice in https://github.com/rapidsai/cudf/pull/19619 -* Cleanup of some libcudf aggregation code by @davidwendt in https://github.com/rapidsai/cudf/pull/20053 -* Run cudf-polars conda unit tests with more than 1 process by @mroeschke in https://github.com/rapidsai/cudf/pull/19980 -* Avoid running pandas unit tests for private functionality with cudf.pandas by @mroeschke in https://github.com/rapidsai/cudf/pull/20115 -* Remove MultiIndex.from_pandas pytest benchmark by @mroeschke in https://github.com/rapidsai/cudf/pull/20112 -* Switch host_vector and host_span dependency by @davidwendt in https://github.com/rapidsai/cudf/pull/20106 -* Have ListColumn.from_sequence go through pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/20098 -* Fix `RAPIDS_BRANCH` version and update script by @galipremsagar in https://github.com/rapidsai/cudf/pull/20091 -* Add pyarrow stubs to mypy environment and fix associated errors by @vyasr in https://github.com/rapidsai/cudf/pull/20118 -* Fix slowdown in cudf-polars distributed tests by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20137 -* Improve performance of string column size computation during parquet reads. by @nvdbaranec in https://github.com/rapidsai/cudf/pull/19986 -* Disable async MR priming in cudf.pandas by @bdice in https://github.com/rapidsai/cudf/pull/20133 -* Rework reduction case statement as dispatch_type_and_aggregation by @davidwendt in https://github.com/rapidsai/cudf/pull/20078 -* Fix type annotations in cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20131 -* Add tests for AUTO and HYBRID (de)compression modes by @vuule in https://github.com/rapidsai/cudf/pull/20126 -* Branch 25.12 merge branch 25.10 by @vyasr in https://github.com/rapidsai/cudf/pull/20152 -* Manual forward merger for Branch 25.12 - branch 25.10 by @galipremsagar in https://github.com/rapidsai/cudf/pull/20157 -* Temporarily disable conda-java-tests by @bdice in https://github.com/rapidsai/cudf/pull/20162 -* Remove unused ColumnBase.view by @mroeschke in https://github.com/rapidsai/cudf/pull/20141 -* Avoid NumericalColumn call from CategoricalColumn.children by @mroeschke in https://github.com/rapidsai/cudf/pull/20153 -* Deprecate legacy public row operators by @PointKernel in https://github.com/rapidsai/cudf/pull/20097 -* Avoid more explicit calls to IntervalColumn and StructColumn by @mroeschke in https://github.com/rapidsai/cudf/pull/20064 -* Run cudf-polars wheels unit tests with more than 1 process by @mroeschke in https://github.com/rapidsai/cudf/pull/20124 -* Trace node execution in cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/19895 -* Make ColumnBase.as_*_column convert via pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/20149 -* Reduce execution times for parquet dictionary tests by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20176 -* Update to rapids-logger 0.2 by @bdice in https://github.com/rapidsai/cudf/pull/20172 -* Adjust rmm pool handling in PDSH benchmarks by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20138 -* Don't assume cudf_polars benchmarking scale factor is always an integer by @mroeschke in https://github.com/rapidsai/cudf/pull/20182 -* Skip filtering Parquet row groups with dictionaries if there are non-dict encoded pages by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20175 -* Remove unnecessary work from `read_parquet_metadata` by @vuule in https://github.com/rapidsai/cudf/pull/20180 -* Improve performance of groupby tdigests gtests by @davidwendt in https://github.com/rapidsai/cudf/pull/20173 -* Revert "Temporarily disable conda-java-tests" by @bdice in https://github.com/rapidsai/cudf/pull/20184 -* Add PDSH benchmark runner for cudf.pandas by @mroeschke in https://github.com/rapidsai/cudf/pull/20164 -* Make Column.set_mask go through pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/20103 -* Pin pydantic<2.12 in ci/test_cudf_polars_polars_tests.sh by @mroeschke in https://github.com/rapidsai/cudf/pull/20200 -* Add an overhead field to cudf-polars tracing by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20198 -* Support binops between float scalar to decimal column by @mroeschke in https://github.com/rapidsai/cudf/pull/20199 -* Reduce output buffer sizes for pruned pages of columns with a `list` parent by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20086 -* Make ListColumn._transform_leaves convert via pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/20151 -* Rename `comparison_binop_generator` to `arg_minmax_binop_generator` and corresponding file to `nested_types_extrema_utils.cuh` by @Copilot in https://github.com/rapidsai/cudf/pull/20212 -* Pin polars version <1.34 and >=1.29 by @Matt711 in https://github.com/rapidsai/cudf/pull/19912 -* Stop using libcudf default parameters in pylibcudf by @vyasr in https://github.com/rapidsai/cudf/pull/20204 -* Fix various typing errors by @vyasr in https://github.com/rapidsai/cudf/pull/20205 -* Cleanup parquet for simple columns by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/19869 -* Configuration for which metrics are enabled during tracing by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20223 -* [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/rapidsai/cudf/pull/20189 -* Fix parquet row number check for page bounds by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20217 -* More mypy and docs fixes by @vyasr in https://github.com/rapidsai/cudf/pull/20224 -* Prevent accidental copies of expensive-to-copy object types by @vuule in https://github.com/rapidsai/cudf/pull/20226 -* Split row operator header by @PointKernel in https://github.com/rapidsai/cudf/pull/20166 -* Standardize setting StructDtype field names post libcudf conversion by @mroeschke in https://github.com/rapidsai/cudf/pull/20235 -* Add arm testing of cudf.pandas unit tests by @vyasr in https://github.com/rapidsai/cudf/pull/20251 -* Enable `sccache-dist` connection pool by @trxcllnt in https://github.com/rapidsai/cudf/pull/20264 -* Run polars tests with the streaming and in-memory executors by @Matt711 in https://github.com/rapidsai/cudf/pull/19354 -* Move and rename ``ScanPartitionPlan`` by @rjzamora in https://github.com/rapidsai/cudf/pull/20248 -* Unpin DuckDB and Ibis in cudf.pandas thirdparty tests by @mroeschke in https://github.com/rapidsai/cudf/pull/20269 -* Add pylibcudf to pre-commit linting and fix outstanding errors by @vyasr in https://github.com/rapidsai/cudf/pull/20250 -* Update ``ConfigOptions`` for rapidsmpf-streaming integration by @rjzamora in https://github.com/rapidsai/cudf/pull/20252 -* Handle unordered grouped windows properly for null filling and cum sums by @Matt711 in https://github.com/rapidsai/cudf/pull/20275 -* Add more type annotations to cudf/core/column subclasses by @mroeschke in https://github.com/rapidsai/cudf/pull/20277 -* Remove extraneous host_memory_resource include by @bdice in https://github.com/rapidsai/cudf/pull/20284 -* Add `MultiIndex.dtypes` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20279 -* Skip mypy in pre-commit.ci by @bdice in https://github.com/rapidsai/cudf/pull/20286 -* Make ColumnBase.deserialize construct via pylibcudf by @mroeschke in https://github.com/rapidsai/cudf/pull/20142 -* Add numpy to the mypy pre-commit environment by @vyasr in https://github.com/rapidsai/cudf/pull/20282 -* Add ability to set the source_info of parquet_reader_options by @wence- in https://github.com/rapidsai/cudf/pull/20253 -* Add more Python type annotations to `cudf/core` by @mroeschke in https://github.com/rapidsai/cudf/pull/20287 -* Use main in RAPIDS_BRANCH by @bdice in https://github.com/rapidsai/cudf/pull/20312 -* Move "All rights reserved" statements to copyright line by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/20313 -* Add `inferred_type` and missing `IntervalIndex` properties by @galipremsagar in https://github.com/rapidsai/cudf/pull/20294 -* Avoid unseeded, random data generation in cuDF classic tests by @mroeschke in https://github.com/rapidsai/cudf/pull/20319 -* Improve hash-based groupby aggregation: direct write to the dense output columns whenever possible by @ttnghia in https://github.com/rapidsai/cudf/pull/19764 -* Avoid accessing range values in cudf::strings::contains_re logic by @davidwendt in https://github.com/rapidsai/cudf/pull/20122 -* Migrate mixed join to use the multiset data structure by @PointKernel in https://github.com/rapidsai/cudf/pull/19989 -* Add benchmark for strings cast to/from integer APIs by @davidwendt in https://github.com/rapidsai/cudf/pull/20247 -* Use main shared-workflows branch by @bdice in https://github.com/rapidsai/cudf/pull/20324 -* Use the thread pool for Parquet metadata processing by @vuule in https://github.com/rapidsai/cudf/pull/20263 -* Add `.dt.day_of_week` and `.dt.daysinmonth` by @galipremsagar in https://github.com/rapidsai/cudf/pull/20298 -* Avoid Column materialization in RangeIndex.nans_to_nulls by @mroeschke in https://github.com/rapidsai/cudf/pull/20331 -* Update the code to be compatible with the new cuco stream-ordered allocator by @PointKernel in https://github.com/rapidsai/cudf/pull/20258 -* Deprecate Series.data by @mroeschke in https://github.com/rapidsai/cudf/pull/20281 -* Align cudf Python's Column constructors by @mroeschke in https://github.com/rapidsai/cudf/pull/20233 -* Make type annotations of ColumnBase.set_mask stricter by @mroeschke in https://github.com/rapidsai/cudf/pull/20261 -* Make type annotations of ColumnBase.find_and_replace stricter by @mroeschke in https://github.com/rapidsai/cudf/pull/20259 -* Make type annotations of ColumnBase.apply_boolean_mask stricter by @mroeschke in https://github.com/rapidsai/cudf/pull/20262 -* Skip Python LZ4 tests when nvCOMP is disabled by @vuule in https://github.com/rapidsai/cudf/pull/20293 -* Move cudf/io/nvcomp_adapter.hpp to cudf/io/detail by @davidwendt in https://github.com/rapidsai/cudf/pull/20327 -* Add context to IR.do_evaluate by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20322 -* Update mypy `# type: ignore` comments according to stricter mypy configs by @mroeschke in https://github.com/rapidsai/cudf/pull/20272 -* Remove duplicated enforce null consistency code by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20342 -* Use SPDX for all copyright headers by @KyleFromNVIDIA in https://github.com/rapidsai/cudf/pull/20321 -* Add more type annotations to `cudf/core/series.py` by @mroeschke in https://github.com/rapidsai/cudf/pull/20304 -* Remove/Replace uses of numba.cuda arrays in pytest benchmarks and tests by @mroeschke in https://github.com/rapidsai/cudf/pull/20359 -* Add duckdb pdsh query queries by @Matt711 in https://github.com/rapidsai/cudf/pull/20257 -* Use stream in cudf_polars.DataFrame.to_polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20323 -* Add `join_streams` to pylibcudf API by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20316 -* Use CUDA streams in all pylibcudf calls made by cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20291 -* Add cudf/io/config_utils.hpp to doxygen by @davidwendt in https://github.com/rapidsai/cudf/pull/20329 -* Test coverage for parallel metadata parsing by @vuule in https://github.com/rapidsai/cudf/pull/20334 -* Support serializing more polars types by @Matt711 in https://github.com/rapidsai/cudf/pull/20347 -* Add CUDAStreamPolicy to cudf-polars configuration by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20366 -* Unskip cudf-polars groupby test by @Matt711 in https://github.com/rapidsai/cudf/pull/20406 -* Deprecate pylibcudf interop arrow APIs by @Matt711 in https://github.com/rapidsai/cudf/pull/20405 -* Get rid of the hashing helper header by @PointKernel in https://github.com/rapidsai/cudf/pull/20360 -* Minor cleanup and fixes for libcudf generate_input.cu by @davidwendt in https://github.com/rapidsai/cudf/pull/20363 -* Ignore assert_produces_warning and shares_memory pandas unit tests for cudf.pandas by @mroeschke in https://github.com/rapidsai/cudf/pull/20434 -* Short circut RangeIndex.append for length 0 input, proxy private attribute by @mroeschke in https://github.com/rapidsai/cudf/pull/20442 -* Mark DataFrame.insert as _external_only_api by @Copilot in https://github.com/rapidsai/cudf/pull/20403 -* Deprecate `get_current_device_resource` in favor of `get_current_device_resource_ref` by @PointKernel in https://github.com/rapidsai/cudf/pull/20386 -* Promote `JoinNoneValue` to public as `JoinNoMatch` for clear non-match Join semantics by @PointKernel in https://github.com/rapidsai/cudf/pull/20440 -* Remove duplicate entries in NODEIDS_THAT_FAIL_WITH_CUDF_PANDAS by @mroeschke in https://github.com/rapidsai/cudf/pull/20447 -* Use the thread pool in the compact protocol reader by @vuule in https://github.com/rapidsai/cudf/pull/20417 -* Update README.md generalizing all cuDF components by @mroeschke in https://github.com/rapidsai/cudf/pull/20357 -* Skip TestDatetimelikeCoercion pandas tests that assert ._value identity by @mroeschke in https://github.com/rapidsai/cudf/pull/20459 -* Add PSDH Q2-9 for cudf.pandas by @mroeschke in https://github.com/rapidsai/cudf/pull/20418 -* Add s3fs to `test_cudf_python` common dependencies by @trxcllnt in https://github.com/rapidsai/cudf/pull/20473 -* Use public pandas APIs in StringColumn.to_pandas by @mroeschke in https://github.com/rapidsai/cudf/pull/20474 -* Expose java GatherMap internals and add toString to AST by @revans2 in https://github.com/rapidsai/cudf/pull/20483 -* Add create_ascii_string_column to the libcudf benchmark data generator by @davidwendt in https://github.com/rapidsai/cudf/pull/20354 -* Skip more pandas unit tests that tests BlockManager, private sparse types by @mroeschke in https://github.com/rapidsai/cudf/pull/20489 -* Add boto3/botocore/aiobotocore to common test dependencies by @trxcllnt in https://github.com/rapidsai/cudf/pull/20490 -* Use a lower bound when estimating the partial file-size by @rjzamora in https://github.com/rapidsai/cudf/pull/20193 -* Performance improvement for nvtext::edit_distance for long strings by @davidwendt in https://github.com/rapidsai/cudf/pull/20268 -* Add MemoryResourceConfig to cudf-polars config by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20042 -* Improve project automation by @vyasr in https://github.com/rapidsai/cudf/pull/20523 -* Fuse simple streaming reductions in cudf-polars by @rjzamora in https://github.com/rapidsai/cudf/pull/18757 -* Migrate to new CCCL memory resource interface by @bdice in https://github.com/rapidsai/cudf/pull/20513 -* Add empty input gtest for cudf::transform by @davidwendt in https://github.com/rapidsai/cudf/pull/20505 -* Rework internal json headers to allow converting gtests files from .cu to .cpp by @davidwendt in https://github.com/rapidsai/cudf/pull/20491 -* Set continue on error in the cudf-polars-rapidsmpf nightly CI job by @Matt711 in https://github.com/rapidsai/cudf/pull/20550 -* Permanently back cuDF column by a pylibcudf.Column by @mroeschke in https://github.com/rapidsai/cudf/pull/20306 -* Skip flaky upstream polars rolling test by @Matt711 in https://github.com/rapidsai/cudf/pull/20552 -* Accelerate data page mask computation on device by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20280 -* Change default rapidsmpf stream policy to 'pool' by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20527 -* Increase gtests coverage for cudf::strings::like patterns by @davidwendt in https://github.com/rapidsai/cudf/pull/20348 -* Add cuda::std::span operator to cudf::column_view by @davidwendt in https://github.com/rapidsai/cudf/pull/20541 -* Update ArrowStringView compare benchmark for gather by @davidwendt in https://github.com/rapidsai/cudf/pull/19935 -* Add pytest stubs and remove ujson usage by @vyasr in https://github.com/rapidsai/cudf/pull/20560 -* Skip arrow array constructor tests by @Matt711 in https://github.com/rapidsai/cudf/pull/20579 -* Add Polars to mypy environment and fix errors by @vyasr in https://github.com/rapidsai/cudf/pull/20563 -* Ensure table chunks are unspilled and available by @madsbk in https://github.com/rapidsai/cudf/pull/20583 -* Skip tests that assert behavior when copy-on-write is False by @Matt711 in https://github.com/rapidsai/cudf/pull/20506 -* Pass streams through `Column.from_array`/`from_iterable_of_py` by @Matt711 in https://github.com/rapidsai/cudf/pull/20569 -* Stop using Dtype annotation by @vyasr in https://github.com/rapidsai/cudf/pull/20590 -* Workaround to enable running PDS-H via WebHDFS by @kingcrimsontianyu in https://github.com/rapidsai/cudf/pull/20132 -* Update RMM includes from `` to `` by @bdice in https://github.com/rapidsai/cudf/pull/20607 -* Stricter typing import for cudf-polars by @TomAugspurger in https://github.com/rapidsai/cudf/pull/20614 -* Avoid the unnecessary H2H copy in the `std::vector` sink by @vuule in https://github.com/rapidsai/cudf/pull/20602 -* Preprocessing offsets for Parquet non-dictionary string columns by @pmattione-nvidia in https://github.com/rapidsai/cudf/pull/20430 -* Move more pandas unit tests that test private APIs by @mroeschke in https://github.com/rapidsai/cudf/pull/20511 -* Use `.plc_column` instead of `.to_pylibcudf` in rolling, string utilties by @mroeschke in https://github.com/rapidsai/cudf/pull/20562 -* Skip TestSetitemNADatetimeLikeDtype pandas unit tests due to private assertion by @mroeschke in https://github.com/rapidsai/cudf/pull/20578 -* Pin Polars version <1.35 by @Matt711 in https://github.com/rapidsai/cudf/pull/20266 -* Skip pandas unit tests in `test_old_base.py` that test private APIs by @mroeschke in https://github.com/rapidsai/cudf/pull/20572 -* Use `.plc_column` attribute instead of `to_pylibcudf` more internally by @mroeschke in https://github.com/rapidsai/cudf/pull/20559 -* Skip arrow-backed arithmetic tests and categorize the remaining failing tests by @Matt711 in https://github.com/rapidsai/cudf/pull/20577 -* Fix a pytest execution that is spawned in a subprocess by @galipremsagar in https://github.com/rapidsai/cudf/pull/20660 -* Accelerated parquet page header decoding when page index is available by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20369 -* feat: add error handling for non-existent columns in parquet reader by @gforsyth in https://github.com/rapidsai/cudf/pull/20659 -* Optimize row mask computation for single filter column by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20335 -* Skip MultiIndex pandas unit tests testing private functionalty, `test_chaining_and_caching.py` by @mroeschke in https://github.com/rapidsai/cudf/pull/20575 -* Address minor comments from recent hybrid scan PRs by @mhaseeb123 in https://github.com/rapidsai/cudf/pull/20672 -* Add a timeout for the rapidsmpf test run by @vyasr in https://github.com/rapidsai/cudf/pull/20681 -* Use `sccache-dist` build cluster for conda and wheel builds by @trxcllnt in https://github.com/rapidsai/cudf/pull/20488 +* Deprecate .from_pandas constructor by @mroeschke in https://github.com/NVIDIA/cudf/pull/19996 +* Prune entries in Sphinx nitpick_ignore by @mroeschke in https://github.com/NVIDIA/cudf/pull/20045 +* Avoid direct CategoricalColumn calls in dask_cudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/20080 +* Fix typing issues in pylibcudf by @vyasr in https://github.com/NVIDIA/cudf/pull/20069 +* Avoid shadowing module names by @vyasr in https://github.com/NVIDIA/cudf/pull/20071 +* Remove calling to `purge_nonempty_nulls` in `make_lists_column` by @ttnghia in https://github.com/NVIDIA/cudf/pull/12873 +* Reduce verbosity of running the pandas test suite by @vyasr in https://github.com/NVIDIA/cudf/pull/20107 +* Clean up detail device atomic logic using atomic_ref by @PointKernel in https://github.com/NVIDIA/cudf/pull/19924 +* Use 8 processes for pandas tests, show top 10 test times by @bdice in https://github.com/NVIDIA/cudf/pull/20109 +* Update nvbench by @bdice in https://github.com/NVIDIA/cudf/pull/19619 +* Cleanup of some libcudf aggregation code by @davidwendt in https://github.com/NVIDIA/cudf/pull/20053 +* Run cudf-polars conda unit tests with more than 1 process by @mroeschke in https://github.com/NVIDIA/cudf/pull/19980 +* Avoid running pandas unit tests for private functionality with cudf.pandas by @mroeschke in https://github.com/NVIDIA/cudf/pull/20115 +* Remove MultiIndex.from_pandas pytest benchmark by @mroeschke in https://github.com/NVIDIA/cudf/pull/20112 +* Switch host_vector and host_span dependency by @davidwendt in https://github.com/NVIDIA/cudf/pull/20106 +* Have ListColumn.from_sequence go through pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/20098 +* Fix `RAPIDS_BRANCH` version and update script by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20091 +* Add pyarrow stubs to mypy environment and fix associated errors by @vyasr in https://github.com/NVIDIA/cudf/pull/20118 +* Fix slowdown in cudf-polars distributed tests by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20137 +* Improve performance of string column size computation during parquet reads. by @nvdbaranec in https://github.com/NVIDIA/cudf/pull/19986 +* Disable async MR priming in cudf.pandas by @bdice in https://github.com/NVIDIA/cudf/pull/20133 +* Rework reduction case statement as dispatch_type_and_aggregation by @davidwendt in https://github.com/NVIDIA/cudf/pull/20078 +* Fix type annotations in cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20131 +* Add tests for AUTO and HYBRID (de)compression modes by @vuule in https://github.com/NVIDIA/cudf/pull/20126 +* Branch 25.12 merge branch 25.10 by @vyasr in https://github.com/NVIDIA/cudf/pull/20152 +* Manual forward merger for Branch 25.12 - branch 25.10 by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20157 +* Temporarily disable conda-java-tests by @bdice in https://github.com/NVIDIA/cudf/pull/20162 +* Remove unused ColumnBase.view by @mroeschke in https://github.com/NVIDIA/cudf/pull/20141 +* Avoid NumericalColumn call from CategoricalColumn.children by @mroeschke in https://github.com/NVIDIA/cudf/pull/20153 +* Deprecate legacy public row operators by @PointKernel in https://github.com/NVIDIA/cudf/pull/20097 +* Avoid more explicit calls to IntervalColumn and StructColumn by @mroeschke in https://github.com/NVIDIA/cudf/pull/20064 +* Run cudf-polars wheels unit tests with more than 1 process by @mroeschke in https://github.com/NVIDIA/cudf/pull/20124 +* Trace node execution in cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/19895 +* Make ColumnBase.as_*_column convert via pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/20149 +* Reduce execution times for parquet dictionary tests by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20176 +* Update to rapids-logger 0.2 by @bdice in https://github.com/NVIDIA/cudf/pull/20172 +* Adjust rmm pool handling in PDSH benchmarks by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20138 +* Don't assume cudf_polars benchmarking scale factor is always an integer by @mroeschke in https://github.com/NVIDIA/cudf/pull/20182 +* Skip filtering Parquet row groups with dictionaries if there are non-dict encoded pages by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20175 +* Remove unnecessary work from `read_parquet_metadata` by @vuule in https://github.com/NVIDIA/cudf/pull/20180 +* Improve performance of groupby tdigests gtests by @davidwendt in https://github.com/NVIDIA/cudf/pull/20173 +* Revert "Temporarily disable conda-java-tests" by @bdice in https://github.com/NVIDIA/cudf/pull/20184 +* Add PDSH benchmark runner for cudf.pandas by @mroeschke in https://github.com/NVIDIA/cudf/pull/20164 +* Make Column.set_mask go through pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/20103 +* Pin pydantic<2.12 in ci/test_cudf_polars_polars_tests.sh by @mroeschke in https://github.com/NVIDIA/cudf/pull/20200 +* Add an overhead field to cudf-polars tracing by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20198 +* Support binops between float scalar to decimal column by @mroeschke in https://github.com/NVIDIA/cudf/pull/20199 +* Reduce output buffer sizes for pruned pages of columns with a `list` parent by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20086 +* Make ListColumn._transform_leaves convert via pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/20151 +* Rename `comparison_binop_generator` to `arg_minmax_binop_generator` and corresponding file to `nested_types_extrema_utils.cuh` by @Copilot in https://github.com/NVIDIA/cudf/pull/20212 +* Pin polars version <1.34 and >=1.29 by @Matt711 in https://github.com/NVIDIA/cudf/pull/19912 +* Stop using libcudf default parameters in pylibcudf by @vyasr in https://github.com/NVIDIA/cudf/pull/20204 +* Fix various typing errors by @vyasr in https://github.com/NVIDIA/cudf/pull/20205 +* Cleanup parquet for simple columns by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/19869 +* Configuration for which metrics are enabled during tracing by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20223 +* [pre-commit.ci] pre-commit autoupdate by @pre-commit-ci[bot] in https://github.com/NVIDIA/cudf/pull/20189 +* Fix parquet row number check for page bounds by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20217 +* More mypy and docs fixes by @vyasr in https://github.com/NVIDIA/cudf/pull/20224 +* Prevent accidental copies of expensive-to-copy object types by @vuule in https://github.com/NVIDIA/cudf/pull/20226 +* Split row operator header by @PointKernel in https://github.com/NVIDIA/cudf/pull/20166 +* Standardize setting StructDtype field names post libcudf conversion by @mroeschke in https://github.com/NVIDIA/cudf/pull/20235 +* Add arm testing of cudf.pandas unit tests by @vyasr in https://github.com/NVIDIA/cudf/pull/20251 +* Enable `sccache-dist` connection pool by @trxcllnt in https://github.com/NVIDIA/cudf/pull/20264 +* Run polars tests with the streaming and in-memory executors by @Matt711 in https://github.com/NVIDIA/cudf/pull/19354 +* Move and rename ``ScanPartitionPlan`` by @rjzamora in https://github.com/NVIDIA/cudf/pull/20248 +* Unpin DuckDB and Ibis in cudf.pandas thirdparty tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/20269 +* Add pylibcudf to pre-commit linting and fix outstanding errors by @vyasr in https://github.com/NVIDIA/cudf/pull/20250 +* Update ``ConfigOptions`` for rapidsmpf-streaming integration by @rjzamora in https://github.com/NVIDIA/cudf/pull/20252 +* Handle unordered grouped windows properly for null filling and cum sums by @Matt711 in https://github.com/NVIDIA/cudf/pull/20275 +* Add more type annotations to cudf/core/column subclasses by @mroeschke in https://github.com/NVIDIA/cudf/pull/20277 +* Remove extraneous host_memory_resource include by @bdice in https://github.com/NVIDIA/cudf/pull/20284 +* Add `MultiIndex.dtypes` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20279 +* Skip mypy in pre-commit.ci by @bdice in https://github.com/NVIDIA/cudf/pull/20286 +* Make ColumnBase.deserialize construct via pylibcudf by @mroeschke in https://github.com/NVIDIA/cudf/pull/20142 +* Add numpy to the mypy pre-commit environment by @vyasr in https://github.com/NVIDIA/cudf/pull/20282 +* Add ability to set the source_info of parquet_reader_options by @wence- in https://github.com/NVIDIA/cudf/pull/20253 +* Add more Python type annotations to `cudf/core` by @mroeschke in https://github.com/NVIDIA/cudf/pull/20287 +* Use main in RAPIDS_BRANCH by @bdice in https://github.com/NVIDIA/cudf/pull/20312 +* Move "All rights reserved" statements to copyright line by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/20313 +* Add `inferred_type` and missing `IntervalIndex` properties by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20294 +* Avoid unseeded, random data generation in cuDF classic tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/20319 +* Improve hash-based groupby aggregation: direct write to the dense output columns whenever possible by @ttnghia in https://github.com/NVIDIA/cudf/pull/19764 +* Avoid accessing range values in cudf::strings::contains_re logic by @davidwendt in https://github.com/NVIDIA/cudf/pull/20122 +* Migrate mixed join to use the multiset data structure by @PointKernel in https://github.com/NVIDIA/cudf/pull/19989 +* Add benchmark for strings cast to/from integer APIs by @davidwendt in https://github.com/NVIDIA/cudf/pull/20247 +* Use main shared-workflows branch by @bdice in https://github.com/NVIDIA/cudf/pull/20324 +* Use the thread pool for Parquet metadata processing by @vuule in https://github.com/NVIDIA/cudf/pull/20263 +* Add `.dt.day_of_week` and `.dt.daysinmonth` by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20298 +* Avoid Column materialization in RangeIndex.nans_to_nulls by @mroeschke in https://github.com/NVIDIA/cudf/pull/20331 +* Update the code to be compatible with the new cuco stream-ordered allocator by @PointKernel in https://github.com/NVIDIA/cudf/pull/20258 +* Deprecate Series.data by @mroeschke in https://github.com/NVIDIA/cudf/pull/20281 +* Align cudf Python's Column constructors by @mroeschke in https://github.com/NVIDIA/cudf/pull/20233 +* Make type annotations of ColumnBase.set_mask stricter by @mroeschke in https://github.com/NVIDIA/cudf/pull/20261 +* Make type annotations of ColumnBase.find_and_replace stricter by @mroeschke in https://github.com/NVIDIA/cudf/pull/20259 +* Make type annotations of ColumnBase.apply_boolean_mask stricter by @mroeschke in https://github.com/NVIDIA/cudf/pull/20262 +* Skip Python LZ4 tests when nvCOMP is disabled by @vuule in https://github.com/NVIDIA/cudf/pull/20293 +* Move cudf/io/nvcomp_adapter.hpp to cudf/io/detail by @davidwendt in https://github.com/NVIDIA/cudf/pull/20327 +* Add context to IR.do_evaluate by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20322 +* Update mypy `# type: ignore` comments according to stricter mypy configs by @mroeschke in https://github.com/NVIDIA/cudf/pull/20272 +* Remove duplicated enforce null consistency code by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20342 +* Use SPDX for all copyright headers by @KyleFromNVIDIA in https://github.com/NVIDIA/cudf/pull/20321 +* Add more type annotations to `cudf/core/series.py` by @mroeschke in https://github.com/NVIDIA/cudf/pull/20304 +* Remove/Replace uses of numba.cuda arrays in pytest benchmarks and tests by @mroeschke in https://github.com/NVIDIA/cudf/pull/20359 +* Add duckdb pdsh query queries by @Matt711 in https://github.com/NVIDIA/cudf/pull/20257 +* Use stream in cudf_polars.DataFrame.to_polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20323 +* Add `join_streams` to pylibcudf API by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20316 +* Use CUDA streams in all pylibcudf calls made by cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20291 +* Add cudf/io/config_utils.hpp to doxygen by @davidwendt in https://github.com/NVIDIA/cudf/pull/20329 +* Test coverage for parallel metadata parsing by @vuule in https://github.com/NVIDIA/cudf/pull/20334 +* Support serializing more polars types by @Matt711 in https://github.com/NVIDIA/cudf/pull/20347 +* Add CUDAStreamPolicy to cudf-polars configuration by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20366 +* Unskip cudf-polars groupby test by @Matt711 in https://github.com/NVIDIA/cudf/pull/20406 +* Deprecate pylibcudf interop arrow APIs by @Matt711 in https://github.com/NVIDIA/cudf/pull/20405 +* Get rid of the hashing helper header by @PointKernel in https://github.com/NVIDIA/cudf/pull/20360 +* Minor cleanup and fixes for libcudf generate_input.cu by @davidwendt in https://github.com/NVIDIA/cudf/pull/20363 +* Ignore assert_produces_warning and shares_memory pandas unit tests for cudf.pandas by @mroeschke in https://github.com/NVIDIA/cudf/pull/20434 +* Short circut RangeIndex.append for length 0 input, proxy private attribute by @mroeschke in https://github.com/NVIDIA/cudf/pull/20442 +* Mark DataFrame.insert as _external_only_api by @Copilot in https://github.com/NVIDIA/cudf/pull/20403 +* Deprecate `get_current_device_resource` in favor of `get_current_device_resource_ref` by @PointKernel in https://github.com/NVIDIA/cudf/pull/20386 +* Promote `JoinNoneValue` to public as `JoinNoMatch` for clear non-match Join semantics by @PointKernel in https://github.com/NVIDIA/cudf/pull/20440 +* Remove duplicate entries in NODEIDS_THAT_FAIL_WITH_CUDF_PANDAS by @mroeschke in https://github.com/NVIDIA/cudf/pull/20447 +* Use the thread pool in the compact protocol reader by @vuule in https://github.com/NVIDIA/cudf/pull/20417 +* Update README.md generalizing all cuDF components by @mroeschke in https://github.com/NVIDIA/cudf/pull/20357 +* Skip TestDatetimelikeCoercion pandas tests that assert ._value identity by @mroeschke in https://github.com/NVIDIA/cudf/pull/20459 +* Add PSDH Q2-9 for cudf.pandas by @mroeschke in https://github.com/NVIDIA/cudf/pull/20418 +* Add s3fs to `test_cudf_python` common dependencies by @trxcllnt in https://github.com/NVIDIA/cudf/pull/20473 +* Use public pandas APIs in StringColumn.to_pandas by @mroeschke in https://github.com/NVIDIA/cudf/pull/20474 +* Expose java GatherMap internals and add toString to AST by @revans2 in https://github.com/NVIDIA/cudf/pull/20483 +* Add create_ascii_string_column to the libcudf benchmark data generator by @davidwendt in https://github.com/NVIDIA/cudf/pull/20354 +* Skip more pandas unit tests that tests BlockManager, private sparse types by @mroeschke in https://github.com/NVIDIA/cudf/pull/20489 +* Add boto3/botocore/aiobotocore to common test dependencies by @trxcllnt in https://github.com/NVIDIA/cudf/pull/20490 +* Use a lower bound when estimating the partial file-size by @rjzamora in https://github.com/NVIDIA/cudf/pull/20193 +* Performance improvement for nvtext::edit_distance for long strings by @davidwendt in https://github.com/NVIDIA/cudf/pull/20268 +* Add MemoryResourceConfig to cudf-polars config by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20042 +* Improve project automation by @vyasr in https://github.com/NVIDIA/cudf/pull/20523 +* Fuse simple streaming reductions in cudf-polars by @rjzamora in https://github.com/NVIDIA/cudf/pull/18757 +* Migrate to new CCCL memory resource interface by @bdice in https://github.com/NVIDIA/cudf/pull/20513 +* Add empty input gtest for cudf::transform by @davidwendt in https://github.com/NVIDIA/cudf/pull/20505 +* Rework internal json headers to allow converting gtests files from .cu to .cpp by @davidwendt in https://github.com/NVIDIA/cudf/pull/20491 +* Set continue on error in the cudf-polars-rapidsmpf nightly CI job by @Matt711 in https://github.com/NVIDIA/cudf/pull/20550 +* Permanently back cuDF column by a pylibcudf.Column by @mroeschke in https://github.com/NVIDIA/cudf/pull/20306 +* Skip flaky upstream polars rolling test by @Matt711 in https://github.com/NVIDIA/cudf/pull/20552 +* Accelerate data page mask computation on device by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20280 +* Change default rapidsmpf stream policy to 'pool' by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20527 +* Increase gtests coverage for cudf::strings::like patterns by @davidwendt in https://github.com/NVIDIA/cudf/pull/20348 +* Add cuda::std::span operator to cudf::column_view by @davidwendt in https://github.com/NVIDIA/cudf/pull/20541 +* Update ArrowStringView compare benchmark for gather by @davidwendt in https://github.com/NVIDIA/cudf/pull/19935 +* Add pytest stubs and remove ujson usage by @vyasr in https://github.com/NVIDIA/cudf/pull/20560 +* Skip arrow array constructor tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/20579 +* Add Polars to mypy environment and fix errors by @vyasr in https://github.com/NVIDIA/cudf/pull/20563 +* Ensure table chunks are unspilled and available by @madsbk in https://github.com/NVIDIA/cudf/pull/20583 +* Skip tests that assert behavior when copy-on-write is False by @Matt711 in https://github.com/NVIDIA/cudf/pull/20506 +* Pass streams through `Column.from_array`/`from_iterable_of_py` by @Matt711 in https://github.com/NVIDIA/cudf/pull/20569 +* Stop using Dtype annotation by @vyasr in https://github.com/NVIDIA/cudf/pull/20590 +* Workaround to enable running PDS-H via WebHDFS by @kingcrimsontianyu in https://github.com/NVIDIA/cudf/pull/20132 +* Update RMM includes from `` to `` by @bdice in https://github.com/NVIDIA/cudf/pull/20607 +* Stricter typing import for cudf-polars by @TomAugspurger in https://github.com/NVIDIA/cudf/pull/20614 +* Avoid the unnecessary H2H copy in the `std::vector` sink by @vuule in https://github.com/NVIDIA/cudf/pull/20602 +* Preprocessing offsets for Parquet non-dictionary string columns by @pmattione-nvidia in https://github.com/NVIDIA/cudf/pull/20430 +* Move more pandas unit tests that test private APIs by @mroeschke in https://github.com/NVIDIA/cudf/pull/20511 +* Use `.plc_column` instead of `.to_pylibcudf` in rolling, string utilties by @mroeschke in https://github.com/NVIDIA/cudf/pull/20562 +* Skip TestSetitemNADatetimeLikeDtype pandas unit tests due to private assertion by @mroeschke in https://github.com/NVIDIA/cudf/pull/20578 +* Pin Polars version <1.35 by @Matt711 in https://github.com/NVIDIA/cudf/pull/20266 +* Skip pandas unit tests in `test_old_base.py` that test private APIs by @mroeschke in https://github.com/NVIDIA/cudf/pull/20572 +* Use `.plc_column` attribute instead of `to_pylibcudf` more internally by @mroeschke in https://github.com/NVIDIA/cudf/pull/20559 +* Skip arrow-backed arithmetic tests and categorize the remaining failing tests by @Matt711 in https://github.com/NVIDIA/cudf/pull/20577 +* Fix a pytest execution that is spawned in a subprocess by @galipremsagar in https://github.com/NVIDIA/cudf/pull/20660 +* Accelerated parquet page header decoding when page index is available by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20369 +* feat: add error handling for non-existent columns in parquet reader by @gforsyth in https://github.com/NVIDIA/cudf/pull/20659 +* Optimize row mask computation for single filter column by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20335 +* Skip MultiIndex pandas unit tests testing private functionalty, `test_chaining_and_caching.py` by @mroeschke in https://github.com/NVIDIA/cudf/pull/20575 +* Address minor comments from recent hybrid scan PRs by @mhaseeb123 in https://github.com/NVIDIA/cudf/pull/20672 +* Add a timeout for the rapidsmpf test run by @vyasr in https://github.com/NVIDIA/cudf/pull/20681 +* Use `sccache-dist` build cluster for conda and wheel builds by @trxcllnt in https://github.com/NVIDIA/cudf/pull/20488 ## New Contributors -* @Copilot made their first contribution in https://github.com/rapidsai/cudf/pull/20212 -* @rockhowse made their first contribution in https://github.com/rapidsai/cudf/pull/20598 +* @Copilot made their first contribution in https://github.com/NVIDIA/cudf/pull/20212 +* @rockhowse made their first contribution in https://github.com/NVIDIA/cudf/pull/20598 -**Full Changelog**: https://github.com/rapidsai/cudf/compare/v25.12.00a...release/25.12 +**Full Changelog**: https://github.com/NVIDIA/cudf/compare/v25.12.00a...release/25.12 # cudf 25.10.00 (8 Oct 2025) ## 🚨 Breaking Changes -- Remove UCX-Py ([#19979](https://github.com/rapidsai/cudf/pull/19979)) [@pentschev](https://github.com/pentschev) -- Revert "Migrate mixed join to use multiset #19660" ([#19933](https://github.com/rapidsai/cudf/pull/19933)) [@PointKernel](https://github.com/PointKernel) -- Fill missing values in `Series/Index.values` for numeric types with np.nan by default ([#19923](https://github.com/rapidsai/cudf/pull/19923)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated `DataFrame.apply_rows`, deprecate `DataFrame.apply_chunks` and `Groupby.apply_grouped` ([#19896](https://github.com/rapidsai/cudf/pull/19896)) [@mroeschke](https://github.com/mroeschke) -- Move prefetching out of experimental and simplify the API ([#19875](https://github.com/rapidsai/cudf/pull/19875)) [@vyasr](https://github.com/vyasr) -- Add join `*_match_context` APIs to hash join ([#19835](https://github.com/rapidsai/cudf/pull/19835)) [@PointKernel](https://github.com/PointKernel) -- Vendor libnvcomp in libcudf ([#19743](https://github.com/rapidsai/cudf/pull/19743)) [@bdice](https://github.com/bdice) -- Migrate mixed join to use multiset ([#19660](https://github.com/rapidsai/cudf/pull/19660)) [@PointKernel](https://github.com/PointKernel) -- Separate row mask and page mask computation and usage ([#19537](https://github.com/rapidsai/cudf/pull/19537)) [@mhaseeb123](https://github.com/mhaseeb123) -- [FEA] Implement null-aware transforms and filters ([#19502](https://github.com/rapidsai/cudf/pull/19502)) [@lamarrr](https://github.com/lamarrr) -- Support output-type for MEDIAN/QUANTILE aggregation in cudf::reduce ([#19267](https://github.com/rapidsai/cudf/pull/19267)) [@davidwendt](https://github.com/davidwendt) +- Remove UCX-Py ([#19979](https://github.com/NVIDIA/cudf/pull/19979)) [@pentschev](https://github.com/pentschev) +- Revert "Migrate mixed join to use multiset #19660" ([#19933](https://github.com/NVIDIA/cudf/pull/19933)) [@PointKernel](https://github.com/PointKernel) +- Fill missing values in `Series/Index.values` for numeric types with np.nan by default ([#19923](https://github.com/NVIDIA/cudf/pull/19923)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated `DataFrame.apply_rows`, deprecate `DataFrame.apply_chunks` and `Groupby.apply_grouped` ([#19896](https://github.com/NVIDIA/cudf/pull/19896)) [@mroeschke](https://github.com/mroeschke) +- Move prefetching out of experimental and simplify the API ([#19875](https://github.com/NVIDIA/cudf/pull/19875)) [@vyasr](https://github.com/vyasr) +- Add join `*_match_context` APIs to hash join ([#19835](https://github.com/NVIDIA/cudf/pull/19835)) [@PointKernel](https://github.com/PointKernel) +- Vendor libnvcomp in libcudf ([#19743](https://github.com/NVIDIA/cudf/pull/19743)) [@bdice](https://github.com/bdice) +- Migrate mixed join to use multiset ([#19660](https://github.com/NVIDIA/cudf/pull/19660)) [@PointKernel](https://github.com/PointKernel) +- Separate row mask and page mask computation and usage ([#19537](https://github.com/NVIDIA/cudf/pull/19537)) [@mhaseeb123](https://github.com/mhaseeb123) +- [FEA] Implement null-aware transforms and filters ([#19502](https://github.com/NVIDIA/cudf/pull/19502)) [@lamarrr](https://github.com/lamarrr) +- Support output-type for MEDIAN/QUANTILE aggregation in cudf::reduce ([#19267](https://github.com/NVIDIA/cudf/pull/19267)) [@davidwendt](https://github.com/davidwendt) ## 🐛 Bug Fixes -- Fix edge cases in statistics collection ([#20094](https://github.com/rapidsai/cudf/pull/20094)) [@rjzamora](https://github.com/rjzamora) -- Fix multi-partition `Filter` bug ([#20075](https://github.com/rapidsai/cudf/pull/20075)) [@rjzamora](https://github.com/rjzamora) -- Fix `reindex` to fill only the reindexed values with `fill_value` ([#20063](https://github.com/rapidsai/cudf/pull/20063)) [@galipremsagar](https://github.com/galipremsagar) -- Fix arrow arrays + numpy ufunc interaction ([#20047](https://github.com/rapidsai/cudf/pull/20047)) [@galipremsagar](https://github.com/galipremsagar) -- Fix race conditions in ORC reader decimal decoding ([#20044](https://github.com/rapidsai/cudf/pull/20044)) [@vuule](https://github.com/vuule) -- Keep mr alive along with arrow tables and columns ([#20028](https://github.com/rapidsai/cudf/pull/20028)) [@vyasr](https://github.com/vyasr) -- Fix `value_counts` missing `nan` bug ([#20026](https://github.com/rapidsai/cudf/pull/20026)) [@galipremsagar](https://github.com/galipremsagar) -- Compatibility for rapidsmpf's unspill_partitions ([#20020](https://github.com/rapidsai/cudf/pull/20020)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix type metadata preservation in `shift` ([#20017](https://github.com/rapidsai/cudf/pull/20017)) [@galipremsagar](https://github.com/galipremsagar) -- Fix incorrect type propagation in dataframe assignment ([#20010](https://github.com/rapidsai/cudf/pull/20010)) [@galipremsagar](https://github.com/galipremsagar) -- Fix OOB memory read in decode_page_data_generic kernel ([#19995](https://github.com/rapidsai/cudf/pull/19995)) [@davidwendt](https://github.com/davidwendt) -- Fix data_type creation in ast::operation::instantiate ([#19994](https://github.com/rapidsai/cudf/pull/19994)) [@davidwendt](https://github.com/davidwendt) -- Skip Narwhals pandas get_dtype_backend[pyarrow] tests after ArrowDtype proxy changes ([#19992](https://github.com/rapidsai/cudf/pull/19992)) [@Matt711](https://github.com/Matt711) -- Make cudf.pandas callables usable with inspect.getfullargspec ([#19988](https://github.com/rapidsai/cudf/pull/19988)) [@mroeschke](https://github.com/mroeschke) -- Align decimal dtypes to schema after parquet IO scan ([#19974](https://github.com/rapidsai/cudf/pull/19974)) [@Matt711](https://github.com/Matt711) -- Avoid undefined numpy protocols on cudf.pandas proxy objects ([#19968](https://github.com/rapidsai/cudf/pull/19968)) [@mroeschke](https://github.com/mroeschke) -- Skip failing polars iceberg test ([#19955](https://github.com/rapidsai/cudf/pull/19955)) [@Matt711](https://github.com/Matt711) -- Revert "Migrate mixed join to use multiset #19660" ([#19933](https://github.com/rapidsai/cudf/pull/19933)) [@PointKernel](https://github.com/PointKernel) -- Define FrozenList proxy independently in cudf.pandas ([#19931](https://github.com/rapidsai/cudf/pull/19931)) [@mroeschke](https://github.com/mroeschke) -- Ignore scalars when broadcasting for horizontal string concatenation in cudf-polars ([#19893](https://github.com/rapidsai/cudf/pull/19893)) [@Matt711](https://github.com/Matt711) -- Fix is_valid_rolling_aggregation for STD aggregation ([#19888](https://github.com/rapidsai/cudf/pull/19888)) [@davidwendt](https://github.com/davidwendt) -- Fix a decompression parameter in the chunked ORC reader ([#19882](https://github.com/rapidsai/cudf/pull/19882)) [@vuule](https://github.com/vuule) -- Skip flaky stats tests pending follow up ([#19881](https://github.com/rapidsai/cudf/pull/19881)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Require list type for is_valid_aggregation and MERGE_LISTS/SETS ([#19876](https://github.com/rapidsai/cudf/pull/19876)) [@davidwendt](https://github.com/davidwendt) -- Temporary solution to ensure data-source/sink stream ordering ([#19874](https://github.com/rapidsai/cudf/pull/19874)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Check for integer overflow in cudf::strings::find_multiple ([#19867](https://github.com/rapidsai/cudf/pull/19867)) [@davidwendt](https://github.com/davidwendt) -- Fix missing stream from cudf::top_k_order ([#19866](https://github.com/rapidsai/cudf/pull/19866)) [@davidwendt](https://github.com/davidwendt) -- Disallow loc.__setitem__ with list-like indexer when list elements not in index ([#19851](https://github.com/rapidsai/cudf/pull/19851)) [@mroeschke](https://github.com/mroeschke) -- Fix .str.replace ignoring n for single character replacements ([#19848](https://github.com/rapidsai/cudf/pull/19848)) [@mroeschke](https://github.com/mroeschke) -- Fix strings::find_instance warp parallel logic ([#19845](https://github.com/rapidsai/cudf/pull/19845)) [@davidwendt](https://github.com/davidwendt) -- Add changed-files to the needs of every job that requires it ([#19830](https://github.com/rapidsai/cudf/pull/19830)) [@Matt711](https://github.com/Matt711) -- xfail polars `decimal(precision=None)` test ([#19821](https://github.com/rapidsai/cudf/pull/19821)) [@Matt711](https://github.com/Matt711) -- Fix empty column returned by cudf::from_arrow_stream_column ([#19812](https://github.com/rapidsai/cudf/pull/19812)) [@davidwendt](https://github.com/davidwendt) -- Filter pandas warning in dask_cudf test ([#19808](https://github.com/rapidsai/cudf/pull/19808)) [@TomAugspurger](https://github.com/TomAugspurger) -- Update identify_stream_usage CUDA runtime hooks to CUDA 13 ([#19807](https://github.com/rapidsai/cudf/pull/19807)) [@robertmaynard](https://github.com/robertmaynard) -- When bundling `libnvcomp.so.X` only append the major version value ([#19786](https://github.com/rapidsai/cudf/pull/19786)) [@robertmaynard](https://github.com/robertmaynard) -- Improvements to `pylibcudf.from_iterable_of_py` ([#19781](https://github.com/rapidsai/cudf/pull/19781)) [@Matt711](https://github.com/Matt711) -- Avoid using multiple `Cache` nodes with the same hash ([#19769](https://github.com/rapidsai/cudf/pull/19769)) [@rjzamora](https://github.com/rjzamora) -- Fix window var() test failures from float rounding ([#19761](https://github.com/rapidsai/cudf/pull/19761)) [@Matt711](https://github.com/Matt711) -- Use `is_compressed` field from Parquet V2 data page headers to determine if they are compressed ([#19755](https://github.com/rapidsai/cudf/pull/19755)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix bug in `eval` function with `nvtx-0.2.11` ([#19754](https://github.com/rapidsai/cudf/pull/19754)) [@galipremsagar](https://github.com/galipremsagar) -- Fix ndsh benchmarks nvtx range usage ([#19753](https://github.com/rapidsai/cudf/pull/19753)) [@davidwendt](https://github.com/davidwendt) -- Support `nan` in non-floating point column in cudf-polars ([#19742](https://github.com/rapidsai/cudf/pull/19742)) [@Matt711](https://github.com/Matt711) -- Fix filter call in benchmark ([#19732](https://github.com/rapidsai/cudf/pull/19732)) [@vyasr](https://github.com/vyasr) -- Suppress NVRTC warning from stdint.h ([#19712](https://github.com/rapidsai/cudf/pull/19712)) [@davidwendt](https://github.com/davidwendt) -- Correctly decode boolean lists in chunked parquet reader ([#19707](https://github.com/rapidsai/cudf/pull/19707)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add new xfails for xarray release ([#19705](https://github.com/rapidsai/cudf/pull/19705)) [@vyasr](https://github.com/vyasr) -- Fix "--executor" pytest parameter for cudf-polars ([#19703](https://github.com/rapidsai/cudf/pull/19703)) [@rjzamora](https://github.com/rjzamora) -- Match polars semantics for rolling-sum with all-null windows (non-empty) ([#19680](https://github.com/rapidsai/cudf/pull/19680)) [@Matt711](https://github.com/Matt711) -- [BUG] Set `query_set` arg when validating/running cudf-polars PDS-DS benchmarks ([#19674](https://github.com/rapidsai/cudf/pull/19674)) [@Matt711](https://github.com/Matt711) -- Fix `group_by().agg()` on non-aggregatable dtypes ([#19669](https://github.com/rapidsai/cudf/pull/19669)) [@Matt711](https://github.com/Matt711) -- Fix broken links in 10min notebook ([#19665](https://github.com/rapidsai/cudf/pull/19665)) [@Matt711](https://github.com/Matt711) -- Skip managed memory test if managed memory not supported in cudf-polars ([#19653](https://github.com/rapidsai/cudf/pull/19653)) [@Matt711](https://github.com/Matt711) -- Fix integer overflow in warp-per-row grid calculation ([#19638](https://github.com/rapidsai/cudf/pull/19638)) [@davidwendt](https://github.com/davidwendt) -- Propagate exceptions thrown in async IO operations ([#19628](https://github.com/rapidsai/cudf/pull/19628)) [@vuule](https://github.com/vuule) -- Make `DataFrame.dtypes` not fallback to CPU always ([#19627](https://github.com/rapidsai/cudf/pull/19627)) [@galipremsagar](https://github.com/galipremsagar) -- Set scalar to valid in range_window_bounds unbounded/current_row ([#19622](https://github.com/rapidsai/cudf/pull/19622)) [@davidwendt](https://github.com/davidwendt) -- Enable data page mask computation for nullable `list` and `struct` columns ([#19617](https://github.com/rapidsai/cudf/pull/19617)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix cudf::sequence() to throw exception for invalid scalar inputs ([#19612](https://github.com/rapidsai/cudf/pull/19612)) [@davidwendt](https://github.com/davidwendt) -- Fix uninitialized variable and misaligned write in parquet generic decoder ([#19601](https://github.com/rapidsai/cudf/pull/19601)) [@mhaseeb123](https://github.com/mhaseeb123) -- Compatibility with rapidsmpf 25.10.0 ([#19591](https://github.com/rapidsai/cudf/pull/19591)) [@TomAugspurger](https://github.com/TomAugspurger) -- Avoid querying device memory on systems without it in dask-cudf ([#19577](https://github.com/rapidsai/cudf/pull/19577)) [@Matt711](https://github.com/Matt711) -- Avoid querying device memory on systems without it in cudf-polars benchmarks ([#19575](https://github.com/rapidsai/cudf/pull/19575)) [@Matt711](https://github.com/Matt711) -- Increase alignment requirement for parquet bloom filter to 256 ([#19573](https://github.com/rapidsai/cudf/pull/19573)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix strftime with non-exact %a, %A, %b, %B ([#19570](https://github.com/rapidsai/cudf/pull/19570)) [@mroeschke](https://github.com/mroeschke) -- Fix OOB memcheck error in group_rank_to_percentage utility ([#19567](https://github.com/rapidsai/cudf/pull/19567)) [@davidwendt](https://github.com/davidwendt) -- Fix logic for number of unique values generated by data profile in benchmarks ([#19540](https://github.com/rapidsai/cudf/pull/19540)) [@shrshi](https://github.com/shrshi) -- Fix contiguous-split nvbench cmake build ([#19534](https://github.com/rapidsai/cudf/pull/19534)) [@davidwendt](https://github.com/davidwendt) -- Fix value counts expression when the column has nulls ([#19524](https://github.com/rapidsai/cudf/pull/19524)) [@Matt711](https://github.com/Matt711) -- Prefer `Column.astype` over `plc.unary.cast` in the fill null unary function expression ([#19479](https://github.com/rapidsai/cudf/pull/19479)) [@Matt711](https://github.com/Matt711) -- Fix missing return in StringFunction.Strptime strict=True path ([#19464](https://github.com/rapidsai/cudf/pull/19464)) [@Matt711](https://github.com/Matt711) -- Make dividing a boolean column return f64 dtype in cudf-polars ([#19443](https://github.com/rapidsai/cudf/pull/19443)) [@Matt711](https://github.com/Matt711) -- branch-25.10-merge-branch-25.08 ([#19429](https://github.com/rapidsai/cudf/pull/19429)) [@davidwendt](https://github.com/davidwendt) -- Replace sprintf with std::format in libcudf parquet tests ([#19364](https://github.com/rapidsai/cudf/pull/19364)) [@davidwendt](https://github.com/davidwendt) +- Fix edge cases in statistics collection ([#20094](https://github.com/NVIDIA/cudf/pull/20094)) [@rjzamora](https://github.com/rjzamora) +- Fix multi-partition `Filter` bug ([#20075](https://github.com/NVIDIA/cudf/pull/20075)) [@rjzamora](https://github.com/rjzamora) +- Fix `reindex` to fill only the reindexed values with `fill_value` ([#20063](https://github.com/NVIDIA/cudf/pull/20063)) [@galipremsagar](https://github.com/galipremsagar) +- Fix arrow arrays + numpy ufunc interaction ([#20047](https://github.com/NVIDIA/cudf/pull/20047)) [@galipremsagar](https://github.com/galipremsagar) +- Fix race conditions in ORC reader decimal decoding ([#20044](https://github.com/NVIDIA/cudf/pull/20044)) [@vuule](https://github.com/vuule) +- Keep mr alive along with arrow tables and columns ([#20028](https://github.com/NVIDIA/cudf/pull/20028)) [@vyasr](https://github.com/vyasr) +- Fix `value_counts` missing `nan` bug ([#20026](https://github.com/NVIDIA/cudf/pull/20026)) [@galipremsagar](https://github.com/galipremsagar) +- Compatibility for rapidsmpf's unspill_partitions ([#20020](https://github.com/NVIDIA/cudf/pull/20020)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix type metadata preservation in `shift` ([#20017](https://github.com/NVIDIA/cudf/pull/20017)) [@galipremsagar](https://github.com/galipremsagar) +- Fix incorrect type propagation in dataframe assignment ([#20010](https://github.com/NVIDIA/cudf/pull/20010)) [@galipremsagar](https://github.com/galipremsagar) +- Fix OOB memory read in decode_page_data_generic kernel ([#19995](https://github.com/NVIDIA/cudf/pull/19995)) [@davidwendt](https://github.com/davidwendt) +- Fix data_type creation in ast::operation::instantiate ([#19994](https://github.com/NVIDIA/cudf/pull/19994)) [@davidwendt](https://github.com/davidwendt) +- Skip Narwhals pandas get_dtype_backend[pyarrow] tests after ArrowDtype proxy changes ([#19992](https://github.com/NVIDIA/cudf/pull/19992)) [@Matt711](https://github.com/Matt711) +- Make cudf.pandas callables usable with inspect.getfullargspec ([#19988](https://github.com/NVIDIA/cudf/pull/19988)) [@mroeschke](https://github.com/mroeschke) +- Align decimal dtypes to schema after parquet IO scan ([#19974](https://github.com/NVIDIA/cudf/pull/19974)) [@Matt711](https://github.com/Matt711) +- Avoid undefined numpy protocols on cudf.pandas proxy objects ([#19968](https://github.com/NVIDIA/cudf/pull/19968)) [@mroeschke](https://github.com/mroeschke) +- Skip failing polars iceberg test ([#19955](https://github.com/NVIDIA/cudf/pull/19955)) [@Matt711](https://github.com/Matt711) +- Revert "Migrate mixed join to use multiset #19660" ([#19933](https://github.com/NVIDIA/cudf/pull/19933)) [@PointKernel](https://github.com/PointKernel) +- Define FrozenList proxy independently in cudf.pandas ([#19931](https://github.com/NVIDIA/cudf/pull/19931)) [@mroeschke](https://github.com/mroeschke) +- Ignore scalars when broadcasting for horizontal string concatenation in cudf-polars ([#19893](https://github.com/NVIDIA/cudf/pull/19893)) [@Matt711](https://github.com/Matt711) +- Fix is_valid_rolling_aggregation for STD aggregation ([#19888](https://github.com/NVIDIA/cudf/pull/19888)) [@davidwendt](https://github.com/davidwendt) +- Fix a decompression parameter in the chunked ORC reader ([#19882](https://github.com/NVIDIA/cudf/pull/19882)) [@vuule](https://github.com/vuule) +- Skip flaky stats tests pending follow up ([#19881](https://github.com/NVIDIA/cudf/pull/19881)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Require list type for is_valid_aggregation and MERGE_LISTS/SETS ([#19876](https://github.com/NVIDIA/cudf/pull/19876)) [@davidwendt](https://github.com/davidwendt) +- Temporary solution to ensure data-source/sink stream ordering ([#19874](https://github.com/NVIDIA/cudf/pull/19874)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Check for integer overflow in cudf::strings::find_multiple ([#19867](https://github.com/NVIDIA/cudf/pull/19867)) [@davidwendt](https://github.com/davidwendt) +- Fix missing stream from cudf::top_k_order ([#19866](https://github.com/NVIDIA/cudf/pull/19866)) [@davidwendt](https://github.com/davidwendt) +- Disallow loc.__setitem__ with list-like indexer when list elements not in index ([#19851](https://github.com/NVIDIA/cudf/pull/19851)) [@mroeschke](https://github.com/mroeschke) +- Fix .str.replace ignoring n for single character replacements ([#19848](https://github.com/NVIDIA/cudf/pull/19848)) [@mroeschke](https://github.com/mroeschke) +- Fix strings::find_instance warp parallel logic ([#19845](https://github.com/NVIDIA/cudf/pull/19845)) [@davidwendt](https://github.com/davidwendt) +- Add changed-files to the needs of every job that requires it ([#19830](https://github.com/NVIDIA/cudf/pull/19830)) [@Matt711](https://github.com/Matt711) +- xfail polars `decimal(precision=None)` test ([#19821](https://github.com/NVIDIA/cudf/pull/19821)) [@Matt711](https://github.com/Matt711) +- Fix empty column returned by cudf::from_arrow_stream_column ([#19812](https://github.com/NVIDIA/cudf/pull/19812)) [@davidwendt](https://github.com/davidwendt) +- Filter pandas warning in dask_cudf test ([#19808](https://github.com/NVIDIA/cudf/pull/19808)) [@TomAugspurger](https://github.com/TomAugspurger) +- Update identify_stream_usage CUDA runtime hooks to CUDA 13 ([#19807](https://github.com/NVIDIA/cudf/pull/19807)) [@robertmaynard](https://github.com/robertmaynard) +- When bundling `libnvcomp.so.X` only append the major version value ([#19786](https://github.com/NVIDIA/cudf/pull/19786)) [@robertmaynard](https://github.com/robertmaynard) +- Improvements to `pylibcudf.from_iterable_of_py` ([#19781](https://github.com/NVIDIA/cudf/pull/19781)) [@Matt711](https://github.com/Matt711) +- Avoid using multiple `Cache` nodes with the same hash ([#19769](https://github.com/NVIDIA/cudf/pull/19769)) [@rjzamora](https://github.com/rjzamora) +- Fix window var() test failures from float rounding ([#19761](https://github.com/NVIDIA/cudf/pull/19761)) [@Matt711](https://github.com/Matt711) +- Use `is_compressed` field from Parquet V2 data page headers to determine if they are compressed ([#19755](https://github.com/NVIDIA/cudf/pull/19755)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix bug in `eval` function with `nvtx-0.2.11` ([#19754](https://github.com/NVIDIA/cudf/pull/19754)) [@galipremsagar](https://github.com/galipremsagar) +- Fix ndsh benchmarks nvtx range usage ([#19753](https://github.com/NVIDIA/cudf/pull/19753)) [@davidwendt](https://github.com/davidwendt) +- Support `nan` in non-floating point column in cudf-polars ([#19742](https://github.com/NVIDIA/cudf/pull/19742)) [@Matt711](https://github.com/Matt711) +- Fix filter call in benchmark ([#19732](https://github.com/NVIDIA/cudf/pull/19732)) [@vyasr](https://github.com/vyasr) +- Suppress NVRTC warning from stdint.h ([#19712](https://github.com/NVIDIA/cudf/pull/19712)) [@davidwendt](https://github.com/davidwendt) +- Correctly decode boolean lists in chunked parquet reader ([#19707](https://github.com/NVIDIA/cudf/pull/19707)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add new xfails for xarray release ([#19705](https://github.com/NVIDIA/cudf/pull/19705)) [@vyasr](https://github.com/vyasr) +- Fix "--executor" pytest parameter for cudf-polars ([#19703](https://github.com/NVIDIA/cudf/pull/19703)) [@rjzamora](https://github.com/rjzamora) +- Match polars semantics for rolling-sum with all-null windows (non-empty) ([#19680](https://github.com/NVIDIA/cudf/pull/19680)) [@Matt711](https://github.com/Matt711) +- [BUG] Set `query_set` arg when validating/running cudf-polars PDS-DS benchmarks ([#19674](https://github.com/NVIDIA/cudf/pull/19674)) [@Matt711](https://github.com/Matt711) +- Fix `group_by().agg()` on non-aggregatable dtypes ([#19669](https://github.com/NVIDIA/cudf/pull/19669)) [@Matt711](https://github.com/Matt711) +- Fix broken links in 10min notebook ([#19665](https://github.com/NVIDIA/cudf/pull/19665)) [@Matt711](https://github.com/Matt711) +- Skip managed memory test if managed memory not supported in cudf-polars ([#19653](https://github.com/NVIDIA/cudf/pull/19653)) [@Matt711](https://github.com/Matt711) +- Fix integer overflow in warp-per-row grid calculation ([#19638](https://github.com/NVIDIA/cudf/pull/19638)) [@davidwendt](https://github.com/davidwendt) +- Propagate exceptions thrown in async IO operations ([#19628](https://github.com/NVIDIA/cudf/pull/19628)) [@vuule](https://github.com/vuule) +- Make `DataFrame.dtypes` not fallback to CPU always ([#19627](https://github.com/NVIDIA/cudf/pull/19627)) [@galipremsagar](https://github.com/galipremsagar) +- Set scalar to valid in range_window_bounds unbounded/current_row ([#19622](https://github.com/NVIDIA/cudf/pull/19622)) [@davidwendt](https://github.com/davidwendt) +- Enable data page mask computation for nullable `list` and `struct` columns ([#19617](https://github.com/NVIDIA/cudf/pull/19617)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix cudf::sequence() to throw exception for invalid scalar inputs ([#19612](https://github.com/NVIDIA/cudf/pull/19612)) [@davidwendt](https://github.com/davidwendt) +- Fix uninitialized variable and misaligned write in parquet generic decoder ([#19601](https://github.com/NVIDIA/cudf/pull/19601)) [@mhaseeb123](https://github.com/mhaseeb123) +- Compatibility with rapidsmpf 25.10.0 ([#19591](https://github.com/NVIDIA/cudf/pull/19591)) [@TomAugspurger](https://github.com/TomAugspurger) +- Avoid querying device memory on systems without it in dask-cudf ([#19577](https://github.com/NVIDIA/cudf/pull/19577)) [@Matt711](https://github.com/Matt711) +- Avoid querying device memory on systems without it in cudf-polars benchmarks ([#19575](https://github.com/NVIDIA/cudf/pull/19575)) [@Matt711](https://github.com/Matt711) +- Increase alignment requirement for parquet bloom filter to 256 ([#19573](https://github.com/NVIDIA/cudf/pull/19573)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix strftime with non-exact %a, %A, %b, %B ([#19570](https://github.com/NVIDIA/cudf/pull/19570)) [@mroeschke](https://github.com/mroeschke) +- Fix OOB memcheck error in group_rank_to_percentage utility ([#19567](https://github.com/NVIDIA/cudf/pull/19567)) [@davidwendt](https://github.com/davidwendt) +- Fix logic for number of unique values generated by data profile in benchmarks ([#19540](https://github.com/NVIDIA/cudf/pull/19540)) [@shrshi](https://github.com/shrshi) +- Fix contiguous-split nvbench cmake build ([#19534](https://github.com/NVIDIA/cudf/pull/19534)) [@davidwendt](https://github.com/davidwendt) +- Fix value counts expression when the column has nulls ([#19524](https://github.com/NVIDIA/cudf/pull/19524)) [@Matt711](https://github.com/Matt711) +- Prefer `Column.astype` over `plc.unary.cast` in the fill null unary function expression ([#19479](https://github.com/NVIDIA/cudf/pull/19479)) [@Matt711](https://github.com/Matt711) +- Fix missing return in StringFunction.Strptime strict=True path ([#19464](https://github.com/NVIDIA/cudf/pull/19464)) [@Matt711](https://github.com/Matt711) +- Make dividing a boolean column return f64 dtype in cudf-polars ([#19443](https://github.com/NVIDIA/cudf/pull/19443)) [@Matt711](https://github.com/Matt711) +- branch-25.10-merge-branch-25.08 ([#19429](https://github.com/NVIDIA/cudf/pull/19429)) [@davidwendt](https://github.com/davidwendt) +- Replace sprintf with std::format in libcudf parquet tests ([#19364](https://github.com/NVIDIA/cudf/pull/19364)) [@davidwendt](https://github.com/davidwendt) ## 📖 Documentation -- Update missing docs ([#19925](https://github.com/rapidsai/cudf/pull/19925)) [@vyasr](https://github.com/vyasr) -- Add examples of null handling to doxygen for cudf::rank ([#19774](https://github.com/rapidsai/cudf/pull/19774)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf-polars dependency list docs ([#19750](https://github.com/rapidsai/cudf/pull/19750)) [@pentschev](https://github.com/pentschev) -- Update cuDF classic testing documention regarding testing organization ([#19745](https://github.com/rapidsai/cudf/pull/19745)) [@mroeschke](https://github.com/mroeschke) -- Improve documentation around why we need no_gc_clear on pylibcudf Scalars ([#19661](https://github.com/rapidsai/cudf/pull/19661)) [@vyasr](https://github.com/vyasr) +- Update missing docs ([#19925](https://github.com/NVIDIA/cudf/pull/19925)) [@vyasr](https://github.com/vyasr) +- Add examples of null handling to doxygen for cudf::rank ([#19774](https://github.com/NVIDIA/cudf/pull/19774)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf-polars dependency list docs ([#19750](https://github.com/NVIDIA/cudf/pull/19750)) [@pentschev](https://github.com/pentschev) +- Update cuDF classic testing documention regarding testing organization ([#19745](https://github.com/NVIDIA/cudf/pull/19745)) [@mroeschke](https://github.com/mroeschke) +- Improve documentation around why we need no_gc_clear on pylibcudf Scalars ([#19661](https://github.com/NVIDIA/cudf/pull/19661)) [@vyasr](https://github.com/vyasr) ## 🚀 New Features -- Add memory resource parameters to interop, merge, and transpose ([#20007](https://github.com/rapidsai/cudf/pull/20007)) [@vyasr](https://github.com/vyasr) -- Add mixed join benchmark with complex AST operators ([#20004](https://github.com/rapidsai/cudf/pull/20004)) [@PointKernel](https://github.com/PointKernel) -- Add memory resource arguments to join, round, and labeling ([#20001](https://github.com/rapidsai/cudf/pull/20001)) [@vyasr](https://github.com/vyasr) -- `cudf-polars` `strptime` format inference ([#19997](https://github.com/rapidsai/cudf/pull/19997)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Filter parquet row groups using byte offset bounds ([#19991](https://github.com/rapidsai/cudf/pull/19991)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add memory resource arguments to concatenate ([#19943](https://github.com/rapidsai/cudf/pull/19943)) [@vyasr](https://github.com/vyasr) -- Use column statistics to generate the physical plan in cuDF-Polars ([#19940](https://github.com/rapidsai/cudf/pull/19940)) [@rjzamora](https://github.com/rjzamora) -- Add all missing stream parameters ([#19922](https://github.com/rapidsai/cudf/pull/19922)) [@vyasr](https://github.com/vyasr) -- Remote IO support in cudf-polars ([#19921](https://github.com/rapidsai/cudf/pull/19921)) [@Matt711](https://github.com/Matt711) -- Add streams to io/timezone and io/text modules ([#19913](https://github.com/rapidsai/cudf/pull/19913)) [@vyasr](https://github.com/vyasr) -- Add stream support to all nvtext modules ([#19911](https://github.com/rapidsai/cudf/pull/19911)) [@vyasr](https://github.com/vyasr) -- Add streams to all top-level strings modules ([#19910](https://github.com/rapidsai/cudf/pull/19910)) [@vyasr](https://github.com/vyasr) -- Update strings split APIs with stream parameters ([#19909](https://github.com/rapidsai/cudf/pull/19909)) [@vyasr](https://github.com/vyasr) -- Support ordered grouped windows in cudf-polars ([#19891](https://github.com/rapidsai/cudf/pull/19891)) [@Matt711](https://github.com/Matt711) -- Add local row-count and unique-count estimates to `explain(... logical=True)` ([#19864](https://github.com/rapidsai/cudf/pull/19864)) [@rjzamora](https://github.com/rjzamora) -- Add join `*_match_context` APIs to hash join ([#19835](https://github.com/rapidsai/cudf/pull/19835)) [@PointKernel](https://github.com/PointKernel) -- Support `rank(...).over(...)` expressions in cudf-polars ([#19803](https://github.com/rapidsai/cudf/pull/19803)) [@Matt711](https://github.com/Matt711) -- Add strings to/from encoded integer APIs ([#19789](https://github.com/rapidsai/cudf/pull/19789)) [@davidwendt](https://github.com/davidwendt) -- Add to_arrow method to pylibcudf core types ([#19787](https://github.com/rapidsai/cudf/pull/19787)) [@Matt711](https://github.com/Matt711) -- Add streams to strings convert APIs ([#19780](https://github.com/rapidsai/cudf/pull/19780)) [@vyasr](https://github.com/vyasr) -- Add an option to support reading ORC timestamp column as UTC time. ([#19773](https://github.com/rapidsai/cudf/pull/19773)) [@res-life](https://github.com/res-life) -- Support null_count in groupby/rolling context ([#19739](https://github.com/rapidsai/cudf/pull/19739)) [@Matt711](https://github.com/Matt711) -- Collect join-key information in cudf-polars ([#19736](https://github.com/rapidsai/cudf/pull/19736)) [@rjzamora](https://github.com/rjzamora) -- Add count aggregation support to cudf::reduce ([#19734](https://github.com/rapidsai/cudf/pull/19734)) [@davidwendt](https://github.com/davidwendt) -- [FEA] Implement AST Expression - JIT codegen ([#19733](https://github.com/rapidsai/cudf/pull/19733)) [@lamarrr](https://github.com/lamarrr) -- Add streams to all scalar factories ([#19729](https://github.com/rapidsai/cudf/pull/19729)) [@vyasr](https://github.com/vyasr) -- Add streams to reshape ([#19728](https://github.com/rapidsai/cudf/pull/19728)) [@vyasr](https://github.com/vyasr) -- Add streams to null mask APIs ([#19727](https://github.com/rapidsai/cudf/pull/19727)) [@vyasr](https://github.com/vyasr) -- Add streams to column APIs ([#19726](https://github.com/rapidsai/cudf/pull/19726)) [@vyasr](https://github.com/vyasr) -- Construct next-gen parquet reader with pre-populated footer ([#19724](https://github.com/rapidsai/cudf/pull/19724)) [@mhaseeb123](https://github.com/mhaseeb123) -- Require `numba-cuda>=0.19.0,<0.20.0a0` ([#19711](https://github.com/rapidsai/cudf/pull/19711)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `over` expression (window mapping) in cudf-polars ([#19684](https://github.com/rapidsai/cudf/pull/19684)) [@Matt711](https://github.com/Matt711) -- Add streams support to all list APIs ([#19683](https://github.com/rapidsai/cudf/pull/19683)) [@vyasr](https://github.com/vyasr) -- [FEA] Add Filter Benchmark ([#19678](https://github.com/rapidsai/cudf/pull/19678)) [@lamarrr](https://github.com/lamarrr) -- Add streams to pylibcudf join APIs ([#19672](https://github.com/rapidsai/cudf/pull/19672)) [@vyasr](https://github.com/vyasr) -- Add streams to sorting APIs ([#19671](https://github.com/rapidsai/cudf/pull/19671)) [@vyasr](https://github.com/vyasr) -- [FEA] Remove excessive copies of JITIFY's ProgramData during JIT kernel launch ([#19667](https://github.com/rapidsai/cudf/pull/19667)) [@lamarrr](https://github.com/lamarrr) -- Add streams to hashing APIs ([#19663](https://github.com/rapidsai/cudf/pull/19663)) [@vyasr](https://github.com/vyasr) -- Use a more robust metric for sorting (de)compression tasks ([#19656](https://github.com/rapidsai/cudf/pull/19656)) [@vuule](https://github.com/vuule) -- Add streams support to datetime APIs ([#19654](https://github.com/rapidsai/cudf/pull/19654)) [@vyasr](https://github.com/vyasr) -- Add streams to stream_compaction ([#19651](https://github.com/rapidsai/cudf/pull/19651)) [@vyasr](https://github.com/vyasr) -- Enable casting `pl.Datetime` to integer types in `cudf-polars` ([#19647](https://github.com/rapidsai/cudf/pull/19647)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add Java JNI interface to get Gpu UUID ([#19646](https://github.com/rapidsai/cudf/pull/19646)) [@res-life](https://github.com/res-life) -- Add reduction with overflow detection ([#19641](https://github.com/rapidsai/cudf/pull/19641)) [@PointKernel](https://github.com/PointKernel) -- Upgrade to nvCOMP 5.0.0.6 ([#19636](https://github.com/rapidsai/cudf/pull/19636)) [@vuule](https://github.com/vuule) -- Use the nvCOMP 5.0 API to better estimate decompression memory requirements ([#19616](https://github.com/rapidsai/cudf/pull/19616)) [@vuule](https://github.com/vuule) -- Add streams to transform and unary ([#19613](https://github.com/rapidsai/cudf/pull/19613)) [@vyasr](https://github.com/vyasr) -- Add streams to all modules with 4-5 functions ([#19609](https://github.com/rapidsai/cudf/pull/19609)) [@vyasr](https://github.com/vyasr) -- Enable casting integer dtypes to `pl.Datetime` via `cudf-polars` ([#19607](https://github.com/rapidsai/cudf/pull/19607)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add fast path for Parquet reading with predicate pushdown via AST filters ([#19605](https://github.com/rapidsai/cudf/pull/19605)) [@Matt711](https://github.com/Matt711) -- Add streams to all modules with three or fewer functions ([#19600](https://github.com/rapidsai/cudf/pull/19600)) [@vyasr](https://github.com/vyasr) -- Add libcudf top_k_segmented APIs ([#19597](https://github.com/rapidsai/cudf/pull/19597)) [@davidwendt](https://github.com/davidwendt) -- Update Arrow bounds to >=15,<22 ([#19592](https://github.com/rapidsai/cudf/pull/19592)) [@bdice](https://github.com/bdice) -- Update cudf to handle CUDA 13 changes ([#19585](https://github.com/rapidsai/cudf/pull/19585)) [@robertmaynard](https://github.com/robertmaynard) -- Support hash-based workflow for `M2` groupby aggregation ([#19569](https://github.com/rapidsai/cudf/pull/19569)) [@ttnghia](https://github.com/ttnghia) -- Expose `filter` and `columns` parquet reader builder options to python ([#19566](https://github.com/rapidsai/cudf/pull/19566)) [@Matt711](https://github.com/Matt711) -- [FEA] Switch to NVIDIA's JITIFY2 ([#19561](https://github.com/rapidsai/cudf/pull/19561)) [@lamarrr](https://github.com/lamarrr) -- Add streams to all single-function modules ([#19559](https://github.com/rapidsai/cudf/pull/19559)) [@vyasr](https://github.com/vyasr) -- Add support for streams to all copying APIs. ([#19553](https://github.com/rapidsai/cudf/pull/19553)) [@vyasr](https://github.com/vyasr) -- Benchmarks comparing Arrow string formats ([#19552](https://github.com/rapidsai/cudf/pull/19552)) [@davidwendt](https://github.com/davidwendt) -- Compile `libcudf_kafka` and `cudf_kafka` with C++20 ([#19543](https://github.com/rapidsai/cudf/pull/19543)) [@vuule](https://github.com/vuule) -- RapidsMPF "single" shuffle integration ([#19530](https://github.com/rapidsai/cudf/pull/19530)) [@rjzamora](https://github.com/rjzamora) -- Make nvCOMP ZLIB (de)compression available by default ([#19528](https://github.com/rapidsai/cudf/pull/19528)) [@vuule](https://github.com/vuule) -- Implement chunking in the next-gen parquet reader ([#19526](https://github.com/rapidsai/cudf/pull/19526)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add primitive row dispatch support for semi/anti join and cudf::contains ([#19518](https://github.com/rapidsai/cudf/pull/19518)) [@PointKernel](https://github.com/PointKernel) -- Derive and use page mask at subpass level for chunked reads ([#19515](https://github.com/rapidsai/cudf/pull/19515)) [@mhaseeb123](https://github.com/mhaseeb123) -- [FEA] Implement null-aware transforms and filters ([#19502](https://github.com/rapidsai/cudf/pull/19502)) [@lamarrr](https://github.com/lamarrr) -- Add PDS-DS queries 2 through 10 to cudf-polars benchmarks ([#19488](https://github.com/rapidsai/cudf/pull/19488)) [@Matt711](https://github.com/Matt711) -- Add API to "initialize" column statistics ([#19447](https://github.com/rapidsai/cudf/pull/19447)) [@rjzamora](https://github.com/rjzamora) -- Implement top k expression in cudf-polars using `cudf::top_k` ([#19431](https://github.com/rapidsai/cudf/pull/19431)) [@Matt711](https://github.com/Matt711) -- Add hash-based SUM_WITH_OVERFLOW aggregation for INT64 values ([#19403](https://github.com/rapidsai/cudf/pull/19403)) [@PointKernel](https://github.com/PointKernel) -- Support rank expression in cudf-polars ([#19340](https://github.com/rapidsai/cudf/pull/19340)) [@Matt711](https://github.com/Matt711) -- Support fill_null with fill strategy in cudf-polars ([#19318](https://github.com/rapidsai/cudf/pull/19318)) [@Matt711](https://github.com/Matt711) -- Support output-type for MEDIAN/QUANTILE aggregation in cudf::reduce ([#19267](https://github.com/rapidsai/cudf/pull/19267)) [@davidwendt](https://github.com/davidwendt) -- Support ternary expression inside groupby/rolling context ([#19242](https://github.com/rapidsai/cudf/pull/19242)) [@Matt711](https://github.com/Matt711) -- Experimental API to read a parquet table, build a custom index column, and apply roaring bitmap deletion vector ([#19237](https://github.com/rapidsai/cudf/pull/19237)) [@mhaseeb123](https://github.com/mhaseeb123) -- Support `cudf-polars` `str.zfill` ([#19081](https://github.com/rapidsai/cudf/pull/19081)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- [FEA] Add chunked Parquet sink support using the libcudf writer ([#19015](https://github.com/rapidsai/cudf/pull/19015)) [@Matt711](https://github.com/Matt711) -- Add multi-column support for primitive row operator dispatch ([#18940](https://github.com/rapidsai/cudf/pull/18940)) [@tgujar](https://github.com/tgujar) +- Add memory resource parameters to interop, merge, and transpose ([#20007](https://github.com/NVIDIA/cudf/pull/20007)) [@vyasr](https://github.com/vyasr) +- Add mixed join benchmark with complex AST operators ([#20004](https://github.com/NVIDIA/cudf/pull/20004)) [@PointKernel](https://github.com/PointKernel) +- Add memory resource arguments to join, round, and labeling ([#20001](https://github.com/NVIDIA/cudf/pull/20001)) [@vyasr](https://github.com/vyasr) +- `cudf-polars` `strptime` format inference ([#19997](https://github.com/NVIDIA/cudf/pull/19997)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Filter parquet row groups using byte offset bounds ([#19991](https://github.com/NVIDIA/cudf/pull/19991)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add memory resource arguments to concatenate ([#19943](https://github.com/NVIDIA/cudf/pull/19943)) [@vyasr](https://github.com/vyasr) +- Use column statistics to generate the physical plan in cuDF-Polars ([#19940](https://github.com/NVIDIA/cudf/pull/19940)) [@rjzamora](https://github.com/rjzamora) +- Add all missing stream parameters ([#19922](https://github.com/NVIDIA/cudf/pull/19922)) [@vyasr](https://github.com/vyasr) +- Remote IO support in cudf-polars ([#19921](https://github.com/NVIDIA/cudf/pull/19921)) [@Matt711](https://github.com/Matt711) +- Add streams to io/timezone and io/text modules ([#19913](https://github.com/NVIDIA/cudf/pull/19913)) [@vyasr](https://github.com/vyasr) +- Add stream support to all nvtext modules ([#19911](https://github.com/NVIDIA/cudf/pull/19911)) [@vyasr](https://github.com/vyasr) +- Add streams to all top-level strings modules ([#19910](https://github.com/NVIDIA/cudf/pull/19910)) [@vyasr](https://github.com/vyasr) +- Update strings split APIs with stream parameters ([#19909](https://github.com/NVIDIA/cudf/pull/19909)) [@vyasr](https://github.com/vyasr) +- Support ordered grouped windows in cudf-polars ([#19891](https://github.com/NVIDIA/cudf/pull/19891)) [@Matt711](https://github.com/Matt711) +- Add local row-count and unique-count estimates to `explain(... logical=True)` ([#19864](https://github.com/NVIDIA/cudf/pull/19864)) [@rjzamora](https://github.com/rjzamora) +- Add join `*_match_context` APIs to hash join ([#19835](https://github.com/NVIDIA/cudf/pull/19835)) [@PointKernel](https://github.com/PointKernel) +- Support `rank(...).over(...)` expressions in cudf-polars ([#19803](https://github.com/NVIDIA/cudf/pull/19803)) [@Matt711](https://github.com/Matt711) +- Add strings to/from encoded integer APIs ([#19789](https://github.com/NVIDIA/cudf/pull/19789)) [@davidwendt](https://github.com/davidwendt) +- Add to_arrow method to pylibcudf core types ([#19787](https://github.com/NVIDIA/cudf/pull/19787)) [@Matt711](https://github.com/Matt711) +- Add streams to strings convert APIs ([#19780](https://github.com/NVIDIA/cudf/pull/19780)) [@vyasr](https://github.com/vyasr) +- Add an option to support reading ORC timestamp column as UTC time. ([#19773](https://github.com/NVIDIA/cudf/pull/19773)) [@res-life](https://github.com/res-life) +- Support null_count in groupby/rolling context ([#19739](https://github.com/NVIDIA/cudf/pull/19739)) [@Matt711](https://github.com/Matt711) +- Collect join-key information in cudf-polars ([#19736](https://github.com/NVIDIA/cudf/pull/19736)) [@rjzamora](https://github.com/rjzamora) +- Add count aggregation support to cudf::reduce ([#19734](https://github.com/NVIDIA/cudf/pull/19734)) [@davidwendt](https://github.com/davidwendt) +- [FEA] Implement AST Expression - JIT codegen ([#19733](https://github.com/NVIDIA/cudf/pull/19733)) [@lamarrr](https://github.com/lamarrr) +- Add streams to all scalar factories ([#19729](https://github.com/NVIDIA/cudf/pull/19729)) [@vyasr](https://github.com/vyasr) +- Add streams to reshape ([#19728](https://github.com/NVIDIA/cudf/pull/19728)) [@vyasr](https://github.com/vyasr) +- Add streams to null mask APIs ([#19727](https://github.com/NVIDIA/cudf/pull/19727)) [@vyasr](https://github.com/vyasr) +- Add streams to column APIs ([#19726](https://github.com/NVIDIA/cudf/pull/19726)) [@vyasr](https://github.com/vyasr) +- Construct next-gen parquet reader with pre-populated footer ([#19724](https://github.com/NVIDIA/cudf/pull/19724)) [@mhaseeb123](https://github.com/mhaseeb123) +- Require `numba-cuda>=0.19.0,<0.20.0a0` ([#19711](https://github.com/NVIDIA/cudf/pull/19711)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `over` expression (window mapping) in cudf-polars ([#19684](https://github.com/NVIDIA/cudf/pull/19684)) [@Matt711](https://github.com/Matt711) +- Add streams support to all list APIs ([#19683](https://github.com/NVIDIA/cudf/pull/19683)) [@vyasr](https://github.com/vyasr) +- [FEA] Add Filter Benchmark ([#19678](https://github.com/NVIDIA/cudf/pull/19678)) [@lamarrr](https://github.com/lamarrr) +- Add streams to pylibcudf join APIs ([#19672](https://github.com/NVIDIA/cudf/pull/19672)) [@vyasr](https://github.com/vyasr) +- Add streams to sorting APIs ([#19671](https://github.com/NVIDIA/cudf/pull/19671)) [@vyasr](https://github.com/vyasr) +- [FEA] Remove excessive copies of JITIFY's ProgramData during JIT kernel launch ([#19667](https://github.com/NVIDIA/cudf/pull/19667)) [@lamarrr](https://github.com/lamarrr) +- Add streams to hashing APIs ([#19663](https://github.com/NVIDIA/cudf/pull/19663)) [@vyasr](https://github.com/vyasr) +- Use a more robust metric for sorting (de)compression tasks ([#19656](https://github.com/NVIDIA/cudf/pull/19656)) [@vuule](https://github.com/vuule) +- Add streams support to datetime APIs ([#19654](https://github.com/NVIDIA/cudf/pull/19654)) [@vyasr](https://github.com/vyasr) +- Add streams to stream_compaction ([#19651](https://github.com/NVIDIA/cudf/pull/19651)) [@vyasr](https://github.com/vyasr) +- Enable casting `pl.Datetime` to integer types in `cudf-polars` ([#19647](https://github.com/NVIDIA/cudf/pull/19647)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add Java JNI interface to get Gpu UUID ([#19646](https://github.com/NVIDIA/cudf/pull/19646)) [@res-life](https://github.com/res-life) +- Add reduction with overflow detection ([#19641](https://github.com/NVIDIA/cudf/pull/19641)) [@PointKernel](https://github.com/PointKernel) +- Upgrade to nvCOMP 5.0.0.6 ([#19636](https://github.com/NVIDIA/cudf/pull/19636)) [@vuule](https://github.com/vuule) +- Use the nvCOMP 5.0 API to better estimate decompression memory requirements ([#19616](https://github.com/NVIDIA/cudf/pull/19616)) [@vuule](https://github.com/vuule) +- Add streams to transform and unary ([#19613](https://github.com/NVIDIA/cudf/pull/19613)) [@vyasr](https://github.com/vyasr) +- Add streams to all modules with 4-5 functions ([#19609](https://github.com/NVIDIA/cudf/pull/19609)) [@vyasr](https://github.com/vyasr) +- Enable casting integer dtypes to `pl.Datetime` via `cudf-polars` ([#19607](https://github.com/NVIDIA/cudf/pull/19607)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add fast path for Parquet reading with predicate pushdown via AST filters ([#19605](https://github.com/NVIDIA/cudf/pull/19605)) [@Matt711](https://github.com/Matt711) +- Add streams to all modules with three or fewer functions ([#19600](https://github.com/NVIDIA/cudf/pull/19600)) [@vyasr](https://github.com/vyasr) +- Add libcudf top_k_segmented APIs ([#19597](https://github.com/NVIDIA/cudf/pull/19597)) [@davidwendt](https://github.com/davidwendt) +- Update Arrow bounds to >=15,<22 ([#19592](https://github.com/NVIDIA/cudf/pull/19592)) [@bdice](https://github.com/bdice) +- Update cudf to handle CUDA 13 changes ([#19585](https://github.com/NVIDIA/cudf/pull/19585)) [@robertmaynard](https://github.com/robertmaynard) +- Support hash-based workflow for `M2` groupby aggregation ([#19569](https://github.com/NVIDIA/cudf/pull/19569)) [@ttnghia](https://github.com/ttnghia) +- Expose `filter` and `columns` parquet reader builder options to python ([#19566](https://github.com/NVIDIA/cudf/pull/19566)) [@Matt711](https://github.com/Matt711) +- [FEA] Switch to NVIDIA's JITIFY2 ([#19561](https://github.com/NVIDIA/cudf/pull/19561)) [@lamarrr](https://github.com/lamarrr) +- Add streams to all single-function modules ([#19559](https://github.com/NVIDIA/cudf/pull/19559)) [@vyasr](https://github.com/vyasr) +- Add support for streams to all copying APIs. ([#19553](https://github.com/NVIDIA/cudf/pull/19553)) [@vyasr](https://github.com/vyasr) +- Benchmarks comparing Arrow string formats ([#19552](https://github.com/NVIDIA/cudf/pull/19552)) [@davidwendt](https://github.com/davidwendt) +- Compile `libcudf_kafka` and `cudf_kafka` with C++20 ([#19543](https://github.com/NVIDIA/cudf/pull/19543)) [@vuule](https://github.com/vuule) +- RapidsMPF "single" shuffle integration ([#19530](https://github.com/NVIDIA/cudf/pull/19530)) [@rjzamora](https://github.com/rjzamora) +- Make nvCOMP ZLIB (de)compression available by default ([#19528](https://github.com/NVIDIA/cudf/pull/19528)) [@vuule](https://github.com/vuule) +- Implement chunking in the next-gen parquet reader ([#19526](https://github.com/NVIDIA/cudf/pull/19526)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add primitive row dispatch support for semi/anti join and cudf::contains ([#19518](https://github.com/NVIDIA/cudf/pull/19518)) [@PointKernel](https://github.com/PointKernel) +- Derive and use page mask at subpass level for chunked reads ([#19515](https://github.com/NVIDIA/cudf/pull/19515)) [@mhaseeb123](https://github.com/mhaseeb123) +- [FEA] Implement null-aware transforms and filters ([#19502](https://github.com/NVIDIA/cudf/pull/19502)) [@lamarrr](https://github.com/lamarrr) +- Add PDS-DS queries 2 through 10 to cudf-polars benchmarks ([#19488](https://github.com/NVIDIA/cudf/pull/19488)) [@Matt711](https://github.com/Matt711) +- Add API to "initialize" column statistics ([#19447](https://github.com/NVIDIA/cudf/pull/19447)) [@rjzamora](https://github.com/rjzamora) +- Implement top k expression in cudf-polars using `cudf::top_k` ([#19431](https://github.com/NVIDIA/cudf/pull/19431)) [@Matt711](https://github.com/Matt711) +- Add hash-based SUM_WITH_OVERFLOW aggregation for INT64 values ([#19403](https://github.com/NVIDIA/cudf/pull/19403)) [@PointKernel](https://github.com/PointKernel) +- Support rank expression in cudf-polars ([#19340](https://github.com/NVIDIA/cudf/pull/19340)) [@Matt711](https://github.com/Matt711) +- Support fill_null with fill strategy in cudf-polars ([#19318](https://github.com/NVIDIA/cudf/pull/19318)) [@Matt711](https://github.com/Matt711) +- Support output-type for MEDIAN/QUANTILE aggregation in cudf::reduce ([#19267](https://github.com/NVIDIA/cudf/pull/19267)) [@davidwendt](https://github.com/davidwendt) +- Support ternary expression inside groupby/rolling context ([#19242](https://github.com/NVIDIA/cudf/pull/19242)) [@Matt711](https://github.com/Matt711) +- Experimental API to read a parquet table, build a custom index column, and apply roaring bitmap deletion vector ([#19237](https://github.com/NVIDIA/cudf/pull/19237)) [@mhaseeb123](https://github.com/mhaseeb123) +- Support `cudf-polars` `str.zfill` ([#19081](https://github.com/NVIDIA/cudf/pull/19081)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- [FEA] Add chunked Parquet sink support using the libcudf writer ([#19015](https://github.com/NVIDIA/cudf/pull/19015)) [@Matt711](https://github.com/Matt711) +- Add multi-column support for primitive row operator dispatch ([#18940](https://github.com/NVIDIA/cudf/pull/18940)) [@tgujar](https://github.com/tgujar) ## 🛠️ Improvements -- Fix CI failures for `pandas-2.3.3` ([#20146](https://github.com/rapidsai/cudf/pull/20146)) [@galipremsagar](https://github.com/galipremsagar) -- Skip passing failures for latest `numexpr` version ([#20092](https://github.com/rapidsai/cudf/pull/20092)) [@galipremsagar](https://github.com/galipremsagar) -- Empty commit to trigger a build ([#20084](https://github.com/rapidsai/cudf/pull/20084)) [@msarahan](https://github.com/msarahan) -- Update the reason to skip for parquet bloom filter test ([#20043](https://github.com/rapidsai/cudf/pull/20043)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove test_scan_hf_url_raises ([#20035](https://github.com/rapidsai/cudf/pull/20035)) [@mroeschke](https://github.com/mroeschke) -- xfail(strict=False) test_scan_hf_url_raises due to rate limiting ([#20027](https://github.com/rapidsai/cudf/pull/20027)) [@mroeschke](https://github.com/mroeschke) -- Deprecate left semi- and anti- join functional APIs ([#20014](https://github.com/rapidsai/cudf/pull/20014)) [@shrshi](https://github.com/shrshi) -- Use to_arrow methods throughout pylibcudf and cudf ([#20013](https://github.com/rapidsai/cudf/pull/20013)) [@Matt711](https://github.com/Matt711) -- Fix chunked reads of list of bools. ([#20000](https://github.com/rapidsai/cudf/pull/20000)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Raise more exceptions for invalid or unsupported cuDF arguments ([#19990](https://github.com/rapidsai/cudf/pull/19990)) [@mroeschke](https://github.com/mroeschke) -- Configure repo for automatic release notes generation ([#19984](https://github.com/rapidsai/cudf/pull/19984)) [@AyodeAwe](https://github.com/AyodeAwe) -- Pin duckdb<1.4 in test_python_narwhals ([#19982](https://github.com/rapidsai/cudf/pull/19982)) [@mroeschke](https://github.com/mroeschke) -- Default to False if `CUDA_ENABLE_NRT` isn't set in config ([#19981](https://github.com/rapidsai/cudf/pull/19981)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Remove UCX-Py ([#19979](https://github.com/rapidsai/cudf/pull/19979)) [@pentschev](https://github.com/pentschev) -- Add support for `attrs` ([#19978](https://github.com/rapidsai/cudf/pull/19978)) [@galipremsagar](https://github.com/galipremsagar) -- Run pytest-benchmarks in CI with --benchmark-disable ([#19969](https://github.com/rapidsai/cudf/pull/19969)) [@mroeschke](https://github.com/mroeschke) -- Change target type so we can test on workflows ([#19963](https://github.com/rapidsai/cudf/pull/19963)) [@vyasr](https://github.com/vyasr) -- Update to actions/labeler v5 ([#19962](https://github.com/rapidsai/cudf/pull/19962)) [@vyasr](https://github.com/vyasr) -- Revert "ci(labeler): update labeler action to [@v5" ([#19961](https://github.com/rapidsai/cudf/pull/19961)) @vyasr](https://github.com/v5" ([#19961](https://github.com/rapidsai/cudf/pull/19961)) @vyasr) -- Add `ArrowDtype` proxy class ([#19960](https://github.com/rapidsai/cudf/pull/19960)) [@galipremsagar](https://github.com/galipremsagar) -- Add missing type stub ([#19958](https://github.com/rapidsai/cudf/pull/19958)) [@vyasr](https://github.com/vyasr) -- Add missing `Styler` attributes ([#19956](https://github.com/rapidsai/cudf/pull/19956)) [@galipremsagar](https://github.com/galipremsagar) -- Allow newer CMake in Java tests ([#19949](https://github.com/rapidsai/cudf/pull/19949)) [@bdice](https://github.com/bdice) -- Make stream a required parameter for from_libcudf methods ([#19945](https://github.com/rapidsai/cudf/pull/19945)) [@vyasr](https://github.com/vyasr) -- Return False instead of NA for comparison ops against NA in cudf.pandas ([#19942](https://github.com/rapidsai/cudf/pull/19942)) [@mroeschke](https://github.com/mroeschke) -- Don't fall back in Series.describe in cudf.pandas for numeric types ([#19941](https://github.com/rapidsai/cudf/pull/19941)) [@mroeschke](https://github.com/mroeschke) -- Move groupby benchmarks to nvbench ([#19930](https://github.com/rapidsai/cudf/pull/19930)) [@davidwendt](https://github.com/davidwendt) -- Perform more input validation in cuDF classic APIs ([#19929](https://github.com/rapidsai/cudf/pull/19929)) [@mroeschke](https://github.com/mroeschke) -- update nvidia-ml-py (>=12), use cuda-toolkit wheels ([#19927](https://github.com/rapidsai/cudf/pull/19927)) [@jameslamb](https://github.com/jameslamb) -- Fill missing values in `Series/Index.values` for numeric types with np.nan by default ([#19923](https://github.com/rapidsai/cudf/pull/19923)) [@mroeschke](https://github.com/mroeschke) -- Add `rmm-release-threshold` to pdsh benchmarks CLI ([#19918](https://github.com/rapidsai/cudf/pull/19918)) [@TomAugspurger](https://github.com/TomAugspurger) -- Also use the CUDA 12 container for nightlies ([#19917](https://github.com/rapidsai/cudf/pull/19917)) [@vyasr](https://github.com/vyasr) -- Move test_binops.py to new cuDF classic directory structure ([#19914](https://github.com/rapidsai/cudf/pull/19914)) [@mroeschke](https://github.com/mroeschke) -- Eagerly load nvCOMP library in `cudf::initialize()` ([#19906](https://github.com/rapidsai/cudf/pull/19906)) [@vuule](https://github.com/vuule) -- Pin to CUDA 12 image for integration tests ([#19903](https://github.com/rapidsai/cudf/pull/19903)) [@vyasr](https://github.com/vyasr) -- Use branch-25.10 again ([#19902](https://github.com/rapidsai/cudf/pull/19902)) [@jameslamb](https://github.com/jameslamb) -- Disable test on non-default stream ([#19901](https://github.com/rapidsai/cudf/pull/19901)) [@vyasr](https://github.com/vyasr) -- Use cupy array instead of numba device array as inputs to jit routines ([#19897](https://github.com/rapidsai/cudf/pull/19897)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated `DataFrame.apply_rows`, deprecate `DataFrame.apply_chunks` and `Groupby.apply_grouped` ([#19896](https://github.com/rapidsai/cudf/pull/19896)) [@mroeschke](https://github.com/mroeschke) -- Move test_dataframe.py to new cuDF classic directory structure ([#19890](https://github.com/rapidsai/cudf/pull/19890)) [@mroeschke](https://github.com/mroeschke) -- Make sure conftest fixture data is valid on exit ([#19889](https://github.com/rapidsai/cudf/pull/19889)) [@vyasr](https://github.com/vyasr) -- Move test_index/multiindex/indexing.py to new cuDF classic directory structure ([#19887](https://github.com/rapidsai/cudf/pull/19887)) [@mroeschke](https://github.com/mroeschke) -- [FEA] Build CUDF with CCCL 3.1.0 ([#19886](https://github.com/rapidsai/cudf/pull/19886)) [@lamarrr](https://github.com/lamarrr) -- Coalesce IO of chunks with different compression when reading Parquet files ([#19884](https://github.com/rapidsai/cudf/pull/19884)) [@vuule](https://github.com/vuule) -- Update boost version to 1.79 for JNI dockerfile ([#19883](https://github.com/rapidsai/cudf/pull/19883)) [@pxLi](https://github.com/pxLi) -- Move test_categorical/dask/serialize.py to new cuDF classic test directory structure ([#19877](https://github.com/rapidsai/cudf/pull/19877)) [@mroeschke](https://github.com/mroeschke) -- Move prefetching out of experimental and simplify the API ([#19875](https://github.com/rapidsai/cudf/pull/19875)) [@vyasr](https://github.com/vyasr) -- Remove `diff.sh` and merge diff generation into `run.sh` ([#19871](https://github.com/rapidsai/cudf/pull/19871)) [@galipremsagar](https://github.com/galipremsagar) -- Remove pyarrow upper bound ([#19870](https://github.com/rapidsai/cudf/pull/19870)) [@vyasr](https://github.com/vyasr) -- Prevent installation of pytest-rerunfailures 16.0.0 ([#19863](https://github.com/rapidsai/cudf/pull/19863)) [@pentschev](https://github.com/pentschev) -- use 'nvidia-ml-py' package for 'pynvml' module ([#19862](https://github.com/rapidsai/cudf/pull/19862)) [@jameslamb](https://github.com/jameslamb) -- Avoid more direct construction of cuDF classic columns ([#19858](https://github.com/rapidsai/cudf/pull/19858)) [@mroeschke](https://github.com/mroeschke) -- Bump pandas supported version to `2.3.2` ([#19856](https://github.com/rapidsai/cudf/pull/19856)) [@galipremsagar](https://github.com/galipremsagar) -- Use cupy arrays instead of numba device arrays for cuDF classic intermediates ([#19855](https://github.com/rapidsai/cudf/pull/19855)) [@mroeschke](https://github.com/mroeschke) -- Move row operators to detail and deprecate legacy ([#19849](https://github.com/rapidsai/cudf/pull/19849)) [@PointKernel](https://github.com/PointKernel) -- Fix flaky DataFrame `to_string` test ([#19847](https://github.com/rapidsai/cudf/pull/19847)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Pin pytest-rerunfailures<16 ([#19846](https://github.com/rapidsai/cudf/pull/19846)) [@mroeschke](https://github.com/mroeschke) -- revert numba CUDA 13 workaround ([#19842](https://github.com/rapidsai/cudf/pull/19842)) [@jameslamb](https://github.com/jameslamb) -- Avoid CategoricalColumn constructors in cuDF classic ([#19837](https://github.com/rapidsai/cudf/pull/19837)) [@mroeschke](https://github.com/mroeschke) -- Construct cuDF classic Decimal32/64Columns from RMM buffers ([#19834](https://github.com/rapidsai/cudf/pull/19834)) [@mroeschke](https://github.com/mroeschke) -- Avoid direct construction of cuDF classic columns ([#19829](https://github.com/rapidsai/cudf/pull/19829)) [@mroeschke](https://github.com/mroeschke) -- Support input filename in ndsh q01 benchmark ([#19820](https://github.com/rapidsai/cudf/pull/19820)) [@davidwendt](https://github.com/davidwendt) -- Run cudf-polars-polars-tests on changes in test_python file group ([#19819](https://github.com/rapidsai/cudf/pull/19819)) [@mroeschke](https://github.com/mroeschke) -- Remove test_mvc.py ([#19816](https://github.com/rapidsai/cudf/pull/19816)) [@mroeschke](https://github.com/mroeschke) -- pin oldest numpy in dask-cudf tests, update dependency floors (cuda-python 12.9.2, cupy 13.6.0, numba 0.60.0) ([#19806](https://github.com/rapidsai/cudf/pull/19806)) [@jameslamb](https://github.com/jameslamb) -- Remove iterative `nan` & `nat` inefficient checks in `as_column` constructor ([#19804](https://github.com/rapidsai/cudf/pull/19804)) [@galipremsagar](https://github.com/galipremsagar) -- Simplify/consolidate from_arrow logic ([#19801](https://github.com/rapidsai/cudf/pull/19801)) [@mroeschke](https://github.com/mroeschke) -- Refactor column_empty to use only pylibcudf APIs ([#19800](https://github.com/rapidsai/cudf/pull/19800)) [@mroeschke](https://github.com/mroeschke) -- Use more cached_property where possible for Index and subclasses ([#19799](https://github.com/rapidsai/cudf/pull/19799)) [@mroeschke](https://github.com/mroeschke) -- Update rapids-dependency-file-generator ([#19796](https://github.com/rapidsai/cudf/pull/19796)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- rearrange dependencies.yaml, other small changes ([#19794](https://github.com/rapidsai/cudf/pull/19794)) [@jameslamb](https://github.com/jameslamb) -- Update exception handling in pdsh benchmarks ([#19793](https://github.com/rapidsai/cudf/pull/19793)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix how nvcomp major version is extracted ([#19791](https://github.com/rapidsai/cudf/pull/19791)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Use KvikIO's unified interface to create remote I/O endpoints ([#19788](https://github.com/rapidsai/cudf/pull/19788)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Add object-oriented APIs for left semi- and anti- join (Part I) ([#19778](https://github.com/rapidsai/cudf/pull/19778)) [@shrshi](https://github.com/shrshi) -- Add nvbench benchmark for cudf::encode API ([#19777](https://github.com/rapidsai/cudf/pull/19777)) [@davidwendt](https://github.com/davidwendt) -- Some clarifications, improvements to GroupedRollingWindows in cudf-polars ([#19776](https://github.com/rapidsai/cudf/pull/19776)) [@Matt711](https://github.com/Matt711) -- Remove validation on import ([#19775](https://github.com/rapidsai/cudf/pull/19775)) [@vyasr](https://github.com/vyasr) -- Move more test_dataframe.py tests to new cudf classic testing directory ([#19770](https://github.com/rapidsai/cudf/pull/19770)) [@mroeschke](https://github.com/mroeschke) -- Build and test with CUDA 13.0.0 ([#19768](https://github.com/rapidsai/cudf/pull/19768)) [@jameslamb](https://github.com/jameslamb) -- Skip polars CPU perf test for with_columns ([#19763](https://github.com/rapidsai/cudf/pull/19763)) [@Matt711](https://github.com/Matt711) -- Optionally capture Shuffle Stats in cudf-polars pdsh benchmarks ([#19762](https://github.com/rapidsai/cudf/pull/19762)) [@TomAugspurger](https://github.com/TomAugspurger) -- Expand compression codec coverage in ORC and Parquet benchmarks ([#19760](https://github.com/rapidsai/cudf/pull/19760)) [@vuule](https://github.com/vuule) -- Add ``ColumnSourceInfo`` convenience layer ([#19752](https://github.com/rapidsai/cudf/pull/19752)) [@rjzamora](https://github.com/rjzamora) -- Support decimal columns in cudf_polars ([#19749](https://github.com/rapidsai/cudf/pull/19749)) [@mroeschke](https://github.com/mroeschke) -- Skip third-party tests when possible ([#19747](https://github.com/rapidsai/cudf/pull/19747)) [@vyasr](https://github.com/vyasr) -- Revert "Support decimal columns in cudf_polars" ([#19746](https://github.com/rapidsai/cudf/pull/19746)) [@mroeschke](https://github.com/mroeschke) -- Vendor libnvcomp in libcudf ([#19743](https://github.com/rapidsai/cudf/pull/19743)) [@bdice](https://github.com/bdice) -- Remove outdated numba workarounds ([#19738](https://github.com/rapidsai/cudf/pull/19738)) [@bdice](https://github.com/bdice) -- Move test_buffer/column/column_accesor/cuda_apply.py to new cudf classic testing directory ([#19737](https://github.com/rapidsai/cudf/pull/19737)) [@mroeschke](https://github.com/mroeschke) -- Move more test_dataframe.py tests to new cudf classic testing directory ([#19731](https://github.com/rapidsai/cudf/pull/19731)) [@mroeschke](https://github.com/mroeschke) -- Move test_udf_masked_ops/test_dropna to new cudf classic testing directory ([#19730](https://github.com/rapidsai/cudf/pull/19730)) [@mroeschke](https://github.com/mroeschke) -- Move test_numerical/{numpy|pandas}_interop/setitem.py to new cudf classic testing directory ([#19725](https://github.com/rapidsai/cudf/pull/19725)) [@mroeschke](https://github.com/mroeschke) -- Move test_timedelta/string/sorting/list/datetime.py to new cudf classic directory structure ([#19723](https://github.com/rapidsai/cudf/pull/19723)) [@mroeschke](https://github.com/mroeschke) -- Warn on fallback in the streaming tests in cudf-polars ([#19721](https://github.com/rapidsai/cudf/pull/19721)) [@Matt711](https://github.com/Matt711) -- Optionally print shuffle stats in pdsh benchmarks ([#19719](https://github.com/rapidsai/cudf/pull/19719)) [@TomAugspurger](https://github.com/TomAugspurger) -- Move test_{io}.py files to new cudf classic test directory ([#19709](https://github.com/rapidsai/cudf/pull/19709)) [@mroeschke](https://github.com/mroeschke) -- Move to pyarrow and numpy to run_constrained ([#19706](https://github.com/rapidsai/cudf/pull/19706)) [@vyasr](https://github.com/vyasr) -- Remove unreachable code in rapidsmpf shuffle ([#19704](https://github.com/rapidsai/cudf/pull/19704)) [@TomAugspurger](https://github.com/TomAugspurger) -- Moves test_options to cudf testing directory, clean up old, stubbed testing files in directory ([#19698](https://github.com/rapidsai/cudf/pull/19698)) [@mroeschke](https://github.com/mroeschke) -- Move (most of) test_index.py to new cudf classic directory structure ([#19696](https://github.com/rapidsai/cudf/pull/19696)) [@mroeschke](https://github.com/mroeschke) -- Improve `M2`, `VARIANCE` and `STD` hash-based groupby aggregations ([#19694](https://github.com/rapidsai/cudf/pull/19694)) [@ttnghia](https://github.com/ttnghia) -- Move quantiles libcudf benchmark to nvbench ([#19692](https://github.com/rapidsai/cudf/pull/19692)) [@davidwendt](https://github.com/davidwendt) -- Handle `TIMESTAMP_DAYS` in rolling window offsets ([#19689](https://github.com/rapidsai/cudf/pull/19689)) [@Matt711](https://github.com/Matt711) -- Move test_groupby to new cudf classic directory structure ([#19688](https://github.com/rapidsai/cudf/pull/19688)) [@mroeschke](https://github.com/mroeschke) -- Move some of test_dataframe.py to new cudf classic directory structure ([#19687](https://github.com/rapidsai/cudf/pull/19687)) [@mroeschke](https://github.com/mroeschke) -- Change nvtext::character_tokenize to return a list column ([#19685](https://github.com/rapidsai/cudf/pull/19685)) [@davidwendt](https://github.com/davidwendt) -- Split up rolling.cuh into separate headers ([#19682](https://github.com/rapidsai/cudf/pull/19682)) [@davidwendt](https://github.com/davidwendt) -- Move test_factorize/drop_duplicates.py to new cudf classic test directory ([#19681](https://github.com/rapidsai/cudf/pull/19681)) [@mroeschke](https://github.com/mroeschke) -- Move test_offset/repr.py to new cudf classic testing directory ([#19677](https://github.com/rapidsai/cudf/pull/19677)) [@mroeschke](https://github.com/mroeschke) -- Move test_stats/reductions/quantile and misc to new cudf classic testing directory ([#19675](https://github.com/rapidsai/cudf/pull/19675)) [@mroeschke](https://github.com/mroeschke) -- Cache hash values to improve hash-based groupby performance with wide/complex table keys ([#19670](https://github.com/rapidsai/cudf/pull/19670)) [@ttnghia](https://github.com/ttnghia) -- Move test_interval/test_dtypes/test_rank.py to new cudf directory structure ([#19668](https://github.com/rapidsai/cudf/pull/19668)) [@mroeschke](https://github.com/mroeschke) -- Clean and move test_join_order/interpolate/onehot.py to new cudf classic test directory structure ([#19662](https://github.com/rapidsai/cudf/pull/19662)) [@mroeschke](https://github.com/mroeschke) -- Migrate mixed join to use multiset ([#19660](https://github.com/rapidsai/cudf/pull/19660)) [@PointKernel](https://github.com/PointKernel) -- Run pylibcudf tests without its optional dependencies ([#19657](https://github.com/rapidsai/cudf/pull/19657)) [@vyasr](https://github.com/vyasr) -- Use build cluster in devcontainers ([#19652](https://github.com/rapidsai/cudf/pull/19652)) [@trxcllnt](https://github.com/trxcllnt) -- Use rapids_cuda_enable_fatbin_compression ([#19650](https://github.com/rapidsai/cudf/pull/19650)) [@robertmaynard](https://github.com/robertmaynard) -- Re-enable Disabled Join Tests ([#19649](https://github.com/rapidsai/cudf/pull/19649)) [@PointKernel](https://github.com/PointKernel) -- Use public Arrow functions for TDigest in PercentileApproxInputTypesTests ([#19648](https://github.com/rapidsai/cudf/pull/19648)) [@davidwendt](https://github.com/davidwendt) -- Use cudaDeviceGetAttribute to get ComputeMode for CUDA13 ([#19645](https://github.com/rapidsai/cudf/pull/19645)) [@GaryShen2008](https://github.com/GaryShen2008) -- remove initial memset of values in parquet reader ([#19643](https://github.com/rapidsai/cudf/pull/19643)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Move ~half of test_groupby.py to new cudf classic test directory structure ([#19640](https://github.com/rapidsai/cudf/pull/19640)) [@mroeschke](https://github.com/mroeschke) -- Move test_csv/feather/json.py to new cudf classic test directory structure ([#19639](https://github.com/rapidsai/cudf/pull/19639)) [@mroeschke](https://github.com/mroeschke) -- Move test_array_function/ufunc to new cudf classic test directory structure ([#19637](https://github.com/rapidsai/cudf/pull/19637)) [@mroeschke](https://github.com/mroeschke) -- Fix anchor naming conventions in dependencies.yaml ([#19635](https://github.com/rapidsai/cudf/pull/19635)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Require `--scale` for PDS-DS benchmarks (due to nonlinear scaling) ([#19631](https://github.com/rapidsai/cudf/pull/19631)) [@Matt711](https://github.com/Matt711) -- Move test_replace.py to new cudf classic directory structure ([#19629](https://github.com/rapidsai/cudf/pull/19629)) [@mroeschke](https://github.com/mroeschke) -- Move test_concat/test_reductions.py to new cudf classic directory structure ([#19626](https://github.com/rapidsai/cudf/pull/19626)) [@mroeschke](https://github.com/mroeschke) -- Update rapids_config to handle user defined branch name ([#19623](https://github.com/rapidsai/cudf/pull/19623)) [@robertmaynard](https://github.com/robertmaynard) -- Add nvtx ranges to public APIs of the experimental parquet reader ([#19618](https://github.com/rapidsai/cudf/pull/19618)) [@mhaseeb123](https://github.com/mhaseeb123) -- Move test_resampling/query/pickling to new cudf classic directory structure ([#19615](https://github.com/rapidsai/cudf/pull/19615)) [@mroeschke](https://github.com/mroeschke) -- Move test_reshape.py to new cudf classic directory strucutre, remove reshape._merge_sorted ([#19614](https://github.com/rapidsai/cudf/pull/19614)) [@mroeschke](https://github.com/mroeschke) -- Move test_rolling/ewm.py to new cudf classic directory structure ([#19611](https://github.com/rapidsai/cudf/pull/19611)) [@mroeschke](https://github.com/mroeschke) -- Simplify cudf::scalar usage in reduce utility ([#19608](https://github.com/rapidsai/cudf/pull/19608)) [@davidwendt](https://github.com/davidwendt) -- Update to numba-cuda>=0.18.0,<0.19.0 ([#19604](https://github.com/rapidsai/cudf/pull/19604)) [@bdice](https://github.com/bdice) -- Update spark-rapdis-jni action to use PR's base.ref and fix issue of ccache version in dockerfile ([#19603](https://github.com/rapidsai/cudf/pull/19603)) [@pxLi](https://github.com/pxLi) -- Multithreaded CPU algorithm for data page mask computation ([#19602](https://github.com/rapidsai/cudf/pull/19602)) [@mhaseeb123](https://github.com/mhaseeb123) -- Move test_cuda_array_interface/cut/dataframe_copy.py to new cudf classic test directories ([#19599](https://github.com/rapidsai/cudf/pull/19599)) [@mroeschke](https://github.com/mroeschke) -- Support decimal columns in cudf_polars ([#19589](https://github.com/rapidsai/cudf/pull/19589)) [@mroeschke](https://github.com/mroeschke) -- Preserve decimal precision in `cudf::interop::column_metadata` ([#19587](https://github.com/rapidsai/cudf/pull/19587)) [@mroeschke](https://github.com/mroeschke) -- Always use strict zipping ([#19584](https://github.com/rapidsai/cudf/pull/19584)) [@vyasr](https://github.com/vyasr) -- Pin polars version to <1.33 ([#19582](https://github.com/rapidsai/cudf/pull/19582)) [@Matt711](https://github.com/Matt711) -- ci(labeler): update labeler action to [@v5 ([#19581](https://github.com/rapidsai/cudf/pull/19581)) @gforsyth](https://github.com/v5 ([#19581](https://github.com/rapidsai/cudf/pull/19581)) @gforsyth) -- Update rapids-build-backend to 0.4.0 ([#19580](https://github.com/rapidsai/cudf/pull/19580)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Move (most of) test_list.py to new cudf classic test directories ([#19574](https://github.com/rapidsai/cudf/pull/19574)) [@mroeschke](https://github.com/mroeschke) -- Move test_monotonic.py to new cudf classic test directory structure ([#19572](https://github.com/rapidsai/cudf/pull/19572)) [@mroeschke](https://github.com/mroeschke) -- Additional gtests error checks for string/timestamp convert libcudf APIs ([#19562](https://github.com/rapidsai/cudf/pull/19562)) [@davidwendt](https://github.com/davidwendt) -- Avoid cudf.pandas fallback for `pandas.array.NumpyExtensionArray` of strings ([#19558](https://github.com/rapidsai/cudf/pull/19558)) [@mroeschke](https://github.com/mroeschke) -- Move str accessor tests in test_string.py to new cudf classic test directory structure ([#19557](https://github.com/rapidsai/cudf/pull/19557)) [@mroeschke](https://github.com/mroeschke) -- Rework fill/repeat benchmark to use nvbench ([#19556](https://github.com/rapidsai/cudf/pull/19556)) [@davidwendt](https://github.com/davidwendt) -- Use no_validity() instead of null_probability(0) in benchmarks profile ([#19554](https://github.com/rapidsai/cudf/pull/19554)) [@davidwendt](https://github.com/davidwendt) -- Move (most of) test_timedelta.py and test_struct.py to new cudf classic test directory structure ([#19551](https://github.com/rapidsai/cudf/pull/19551)) [@mroeschke](https://github.com/mroeschke) -- Capture commit hashes in pdsh benchmarks ([#19548](https://github.com/rapidsai/cudf/pull/19548)) [@TomAugspurger](https://github.com/TomAugspurger) -- Simplify clang dependency spec ([#19546](https://github.com/rapidsai/cudf/pull/19546)) [@vyasr](https://github.com/vyasr) -- Move timeout in cudf.pandas pandas unit tests script to ci script ([#19542](https://github.com/rapidsai/cudf/pull/19542)) [@mroeschke](https://github.com/mroeschke) -- [FEA] Refactor AST `operator_functor`s for use in JIT-compiled CUDA ([#19541](https://github.com/rapidsai/cudf/pull/19541)) [@lamarrr](https://github.com/lamarrr) -- Construct cuDF classic columns with __array_interface__ through pylibcudf ([#19538](https://github.com/rapidsai/cudf/pull/19538)) [@mroeschke](https://github.com/mroeschke) -- Separate row mask and page mask computation and usage ([#19537](https://github.com/rapidsai/cudf/pull/19537)) [@mhaseeb123](https://github.com/mhaseeb123) -- Get rid of CG logic in the mixed semi-join kernel ([#19536](https://github.com/rapidsai/cudf/pull/19536)) [@PointKernel](https://github.com/PointKernel) -- Construct more cuDF classic Columns with pylibcudf instead of using Buffers ([#19535](https://github.com/rapidsai/cudf/pull/19535)) [@mroeschke](https://github.com/mroeschke) -- Fix clang-tools version pinning ([#19529](https://github.com/rapidsai/cudf/pull/19529)) [@wence-](https://github.com/wence-) -- Add cudf_polars unit test for `is_in([])` expr ([#19525](https://github.com/rapidsai/cudf/pull/19525)) [@mroeschke](https://github.com/mroeschke) -- Expose `nvtext::letter_type` to python ([#19520](https://github.com/rapidsai/cudf/pull/19520)) [@Matt711](https://github.com/Matt711) -- Remove c++ stringview interop example ([#19516](https://github.com/rapidsai/cudf/pull/19516)) [@davidwendt](https://github.com/davidwendt) -- Remove cudf/_fuzz_testing directory ([#19510](https://github.com/rapidsai/cudf/pull/19510)) [@mroeschke](https://github.com/mroeschke) -- Add missing import of pyarrow.parquet when reading specified row_groups. ([#19509](https://github.com/rapidsai/cudf/pull/19509)) [@bdice](https://github.com/bdice) -- Don't run serial cudf_pandas tests when testing multiple pandas versions ([#19507](https://github.com/rapidsai/cudf/pull/19507)) [@mroeschke](https://github.com/mroeschke) -- Clean testing/_utils.py ([#19506](https://github.com/rapidsai/cudf/pull/19506)) [@mroeschke](https://github.com/mroeschke) -- Move some test_datetime.py tests to new cudf classic test directory structure ([#19505](https://github.com/rapidsai/cudf/pull/19505)) [@mroeschke](https://github.com/mroeschke) -- Move test_joining to new cudf classic test directory structure ([#19501](https://github.com/rapidsai/cudf/pull/19501)) [@mroeschke](https://github.com/mroeschke) -- Upgrade `gcc-toolset` for Java/JNI build to version 14 ([#19500](https://github.com/rapidsai/cudf/pull/19500)) [@ttnghia](https://github.com/ttnghia) -- Remove deprecated subword-tokenizer APIs ([#19498](https://github.com/rapidsai/cudf/pull/19498)) [@davidwendt](https://github.com/davidwendt) -- Move some test_multiindex.py to new cudf classic test directory structure ([#19496](https://github.com/rapidsai/cudf/pull/19496)) [@mroeschke](https://github.com/mroeschke) -- Add nvtx ranges and minor fix for `lists` types in the next-gen parquet reader ([#19493](https://github.com/rapidsai/cudf/pull/19493)) [@mhaseeb123](https://github.com/mhaseeb123) -- Move test_search/test_scan/test_seriesmap.py to new cudf classic test directory structure ([#19492](https://github.com/rapidsai/cudf/pull/19492)) [@mroeschke](https://github.com/mroeschke) -- Improve support for sliced input on from_arrow_host APIs ([#19491](https://github.com/rapidsai/cudf/pull/19491)) [@davidwendt](https://github.com/davidwendt) -- Move test_avro/test_api_types.py and some DataFrame tests to new cudf classic test directory structure ([#19490](https://github.com/rapidsai/cudf/pull/19490)) [@mroeschke](https://github.com/mroeschke) -- Move test_series.py to new cudf classic test directory structure ([#19485](https://github.com/rapidsai/cudf/pull/19485)) [@mroeschke](https://github.com/mroeschke) -- Move test_testing.py to new cudf classic test directory structure ([#19481](https://github.com/rapidsai/cudf/pull/19481)) [@mroeschke](https://github.com/mroeschke) -- Allow latest OS in devcontainers ([#19480](https://github.com/rapidsai/cudf/pull/19480)) [@bdice](https://github.com/bdice) -- Move test_unaops/test_unique/test_transform.py to new cudf classic test directory structure ([#19477](https://github.com/rapidsai/cudf/pull/19477)) [@mroeschke](https://github.com/mroeschke) -- Branch 25.10 merge branch 25.08 ([#19475](https://github.com/rapidsai/cudf/pull/19475)) [@davidwendt](https://github.com/davidwendt) -- Use more pytest fixtures and clean data files cuDF classic tests subdirectories ([#19474](https://github.com/rapidsai/cudf/pull/19474)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in test_binops/column/column_accessor/contains.py and more ([#19473](https://github.com/rapidsai/cudf/pull/19473)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in test_csv/cuda_*/cut.py and more ([#19463](https://github.com/rapidsai/cudf/pull/19463)) [@mroeschke](https://github.com/mroeschke) -- Improve readability when printing pylibcudf enums ([#19451](https://github.com/rapidsai/cudf/pull/19451)) [@Matt711](https://github.com/Matt711) -- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19450](https://github.com/rapidsai/cudf/pull/19450)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in test_dropna/factorize.py and more ([#19449](https://github.com/rapidsai/cudf/pull/19449)) [@mroeschke](https://github.com/mroeschke) -- Update build infra to support new branching strategy ([#19445](https://github.com/rapidsai/cudf/pull/19445)) [@robertmaynard](https://github.com/robertmaynard) -- Updated libcudf-example conda package to preserve directories structure ([#19440](https://github.com/rapidsai/cudf/pull/19440)) [@Avinash-Raj](https://github.com/Avinash-Raj) -- Use more pytest fixtures and avoid GPU parameterization in test_groupby/index.py ([#19438](https://github.com/rapidsai/cudf/pull/19438)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in test_indexing/joining/monotonic/multiindex.py ([#19437](https://github.com/rapidsai/cudf/pull/19437)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19436](https://github.com/rapidsai/cudf/pull/19436)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in test_query/rank/reduction/repr.py ([#19434](https://github.com/rapidsai/cudf/pull/19434)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in test_replace/reshape/rolling.py ([#19426](https://github.com/rapidsai/cudf/pull/19426)) [@mroeschke](https://github.com/mroeschke) -- Update s3 Bucket fixture creation in test_s3 ([#19424](https://github.com/rapidsai/cudf/pull/19424)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19419](https://github.com/rapidsai/cudf/pull/19419)) [@mroeschke](https://github.com/mroeschke) -- Fix various pandas test failures in `cudf.pandas` ([#19372](https://github.com/rapidsai/cudf/pull/19372)) [@galipremsagar](https://github.com/galipremsagar) -- Pin Narwhals to 1.47 ([#19358](https://github.com/rapidsai/cudf/pull/19358)) [@Matt711](https://github.com/Matt711) -- Run cudf-polars tests with all supported polars versions ([#19353](https://github.com/rapidsai/cudf/pull/19353)) [@Matt711](https://github.com/Matt711) -- Update `pandas-tests-diff` to only display GPU/CPU usage metrics ([#19210](https://github.com/rapidsai/cudf/pull/19210)) [@galipremsagar](https://github.com/galipremsagar) -- Use GCC 14 in conda builds. ([#19192](https://github.com/rapidsai/cudf/pull/19192)) [@vyasr](https://github.com/vyasr) -- Use KvikIO's implementation of file-backed memory mapping ([#19164](https://github.com/rapidsai/cudf/pull/19164)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Replace `rmm::device_scalar` with `cudf::detail::device_scalar` due to unnecessary synchronization (Part 3 of miss-sync) ([#19119](https://github.com/rapidsai/cudf/pull/19119)) [@JigaoLuo](https://github.com/JigaoLuo) -- Implement distributed sorted for ``cudf_polars`` ([#18912](https://github.com/rapidsai/cudf/pull/18912)) [@seberg](https://github.com/seberg) +- Fix CI failures for `pandas-2.3.3` ([#20146](https://github.com/NVIDIA/cudf/pull/20146)) [@galipremsagar](https://github.com/galipremsagar) +- Skip passing failures for latest `numexpr` version ([#20092](https://github.com/NVIDIA/cudf/pull/20092)) [@galipremsagar](https://github.com/galipremsagar) +- Empty commit to trigger a build ([#20084](https://github.com/NVIDIA/cudf/pull/20084)) [@msarahan](https://github.com/msarahan) +- Update the reason to skip for parquet bloom filter test ([#20043](https://github.com/NVIDIA/cudf/pull/20043)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove test_scan_hf_url_raises ([#20035](https://github.com/NVIDIA/cudf/pull/20035)) [@mroeschke](https://github.com/mroeschke) +- xfail(strict=False) test_scan_hf_url_raises due to rate limiting ([#20027](https://github.com/NVIDIA/cudf/pull/20027)) [@mroeschke](https://github.com/mroeschke) +- Deprecate left semi- and anti- join functional APIs ([#20014](https://github.com/NVIDIA/cudf/pull/20014)) [@shrshi](https://github.com/shrshi) +- Use to_arrow methods throughout pylibcudf and cudf ([#20013](https://github.com/NVIDIA/cudf/pull/20013)) [@Matt711](https://github.com/Matt711) +- Fix chunked reads of list of bools. ([#20000](https://github.com/NVIDIA/cudf/pull/20000)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Raise more exceptions for invalid or unsupported cuDF arguments ([#19990](https://github.com/NVIDIA/cudf/pull/19990)) [@mroeschke](https://github.com/mroeschke) +- Configure repo for automatic release notes generation ([#19984](https://github.com/NVIDIA/cudf/pull/19984)) [@AyodeAwe](https://github.com/AyodeAwe) +- Pin duckdb<1.4 in test_python_narwhals ([#19982](https://github.com/NVIDIA/cudf/pull/19982)) [@mroeschke](https://github.com/mroeschke) +- Default to False if `CUDA_ENABLE_NRT` isn't set in config ([#19981](https://github.com/NVIDIA/cudf/pull/19981)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Remove UCX-Py ([#19979](https://github.com/NVIDIA/cudf/pull/19979)) [@pentschev](https://github.com/pentschev) +- Add support for `attrs` ([#19978](https://github.com/NVIDIA/cudf/pull/19978)) [@galipremsagar](https://github.com/galipremsagar) +- Run pytest-benchmarks in CI with --benchmark-disable ([#19969](https://github.com/NVIDIA/cudf/pull/19969)) [@mroeschke](https://github.com/mroeschke) +- Change target type so we can test on workflows ([#19963](https://github.com/NVIDIA/cudf/pull/19963)) [@vyasr](https://github.com/vyasr) +- Update to actions/labeler v5 ([#19962](https://github.com/NVIDIA/cudf/pull/19962)) [@vyasr](https://github.com/vyasr) +- Revert "ci(labeler): update labeler action to [@v5" ([#19961](https://github.com/NVIDIA/cudf/pull/19961)) @vyasr](https://github.com/v5" ([#19961](https://github.com/NVIDIA/cudf/pull/19961)) @vyasr) +- Add `ArrowDtype` proxy class ([#19960](https://github.com/NVIDIA/cudf/pull/19960)) [@galipremsagar](https://github.com/galipremsagar) +- Add missing type stub ([#19958](https://github.com/NVIDIA/cudf/pull/19958)) [@vyasr](https://github.com/vyasr) +- Add missing `Styler` attributes ([#19956](https://github.com/NVIDIA/cudf/pull/19956)) [@galipremsagar](https://github.com/galipremsagar) +- Allow newer CMake in Java tests ([#19949](https://github.com/NVIDIA/cudf/pull/19949)) [@bdice](https://github.com/bdice) +- Make stream a required parameter for from_libcudf methods ([#19945](https://github.com/NVIDIA/cudf/pull/19945)) [@vyasr](https://github.com/vyasr) +- Return False instead of NA for comparison ops against NA in cudf.pandas ([#19942](https://github.com/NVIDIA/cudf/pull/19942)) [@mroeschke](https://github.com/mroeschke) +- Don't fall back in Series.describe in cudf.pandas for numeric types ([#19941](https://github.com/NVIDIA/cudf/pull/19941)) [@mroeschke](https://github.com/mroeschke) +- Move groupby benchmarks to nvbench ([#19930](https://github.com/NVIDIA/cudf/pull/19930)) [@davidwendt](https://github.com/davidwendt) +- Perform more input validation in cuDF classic APIs ([#19929](https://github.com/NVIDIA/cudf/pull/19929)) [@mroeschke](https://github.com/mroeschke) +- update nvidia-ml-py (>=12), use cuda-toolkit wheels ([#19927](https://github.com/NVIDIA/cudf/pull/19927)) [@jameslamb](https://github.com/jameslamb) +- Fill missing values in `Series/Index.values` for numeric types with np.nan by default ([#19923](https://github.com/NVIDIA/cudf/pull/19923)) [@mroeschke](https://github.com/mroeschke) +- Add `rmm-release-threshold` to pdsh benchmarks CLI ([#19918](https://github.com/NVIDIA/cudf/pull/19918)) [@TomAugspurger](https://github.com/TomAugspurger) +- Also use the CUDA 12 container for nightlies ([#19917](https://github.com/NVIDIA/cudf/pull/19917)) [@vyasr](https://github.com/vyasr) +- Move test_binops.py to new cuDF classic directory structure ([#19914](https://github.com/NVIDIA/cudf/pull/19914)) [@mroeschke](https://github.com/mroeschke) +- Eagerly load nvCOMP library in `cudf::initialize()` ([#19906](https://github.com/NVIDIA/cudf/pull/19906)) [@vuule](https://github.com/vuule) +- Pin to CUDA 12 image for integration tests ([#19903](https://github.com/NVIDIA/cudf/pull/19903)) [@vyasr](https://github.com/vyasr) +- Use branch-25.10 again ([#19902](https://github.com/NVIDIA/cudf/pull/19902)) [@jameslamb](https://github.com/jameslamb) +- Disable test on non-default stream ([#19901](https://github.com/NVIDIA/cudf/pull/19901)) [@vyasr](https://github.com/vyasr) +- Use cupy array instead of numba device array as inputs to jit routines ([#19897](https://github.com/NVIDIA/cudf/pull/19897)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated `DataFrame.apply_rows`, deprecate `DataFrame.apply_chunks` and `Groupby.apply_grouped` ([#19896](https://github.com/NVIDIA/cudf/pull/19896)) [@mroeschke](https://github.com/mroeschke) +- Move test_dataframe.py to new cuDF classic directory structure ([#19890](https://github.com/NVIDIA/cudf/pull/19890)) [@mroeschke](https://github.com/mroeschke) +- Make sure conftest fixture data is valid on exit ([#19889](https://github.com/NVIDIA/cudf/pull/19889)) [@vyasr](https://github.com/vyasr) +- Move test_index/multiindex/indexing.py to new cuDF classic directory structure ([#19887](https://github.com/NVIDIA/cudf/pull/19887)) [@mroeschke](https://github.com/mroeschke) +- [FEA] Build CUDF with CCCL 3.1.0 ([#19886](https://github.com/NVIDIA/cudf/pull/19886)) [@lamarrr](https://github.com/lamarrr) +- Coalesce IO of chunks with different compression when reading Parquet files ([#19884](https://github.com/NVIDIA/cudf/pull/19884)) [@vuule](https://github.com/vuule) +- Update boost version to 1.79 for JNI dockerfile ([#19883](https://github.com/NVIDIA/cudf/pull/19883)) [@pxLi](https://github.com/pxLi) +- Move test_categorical/dask/serialize.py to new cuDF classic test directory structure ([#19877](https://github.com/NVIDIA/cudf/pull/19877)) [@mroeschke](https://github.com/mroeschke) +- Move prefetching out of experimental and simplify the API ([#19875](https://github.com/NVIDIA/cudf/pull/19875)) [@vyasr](https://github.com/vyasr) +- Remove `diff.sh` and merge diff generation into `run.sh` ([#19871](https://github.com/NVIDIA/cudf/pull/19871)) [@galipremsagar](https://github.com/galipremsagar) +- Remove pyarrow upper bound ([#19870](https://github.com/NVIDIA/cudf/pull/19870)) [@vyasr](https://github.com/vyasr) +- Prevent installation of pytest-rerunfailures 16.0.0 ([#19863](https://github.com/NVIDIA/cudf/pull/19863)) [@pentschev](https://github.com/pentschev) +- use 'nvidia-ml-py' package for 'pynvml' module ([#19862](https://github.com/NVIDIA/cudf/pull/19862)) [@jameslamb](https://github.com/jameslamb) +- Avoid more direct construction of cuDF classic columns ([#19858](https://github.com/NVIDIA/cudf/pull/19858)) [@mroeschke](https://github.com/mroeschke) +- Bump pandas supported version to `2.3.2` ([#19856](https://github.com/NVIDIA/cudf/pull/19856)) [@galipremsagar](https://github.com/galipremsagar) +- Use cupy arrays instead of numba device arrays for cuDF classic intermediates ([#19855](https://github.com/NVIDIA/cudf/pull/19855)) [@mroeschke](https://github.com/mroeschke) +- Move row operators to detail and deprecate legacy ([#19849](https://github.com/NVIDIA/cudf/pull/19849)) [@PointKernel](https://github.com/PointKernel) +- Fix flaky DataFrame `to_string` test ([#19847](https://github.com/NVIDIA/cudf/pull/19847)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Pin pytest-rerunfailures<16 ([#19846](https://github.com/NVIDIA/cudf/pull/19846)) [@mroeschke](https://github.com/mroeschke) +- revert numba CUDA 13 workaround ([#19842](https://github.com/NVIDIA/cudf/pull/19842)) [@jameslamb](https://github.com/jameslamb) +- Avoid CategoricalColumn constructors in cuDF classic ([#19837](https://github.com/NVIDIA/cudf/pull/19837)) [@mroeschke](https://github.com/mroeschke) +- Construct cuDF classic Decimal32/64Columns from RMM buffers ([#19834](https://github.com/NVIDIA/cudf/pull/19834)) [@mroeschke](https://github.com/mroeschke) +- Avoid direct construction of cuDF classic columns ([#19829](https://github.com/NVIDIA/cudf/pull/19829)) [@mroeschke](https://github.com/mroeschke) +- Support input filename in ndsh q01 benchmark ([#19820](https://github.com/NVIDIA/cudf/pull/19820)) [@davidwendt](https://github.com/davidwendt) +- Run cudf-polars-polars-tests on changes in test_python file group ([#19819](https://github.com/NVIDIA/cudf/pull/19819)) [@mroeschke](https://github.com/mroeschke) +- Remove test_mvc.py ([#19816](https://github.com/NVIDIA/cudf/pull/19816)) [@mroeschke](https://github.com/mroeschke) +- pin oldest numpy in dask-cudf tests, update dependency floors (cuda-python 12.9.2, cupy 13.6.0, numba 0.60.0) ([#19806](https://github.com/NVIDIA/cudf/pull/19806)) [@jameslamb](https://github.com/jameslamb) +- Remove iterative `nan` & `nat` inefficient checks in `as_column` constructor ([#19804](https://github.com/NVIDIA/cudf/pull/19804)) [@galipremsagar](https://github.com/galipremsagar) +- Simplify/consolidate from_arrow logic ([#19801](https://github.com/NVIDIA/cudf/pull/19801)) [@mroeschke](https://github.com/mroeschke) +- Refactor column_empty to use only pylibcudf APIs ([#19800](https://github.com/NVIDIA/cudf/pull/19800)) [@mroeschke](https://github.com/mroeschke) +- Use more cached_property where possible for Index and subclasses ([#19799](https://github.com/NVIDIA/cudf/pull/19799)) [@mroeschke](https://github.com/mroeschke) +- Update rapids-dependency-file-generator ([#19796](https://github.com/NVIDIA/cudf/pull/19796)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- rearrange dependencies.yaml, other small changes ([#19794](https://github.com/NVIDIA/cudf/pull/19794)) [@jameslamb](https://github.com/jameslamb) +- Update exception handling in pdsh benchmarks ([#19793](https://github.com/NVIDIA/cudf/pull/19793)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix how nvcomp major version is extracted ([#19791](https://github.com/NVIDIA/cudf/pull/19791)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Use KvikIO's unified interface to create remote I/O endpoints ([#19788](https://github.com/NVIDIA/cudf/pull/19788)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Add object-oriented APIs for left semi- and anti- join (Part I) ([#19778](https://github.com/NVIDIA/cudf/pull/19778)) [@shrshi](https://github.com/shrshi) +- Add nvbench benchmark for cudf::encode API ([#19777](https://github.com/NVIDIA/cudf/pull/19777)) [@davidwendt](https://github.com/davidwendt) +- Some clarifications, improvements to GroupedRollingWindows in cudf-polars ([#19776](https://github.com/NVIDIA/cudf/pull/19776)) [@Matt711](https://github.com/Matt711) +- Remove validation on import ([#19775](https://github.com/NVIDIA/cudf/pull/19775)) [@vyasr](https://github.com/vyasr) +- Move more test_dataframe.py tests to new cudf classic testing directory ([#19770](https://github.com/NVIDIA/cudf/pull/19770)) [@mroeschke](https://github.com/mroeschke) +- Build and test with CUDA 13.0.0 ([#19768](https://github.com/NVIDIA/cudf/pull/19768)) [@jameslamb](https://github.com/jameslamb) +- Skip polars CPU perf test for with_columns ([#19763](https://github.com/NVIDIA/cudf/pull/19763)) [@Matt711](https://github.com/Matt711) +- Optionally capture Shuffle Stats in cudf-polars pdsh benchmarks ([#19762](https://github.com/NVIDIA/cudf/pull/19762)) [@TomAugspurger](https://github.com/TomAugspurger) +- Expand compression codec coverage in ORC and Parquet benchmarks ([#19760](https://github.com/NVIDIA/cudf/pull/19760)) [@vuule](https://github.com/vuule) +- Add ``ColumnSourceInfo`` convenience layer ([#19752](https://github.com/NVIDIA/cudf/pull/19752)) [@rjzamora](https://github.com/rjzamora) +- Support decimal columns in cudf_polars ([#19749](https://github.com/NVIDIA/cudf/pull/19749)) [@mroeschke](https://github.com/mroeschke) +- Skip third-party tests when possible ([#19747](https://github.com/NVIDIA/cudf/pull/19747)) [@vyasr](https://github.com/vyasr) +- Revert "Support decimal columns in cudf_polars" ([#19746](https://github.com/NVIDIA/cudf/pull/19746)) [@mroeschke](https://github.com/mroeschke) +- Vendor libnvcomp in libcudf ([#19743](https://github.com/NVIDIA/cudf/pull/19743)) [@bdice](https://github.com/bdice) +- Remove outdated numba workarounds ([#19738](https://github.com/NVIDIA/cudf/pull/19738)) [@bdice](https://github.com/bdice) +- Move test_buffer/column/column_accesor/cuda_apply.py to new cudf classic testing directory ([#19737](https://github.com/NVIDIA/cudf/pull/19737)) [@mroeschke](https://github.com/mroeschke) +- Move more test_dataframe.py tests to new cudf classic testing directory ([#19731](https://github.com/NVIDIA/cudf/pull/19731)) [@mroeschke](https://github.com/mroeschke) +- Move test_udf_masked_ops/test_dropna to new cudf classic testing directory ([#19730](https://github.com/NVIDIA/cudf/pull/19730)) [@mroeschke](https://github.com/mroeschke) +- Move test_numerical/{numpy|pandas}_interop/setitem.py to new cudf classic testing directory ([#19725](https://github.com/NVIDIA/cudf/pull/19725)) [@mroeschke](https://github.com/mroeschke) +- Move test_timedelta/string/sorting/list/datetime.py to new cudf classic directory structure ([#19723](https://github.com/NVIDIA/cudf/pull/19723)) [@mroeschke](https://github.com/mroeschke) +- Warn on fallback in the streaming tests in cudf-polars ([#19721](https://github.com/NVIDIA/cudf/pull/19721)) [@Matt711](https://github.com/Matt711) +- Optionally print shuffle stats in pdsh benchmarks ([#19719](https://github.com/NVIDIA/cudf/pull/19719)) [@TomAugspurger](https://github.com/TomAugspurger) +- Move test_{io}.py files to new cudf classic test directory ([#19709](https://github.com/NVIDIA/cudf/pull/19709)) [@mroeschke](https://github.com/mroeschke) +- Move to pyarrow and numpy to run_constrained ([#19706](https://github.com/NVIDIA/cudf/pull/19706)) [@vyasr](https://github.com/vyasr) +- Remove unreachable code in rapidsmpf shuffle ([#19704](https://github.com/NVIDIA/cudf/pull/19704)) [@TomAugspurger](https://github.com/TomAugspurger) +- Moves test_options to cudf testing directory, clean up old, stubbed testing files in directory ([#19698](https://github.com/NVIDIA/cudf/pull/19698)) [@mroeschke](https://github.com/mroeschke) +- Move (most of) test_index.py to new cudf classic directory structure ([#19696](https://github.com/NVIDIA/cudf/pull/19696)) [@mroeschke](https://github.com/mroeschke) +- Improve `M2`, `VARIANCE` and `STD` hash-based groupby aggregations ([#19694](https://github.com/NVIDIA/cudf/pull/19694)) [@ttnghia](https://github.com/ttnghia) +- Move quantiles libcudf benchmark to nvbench ([#19692](https://github.com/NVIDIA/cudf/pull/19692)) [@davidwendt](https://github.com/davidwendt) +- Handle `TIMESTAMP_DAYS` in rolling window offsets ([#19689](https://github.com/NVIDIA/cudf/pull/19689)) [@Matt711](https://github.com/Matt711) +- Move test_groupby to new cudf classic directory structure ([#19688](https://github.com/NVIDIA/cudf/pull/19688)) [@mroeschke](https://github.com/mroeschke) +- Move some of test_dataframe.py to new cudf classic directory structure ([#19687](https://github.com/NVIDIA/cudf/pull/19687)) [@mroeschke](https://github.com/mroeschke) +- Change nvtext::character_tokenize to return a list column ([#19685](https://github.com/NVIDIA/cudf/pull/19685)) [@davidwendt](https://github.com/davidwendt) +- Split up rolling.cuh into separate headers ([#19682](https://github.com/NVIDIA/cudf/pull/19682)) [@davidwendt](https://github.com/davidwendt) +- Move test_factorize/drop_duplicates.py to new cudf classic test directory ([#19681](https://github.com/NVIDIA/cudf/pull/19681)) [@mroeschke](https://github.com/mroeschke) +- Move test_offset/repr.py to new cudf classic testing directory ([#19677](https://github.com/NVIDIA/cudf/pull/19677)) [@mroeschke](https://github.com/mroeschke) +- Move test_stats/reductions/quantile and misc to new cudf classic testing directory ([#19675](https://github.com/NVIDIA/cudf/pull/19675)) [@mroeschke](https://github.com/mroeschke) +- Cache hash values to improve hash-based groupby performance with wide/complex table keys ([#19670](https://github.com/NVIDIA/cudf/pull/19670)) [@ttnghia](https://github.com/ttnghia) +- Move test_interval/test_dtypes/test_rank.py to new cudf directory structure ([#19668](https://github.com/NVIDIA/cudf/pull/19668)) [@mroeschke](https://github.com/mroeschke) +- Clean and move test_join_order/interpolate/onehot.py to new cudf classic test directory structure ([#19662](https://github.com/NVIDIA/cudf/pull/19662)) [@mroeschke](https://github.com/mroeschke) +- Migrate mixed join to use multiset ([#19660](https://github.com/NVIDIA/cudf/pull/19660)) [@PointKernel](https://github.com/PointKernel) +- Run pylibcudf tests without its optional dependencies ([#19657](https://github.com/NVIDIA/cudf/pull/19657)) [@vyasr](https://github.com/vyasr) +- Use build cluster in devcontainers ([#19652](https://github.com/NVIDIA/cudf/pull/19652)) [@trxcllnt](https://github.com/trxcllnt) +- Use rapids_cuda_enable_fatbin_compression ([#19650](https://github.com/NVIDIA/cudf/pull/19650)) [@robertmaynard](https://github.com/robertmaynard) +- Re-enable Disabled Join Tests ([#19649](https://github.com/NVIDIA/cudf/pull/19649)) [@PointKernel](https://github.com/PointKernel) +- Use public Arrow functions for TDigest in PercentileApproxInputTypesTests ([#19648](https://github.com/NVIDIA/cudf/pull/19648)) [@davidwendt](https://github.com/davidwendt) +- Use cudaDeviceGetAttribute to get ComputeMode for CUDA13 ([#19645](https://github.com/NVIDIA/cudf/pull/19645)) [@GaryShen2008](https://github.com/GaryShen2008) +- remove initial memset of values in parquet reader ([#19643](https://github.com/NVIDIA/cudf/pull/19643)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Move ~half of test_groupby.py to new cudf classic test directory structure ([#19640](https://github.com/NVIDIA/cudf/pull/19640)) [@mroeschke](https://github.com/mroeschke) +- Move test_csv/feather/json.py to new cudf classic test directory structure ([#19639](https://github.com/NVIDIA/cudf/pull/19639)) [@mroeschke](https://github.com/mroeschke) +- Move test_array_function/ufunc to new cudf classic test directory structure ([#19637](https://github.com/NVIDIA/cudf/pull/19637)) [@mroeschke](https://github.com/mroeschke) +- Fix anchor naming conventions in dependencies.yaml ([#19635](https://github.com/NVIDIA/cudf/pull/19635)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Require `--scale` for PDS-DS benchmarks (due to nonlinear scaling) ([#19631](https://github.com/NVIDIA/cudf/pull/19631)) [@Matt711](https://github.com/Matt711) +- Move test_replace.py to new cudf classic directory structure ([#19629](https://github.com/NVIDIA/cudf/pull/19629)) [@mroeschke](https://github.com/mroeschke) +- Move test_concat/test_reductions.py to new cudf classic directory structure ([#19626](https://github.com/NVIDIA/cudf/pull/19626)) [@mroeschke](https://github.com/mroeschke) +- Update rapids_config to handle user defined branch name ([#19623](https://github.com/NVIDIA/cudf/pull/19623)) [@robertmaynard](https://github.com/robertmaynard) +- Add nvtx ranges to public APIs of the experimental parquet reader ([#19618](https://github.com/NVIDIA/cudf/pull/19618)) [@mhaseeb123](https://github.com/mhaseeb123) +- Move test_resampling/query/pickling to new cudf classic directory structure ([#19615](https://github.com/NVIDIA/cudf/pull/19615)) [@mroeschke](https://github.com/mroeschke) +- Move test_reshape.py to new cudf classic directory strucutre, remove reshape._merge_sorted ([#19614](https://github.com/NVIDIA/cudf/pull/19614)) [@mroeschke](https://github.com/mroeschke) +- Move test_rolling/ewm.py to new cudf classic directory structure ([#19611](https://github.com/NVIDIA/cudf/pull/19611)) [@mroeschke](https://github.com/mroeschke) +- Simplify cudf::scalar usage in reduce utility ([#19608](https://github.com/NVIDIA/cudf/pull/19608)) [@davidwendt](https://github.com/davidwendt) +- Update to numba-cuda>=0.18.0,<0.19.0 ([#19604](https://github.com/NVIDIA/cudf/pull/19604)) [@bdice](https://github.com/bdice) +- Update spark-rapdis-jni action to use PR's base.ref and fix issue of ccache version in dockerfile ([#19603](https://github.com/NVIDIA/cudf/pull/19603)) [@pxLi](https://github.com/pxLi) +- Multithreaded CPU algorithm for data page mask computation ([#19602](https://github.com/NVIDIA/cudf/pull/19602)) [@mhaseeb123](https://github.com/mhaseeb123) +- Move test_cuda_array_interface/cut/dataframe_copy.py to new cudf classic test directories ([#19599](https://github.com/NVIDIA/cudf/pull/19599)) [@mroeschke](https://github.com/mroeschke) +- Support decimal columns in cudf_polars ([#19589](https://github.com/NVIDIA/cudf/pull/19589)) [@mroeschke](https://github.com/mroeschke) +- Preserve decimal precision in `cudf::interop::column_metadata` ([#19587](https://github.com/NVIDIA/cudf/pull/19587)) [@mroeschke](https://github.com/mroeschke) +- Always use strict zipping ([#19584](https://github.com/NVIDIA/cudf/pull/19584)) [@vyasr](https://github.com/vyasr) +- Pin polars version to <1.33 ([#19582](https://github.com/NVIDIA/cudf/pull/19582)) [@Matt711](https://github.com/Matt711) +- ci(labeler): update labeler action to [@v5 ([#19581](https://github.com/NVIDIA/cudf/pull/19581)) @gforsyth](https://github.com/v5 ([#19581](https://github.com/NVIDIA/cudf/pull/19581)) @gforsyth) +- Update rapids-build-backend to 0.4.0 ([#19580](https://github.com/NVIDIA/cudf/pull/19580)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Move (most of) test_list.py to new cudf classic test directories ([#19574](https://github.com/NVIDIA/cudf/pull/19574)) [@mroeschke](https://github.com/mroeschke) +- Move test_monotonic.py to new cudf classic test directory structure ([#19572](https://github.com/NVIDIA/cudf/pull/19572)) [@mroeschke](https://github.com/mroeschke) +- Additional gtests error checks for string/timestamp convert libcudf APIs ([#19562](https://github.com/NVIDIA/cudf/pull/19562)) [@davidwendt](https://github.com/davidwendt) +- Avoid cudf.pandas fallback for `pandas.array.NumpyExtensionArray` of strings ([#19558](https://github.com/NVIDIA/cudf/pull/19558)) [@mroeschke](https://github.com/mroeschke) +- Move str accessor tests in test_string.py to new cudf classic test directory structure ([#19557](https://github.com/NVIDIA/cudf/pull/19557)) [@mroeschke](https://github.com/mroeschke) +- Rework fill/repeat benchmark to use nvbench ([#19556](https://github.com/NVIDIA/cudf/pull/19556)) [@davidwendt](https://github.com/davidwendt) +- Use no_validity() instead of null_probability(0) in benchmarks profile ([#19554](https://github.com/NVIDIA/cudf/pull/19554)) [@davidwendt](https://github.com/davidwendt) +- Move (most of) test_timedelta.py and test_struct.py to new cudf classic test directory structure ([#19551](https://github.com/NVIDIA/cudf/pull/19551)) [@mroeschke](https://github.com/mroeschke) +- Capture commit hashes in pdsh benchmarks ([#19548](https://github.com/NVIDIA/cudf/pull/19548)) [@TomAugspurger](https://github.com/TomAugspurger) +- Simplify clang dependency spec ([#19546](https://github.com/NVIDIA/cudf/pull/19546)) [@vyasr](https://github.com/vyasr) +- Move timeout in cudf.pandas pandas unit tests script to ci script ([#19542](https://github.com/NVIDIA/cudf/pull/19542)) [@mroeschke](https://github.com/mroeschke) +- [FEA] Refactor AST `operator_functor`s for use in JIT-compiled CUDA ([#19541](https://github.com/NVIDIA/cudf/pull/19541)) [@lamarrr](https://github.com/lamarrr) +- Construct cuDF classic columns with __array_interface__ through pylibcudf ([#19538](https://github.com/NVIDIA/cudf/pull/19538)) [@mroeschke](https://github.com/mroeschke) +- Separate row mask and page mask computation and usage ([#19537](https://github.com/NVIDIA/cudf/pull/19537)) [@mhaseeb123](https://github.com/mhaseeb123) +- Get rid of CG logic in the mixed semi-join kernel ([#19536](https://github.com/NVIDIA/cudf/pull/19536)) [@PointKernel](https://github.com/PointKernel) +- Construct more cuDF classic Columns with pylibcudf instead of using Buffers ([#19535](https://github.com/NVIDIA/cudf/pull/19535)) [@mroeschke](https://github.com/mroeschke) +- Fix clang-tools version pinning ([#19529](https://github.com/NVIDIA/cudf/pull/19529)) [@wence-](https://github.com/wence-) +- Add cudf_polars unit test for `is_in([])` expr ([#19525](https://github.com/NVIDIA/cudf/pull/19525)) [@mroeschke](https://github.com/mroeschke) +- Expose `nvtext::letter_type` to python ([#19520](https://github.com/NVIDIA/cudf/pull/19520)) [@Matt711](https://github.com/Matt711) +- Remove c++ stringview interop example ([#19516](https://github.com/NVIDIA/cudf/pull/19516)) [@davidwendt](https://github.com/davidwendt) +- Remove cudf/_fuzz_testing directory ([#19510](https://github.com/NVIDIA/cudf/pull/19510)) [@mroeschke](https://github.com/mroeschke) +- Add missing import of pyarrow.parquet when reading specified row_groups. ([#19509](https://github.com/NVIDIA/cudf/pull/19509)) [@bdice](https://github.com/bdice) +- Don't run serial cudf_pandas tests when testing multiple pandas versions ([#19507](https://github.com/NVIDIA/cudf/pull/19507)) [@mroeschke](https://github.com/mroeschke) +- Clean testing/_utils.py ([#19506](https://github.com/NVIDIA/cudf/pull/19506)) [@mroeschke](https://github.com/mroeschke) +- Move some test_datetime.py tests to new cudf classic test directory structure ([#19505](https://github.com/NVIDIA/cudf/pull/19505)) [@mroeschke](https://github.com/mroeschke) +- Move test_joining to new cudf classic test directory structure ([#19501](https://github.com/NVIDIA/cudf/pull/19501)) [@mroeschke](https://github.com/mroeschke) +- Upgrade `gcc-toolset` for Java/JNI build to version 14 ([#19500](https://github.com/NVIDIA/cudf/pull/19500)) [@ttnghia](https://github.com/ttnghia) +- Remove deprecated subword-tokenizer APIs ([#19498](https://github.com/NVIDIA/cudf/pull/19498)) [@davidwendt](https://github.com/davidwendt) +- Move some test_multiindex.py to new cudf classic test directory structure ([#19496](https://github.com/NVIDIA/cudf/pull/19496)) [@mroeschke](https://github.com/mroeschke) +- Add nvtx ranges and minor fix for `lists` types in the next-gen parquet reader ([#19493](https://github.com/NVIDIA/cudf/pull/19493)) [@mhaseeb123](https://github.com/mhaseeb123) +- Move test_search/test_scan/test_seriesmap.py to new cudf classic test directory structure ([#19492](https://github.com/NVIDIA/cudf/pull/19492)) [@mroeschke](https://github.com/mroeschke) +- Improve support for sliced input on from_arrow_host APIs ([#19491](https://github.com/NVIDIA/cudf/pull/19491)) [@davidwendt](https://github.com/davidwendt) +- Move test_avro/test_api_types.py and some DataFrame tests to new cudf classic test directory structure ([#19490](https://github.com/NVIDIA/cudf/pull/19490)) [@mroeschke](https://github.com/mroeschke) +- Move test_series.py to new cudf classic test directory structure ([#19485](https://github.com/NVIDIA/cudf/pull/19485)) [@mroeschke](https://github.com/mroeschke) +- Move test_testing.py to new cudf classic test directory structure ([#19481](https://github.com/NVIDIA/cudf/pull/19481)) [@mroeschke](https://github.com/mroeschke) +- Allow latest OS in devcontainers ([#19480](https://github.com/NVIDIA/cudf/pull/19480)) [@bdice](https://github.com/bdice) +- Move test_unaops/test_unique/test_transform.py to new cudf classic test directory structure ([#19477](https://github.com/NVIDIA/cudf/pull/19477)) [@mroeschke](https://github.com/mroeschke) +- Branch 25.10 merge branch 25.08 ([#19475](https://github.com/NVIDIA/cudf/pull/19475)) [@davidwendt](https://github.com/davidwendt) +- Use more pytest fixtures and clean data files cuDF classic tests subdirectories ([#19474](https://github.com/NVIDIA/cudf/pull/19474)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in test_binops/column/column_accessor/contains.py and more ([#19473](https://github.com/NVIDIA/cudf/pull/19473)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in test_csv/cuda_*/cut.py and more ([#19463](https://github.com/NVIDIA/cudf/pull/19463)) [@mroeschke](https://github.com/mroeschke) +- Improve readability when printing pylibcudf enums ([#19451](https://github.com/NVIDIA/cudf/pull/19451)) [@Matt711](https://github.com/Matt711) +- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19450](https://github.com/NVIDIA/cudf/pull/19450)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in test_dropna/factorize.py and more ([#19449](https://github.com/NVIDIA/cudf/pull/19449)) [@mroeschke](https://github.com/mroeschke) +- Update build infra to support new branching strategy ([#19445](https://github.com/NVIDIA/cudf/pull/19445)) [@robertmaynard](https://github.com/robertmaynard) +- Updated libcudf-example conda package to preserve directories structure ([#19440](https://github.com/NVIDIA/cudf/pull/19440)) [@Avinash-Raj](https://github.com/Avinash-Raj) +- Use more pytest fixtures and avoid GPU parameterization in test_groupby/index.py ([#19438](https://github.com/NVIDIA/cudf/pull/19438)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in test_indexing/joining/monotonic/multiindex.py ([#19437](https://github.com/NVIDIA/cudf/pull/19437)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19436](https://github.com/NVIDIA/cudf/pull/19436)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in test_query/rank/reduction/repr.py ([#19434](https://github.com/NVIDIA/cudf/pull/19434)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in test_replace/reshape/rolling.py ([#19426](https://github.com/NVIDIA/cudf/pull/19426)) [@mroeschke](https://github.com/mroeschke) +- Update s3 Bucket fixture creation in test_s3 ([#19424](https://github.com/NVIDIA/cudf/pull/19424)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19419](https://github.com/NVIDIA/cudf/pull/19419)) [@mroeschke](https://github.com/mroeschke) +- Fix various pandas test failures in `cudf.pandas` ([#19372](https://github.com/NVIDIA/cudf/pull/19372)) [@galipremsagar](https://github.com/galipremsagar) +- Pin Narwhals to 1.47 ([#19358](https://github.com/NVIDIA/cudf/pull/19358)) [@Matt711](https://github.com/Matt711) +- Run cudf-polars tests with all supported polars versions ([#19353](https://github.com/NVIDIA/cudf/pull/19353)) [@Matt711](https://github.com/Matt711) +- Update `pandas-tests-diff` to only display GPU/CPU usage metrics ([#19210](https://github.com/NVIDIA/cudf/pull/19210)) [@galipremsagar](https://github.com/galipremsagar) +- Use GCC 14 in conda builds. ([#19192](https://github.com/NVIDIA/cudf/pull/19192)) [@vyasr](https://github.com/vyasr) +- Use KvikIO's implementation of file-backed memory mapping ([#19164](https://github.com/NVIDIA/cudf/pull/19164)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Replace `rmm::device_scalar` with `cudf::detail::device_scalar` due to unnecessary synchronization (Part 3 of miss-sync) ([#19119](https://github.com/NVIDIA/cudf/pull/19119)) [@JigaoLuo](https://github.com/JigaoLuo) +- Implement distributed sorted for ``cudf_polars`` ([#18912](https://github.com/NVIDIA/cudf/pull/18912)) [@seberg](https://github.com/seberg) # cudf 25.08.00 (6 Aug 2025) ## 🚨 Breaking Changes -- Allow `np.dtype('object')` for cases that are valid ([#19478](https://github.com/rapidsai/cudf/pull/19478)) [@galipremsagar](https://github.com/galipremsagar) -- [FEA] Remove CUDA JIT-Compatibility Checks & CCCL WARs ([#19470](https://github.com/rapidsai/cudf/pull/19470)) [@lamarrr](https://github.com/lamarrr) -- Drop cuda 11 usages ([#19386](https://github.com/rapidsai/cudf/pull/19386)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate cudf::round for float types ([#19298](https://github.com/rapidsai/cudf/pull/19298)) [@davidwendt](https://github.com/davidwendt) -- Support output_dtype in cudf::reduce for nunique aggregation ([#19265](https://github.com/rapidsai/cudf/pull/19265)) [@davidwendt](https://github.com/davidwendt) -- Change default cudf-polars executor to "streaming" ([#19263](https://github.com/rapidsai/cudf/pull/19263)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix Handling of Complex Types in AST ([#19248](https://github.com/rapidsai/cudf/pull/19248)) [@lamarrr](https://github.com/lamarrr) -- Enable chunked reading of PQ sources with `>2B` rows ([#19245](https://github.com/rapidsai/cudf/pull/19245)) [@mhaseeb123](https://github.com/mhaseeb123) -- Refactor `grid_1d` class ([#19211](https://github.com/rapidsai/cudf/pull/19211)) [@lamarrr](https://github.com/lamarrr) -- Return valid for all-nulls in reduce() with nunique include-nulls aggregation ([#19196](https://github.com/rapidsai/cudf/pull/19196)) [@davidwendt](https://github.com/davidwendt) -- Refactor JNI error handling ([#19149](https://github.com/rapidsai/cudf/pull/19149)) [@ttnghia](https://github.com/ttnghia) -- Remove CUDA 11 from dependencies.yaml ([#19139](https://github.com/rapidsai/cudf/pull/19139)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Quick fixes of `modernize-use-constraints` rule ([#19105](https://github.com/rapidsai/cudf/pull/19105)) [@vuule](https://github.com/vuule) -- Filter Parquet row groups using row bounds ([#19082](https://github.com/rapidsai/cudf/pull/19082)) [@mhaseeb123](https://github.com/mhaseeb123) -- Temporarily revert "Refactor JNI error handling ([#18983)" (#19076](https://github.com/rapidsai/cudf/pull/18983)" (#19076)) [@abellina](https://github.com/abellina) -- Rename `parquet_chunked_writer` to `chunked_parquet_writer` for consistency with the reader ([#19047](https://github.com/rapidsai/cudf/pull/19047)) [@mhaseeb123](https://github.com/mhaseeb123) -- Compile libcudf using C++20 Standard ([#19045](https://github.com/rapidsai/cudf/pull/19045)) [@vuule](https://github.com/vuule) -- Refactor JNI error handling ([#18983](https://github.com/rapidsai/cudf/pull/18983)) [@ttnghia](https://github.com/ttnghia) -- stop uploading packages to downloads.rapids.ai ([#18973](https://github.com/rapidsai/cudf/pull/18973)) [@jameslamb](https://github.com/jameslamb) -- Remove deprecated Series methods, isclose ([#18947](https://github.com/rapidsai/cudf/pull/18947)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated groupby.collect ([#18946](https://github.com/rapidsai/cudf/pull/18946)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated get_dummies(cats=, ...) ([#18944](https://github.com/rapidsai/cudf/pull/18944)) [@mroeschke](https://github.com/mroeschke) -- Add pylibcudf.Column.from_arrow factory method ([#18937](https://github.com/rapidsai/cudf/pull/18937)) [@Matt711](https://github.com/Matt711) -- Add pylibcudf.Table.from_arrow factory method ([#18936](https://github.com/rapidsai/cudf/pull/18936)) [@Matt711](https://github.com/Matt711) -- Remove deprecated APIs ([#18933](https://github.com/rapidsai/cudf/pull/18933)) [@vuule](https://github.com/vuule) -- Remove cudf.Scalar ([#18927](https://github.com/rapidsai/cudf/pull/18927)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated `cudf::io::host_buffer` ([#18881](https://github.com/rapidsai/cudf/pull/18881)) [@Matt711](https://github.com/Matt711) -- Null-handling for Transforms ([#18845](https://github.com/rapidsai/cudf/pull/18845)) [@lamarrr](https://github.com/lamarrr) -- Enable `skip_rows` in the chunked parquet reader. ([#18130](https://github.com/rapidsai/cudf/pull/18130)) [@mhaseeb123](https://github.com/mhaseeb123) +- Allow `np.dtype('object')` for cases that are valid ([#19478](https://github.com/NVIDIA/cudf/pull/19478)) [@galipremsagar](https://github.com/galipremsagar) +- [FEA] Remove CUDA JIT-Compatibility Checks & CCCL WARs ([#19470](https://github.com/NVIDIA/cudf/pull/19470)) [@lamarrr](https://github.com/lamarrr) +- Drop cuda 11 usages ([#19386](https://github.com/NVIDIA/cudf/pull/19386)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate cudf::round for float types ([#19298](https://github.com/NVIDIA/cudf/pull/19298)) [@davidwendt](https://github.com/davidwendt) +- Support output_dtype in cudf::reduce for nunique aggregation ([#19265](https://github.com/NVIDIA/cudf/pull/19265)) [@davidwendt](https://github.com/davidwendt) +- Change default cudf-polars executor to "streaming" ([#19263](https://github.com/NVIDIA/cudf/pull/19263)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix Handling of Complex Types in AST ([#19248](https://github.com/NVIDIA/cudf/pull/19248)) [@lamarrr](https://github.com/lamarrr) +- Enable chunked reading of PQ sources with `>2B` rows ([#19245](https://github.com/NVIDIA/cudf/pull/19245)) [@mhaseeb123](https://github.com/mhaseeb123) +- Refactor `grid_1d` class ([#19211](https://github.com/NVIDIA/cudf/pull/19211)) [@lamarrr](https://github.com/lamarrr) +- Return valid for all-nulls in reduce() with nunique include-nulls aggregation ([#19196](https://github.com/NVIDIA/cudf/pull/19196)) [@davidwendt](https://github.com/davidwendt) +- Refactor JNI error handling ([#19149](https://github.com/NVIDIA/cudf/pull/19149)) [@ttnghia](https://github.com/ttnghia) +- Remove CUDA 11 from dependencies.yaml ([#19139](https://github.com/NVIDIA/cudf/pull/19139)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Quick fixes of `modernize-use-constraints` rule ([#19105](https://github.com/NVIDIA/cudf/pull/19105)) [@vuule](https://github.com/vuule) +- Filter Parquet row groups using row bounds ([#19082](https://github.com/NVIDIA/cudf/pull/19082)) [@mhaseeb123](https://github.com/mhaseeb123) +- Temporarily revert "Refactor JNI error handling ([#18983)" (#19076](https://github.com/NVIDIA/cudf/pull/18983)" (#19076)) [@abellina](https://github.com/abellina) +- Rename `parquet_chunked_writer` to `chunked_parquet_writer` for consistency with the reader ([#19047](https://github.com/NVIDIA/cudf/pull/19047)) [@mhaseeb123](https://github.com/mhaseeb123) +- Compile libcudf using C++20 Standard ([#19045](https://github.com/NVIDIA/cudf/pull/19045)) [@vuule](https://github.com/vuule) +- Refactor JNI error handling ([#18983](https://github.com/NVIDIA/cudf/pull/18983)) [@ttnghia](https://github.com/ttnghia) +- stop uploading packages to downloads.rapids.ai ([#18973](https://github.com/NVIDIA/cudf/pull/18973)) [@jameslamb](https://github.com/jameslamb) +- Remove deprecated Series methods, isclose ([#18947](https://github.com/NVIDIA/cudf/pull/18947)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated groupby.collect ([#18946](https://github.com/NVIDIA/cudf/pull/18946)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated get_dummies(cats=, ...) ([#18944](https://github.com/NVIDIA/cudf/pull/18944)) [@mroeschke](https://github.com/mroeschke) +- Add pylibcudf.Column.from_arrow factory method ([#18937](https://github.com/NVIDIA/cudf/pull/18937)) [@Matt711](https://github.com/Matt711) +- Add pylibcudf.Table.from_arrow factory method ([#18936](https://github.com/NVIDIA/cudf/pull/18936)) [@Matt711](https://github.com/Matt711) +- Remove deprecated APIs ([#18933](https://github.com/NVIDIA/cudf/pull/18933)) [@vuule](https://github.com/vuule) +- Remove cudf.Scalar ([#18927](https://github.com/NVIDIA/cudf/pull/18927)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated `cudf::io::host_buffer` ([#18881](https://github.com/NVIDIA/cudf/pull/18881)) [@Matt711](https://github.com/Matt711) +- Null-handling for Transforms ([#18845](https://github.com/NVIDIA/cudf/pull/18845)) [@lamarrr](https://github.com/lamarrr) +- Enable `skip_rows` in the chunked parquet reader. ([#18130](https://github.com/NVIDIA/cudf/pull/18130)) [@mhaseeb123](https://github.com/mhaseeb123) ## 🐛 Bug Fixes -- Increase alignment requirement for parquet bloom filter to 256 ([#19595](https://github.com/rapidsai/cudf/pull/19595)) [@mhaseeb123](https://github.com/mhaseeb123) -- Revert "Add primitive row dispatch support for semi/anti join and cudf::contains" ([#19503](https://github.com/rapidsai/cudf/pull/19503)) [@PointKernel](https://github.com/PointKernel) -- Allow `np.dtype('object')` for cases that are valid ([#19478](https://github.com/rapidsai/cudf/pull/19478)) [@galipremsagar](https://github.com/galipremsagar) -- Add conda dependency on nvidia-ml-py. ([#19454](https://github.com/rapidsai/cudf/pull/19454)) [@bdice](https://github.com/bdice) -- Mark `cudf.pandas` notebook repr test as flaky ([#19441](https://github.com/rapidsai/cudf/pull/19441)) [@Matt711](https://github.com/Matt711) -- Fix pytest to properly expose a bug ([#19433](https://github.com/rapidsai/cudf/pull/19433)) [@galipremsagar](https://github.com/galipremsagar) -- Switch from `thrust::sort` to `cub::DeviceRadixSort` in Parquet chunked reader ([#19414](https://github.com/rapidsai/cudf/pull/19414)) [@ttnghia](https://github.com/ttnghia) -- Use numba-cuda>=0.15.2,<0.16 ([#19413](https://github.com/rapidsai/cudf/pull/19413)) [@bdice](https://github.com/bdice) -- Update String Transform Examples ([#19407](https://github.com/rapidsai/cudf/pull/19407)) [@lamarrr](https://github.com/lamarrr) -- [BUG] Make floor division and modulo by 0 match CPU polars ([#19406](https://github.com/rapidsai/cudf/pull/19406)) [@Matt711](https://github.com/Matt711) -- Handle empty input in cudf::strings::extract APIs ([#19398](https://github.com/rapidsai/cudf/pull/19398)) [@davidwendt](https://github.com/davidwendt) -- Fix jitify error on exit from FILTER_TEST ([#19395](https://github.com/rapidsai/cudf/pull/19395)) [@davidwendt](https://github.com/davidwendt) -- Update cudf.pandas tests to silence deprecation warnings ([#19377](https://github.com/rapidsai/cudf/pull/19377)) [@Matt711](https://github.com/Matt711) -- Replace sprintf with snprintf in libcudf parquet tests ([#19371](https://github.com/rapidsai/cudf/pull/19371)) [@davidwendt](https://github.com/davidwendt) -- Make DateOffset respect timezone ([#19366](https://github.com/rapidsai/cudf/pull/19366)) [@Matt711](https://github.com/Matt711) -- Fix flaky tests in `cudf.pandas` ([#19345](https://github.com/rapidsai/cudf/pull/19345)) [@TomAugspurger](https://github.com/TomAugspurger) -- Update protocol choices for ucxx in PDSH benchmark ([#19343](https://github.com/rapidsai/cudf/pull/19343)) [@TomAugspurger](https://github.com/TomAugspurger) -- Remove passing pandas tests from xfail list ([#19341](https://github.com/rapidsai/cudf/pull/19341)) [@Matt711](https://github.com/Matt711) -- Fix Union-Slice bug ([#19336](https://github.com/rapidsai/cudf/pull/19336)) [@Matt711](https://github.com/Matt711) -- Fix bit shift overflow in segmented_offset_bitmask_binop utility ([#19329](https://github.com/rapidsai/cudf/pull/19329)) [@davidwendt](https://github.com/davidwendt) -- Fix job filters for `pandas-tests` ([#19322](https://github.com/rapidsai/cudf/pull/19322)) [@galipremsagar](https://github.com/galipremsagar) -- Fix compile warning in interop_stringview.cpp ([#19320](https://github.com/rapidsai/cudf/pull/19320)) [@davidwendt](https://github.com/davidwendt) -- Fix a use-after-free issue in TDigest aggregation code. ([#19311](https://github.com/rapidsai/cudf/pull/19311)) [@nvdbaranec](https://github.com/nvdbaranec) -- Always represent datetime aware data as UTC in strftime ([#19304](https://github.com/rapidsai/cudf/pull/19304)) [@mroeschke](https://github.com/mroeschke) -- Do not pass cupy objects objects to numba kernels directly ([#19283](https://github.com/rapidsai/cudf/pull/19283)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Correct docstring for `DataFrame.apply` to match code ([#19262](https://github.com/rapidsai/cudf/pull/19262)) [@dagardner-nv](https://github.com/dagardner-nv) -- Cast `n_unique` aggregation result to match polars ([#19256](https://github.com/rapidsai/cudf/pull/19256)) [@Matt711](https://github.com/Matt711) -- Fix Handling of Complex Types in AST ([#19248](https://github.com/rapidsai/cudf/pull/19248)) [@lamarrr](https://github.com/lamarrr) -- Add missing include ([#19239](https://github.com/rapidsai/cudf/pull/19239)) [@vyasr](https://github.com/vyasr) -- Raised `MixedTypeErrors` for condition that lead to mixed types ([#19232](https://github.com/rapidsai/cudf/pull/19232)) [@galipremsagar](https://github.com/galipremsagar) -- Fix errors in the nvCOMP adapter ([#19221](https://github.com/rapidsai/cudf/pull/19221)) [@vuule](https://github.com/vuule) -- Remove nvToolsExt usage ([#19209](https://github.com/rapidsai/cudf/pull/19209)) [@vyasr](https://github.com/vyasr) -- Fix a pair of bugs in get_decompression_scratch() size. ([#19207](https://github.com/rapidsai/cudf/pull/19207)) [@nvdbaranec](https://github.com/nvdbaranec) -- Allow `is_list_like` to return correct values by disabling it ([#19188](https://github.com/rapidsai/cudf/pull/19188)) [@galipremsagar](https://github.com/galipremsagar) -- Fix slicing after `Join` and `GroupBy` in streaming cudf-polars ([#19187](https://github.com/rapidsai/cudf/pull/19187)) [@rjzamora](https://github.com/rjzamora) -- Fix `binops` type preservation for some dtypes ([#19183](https://github.com/rapidsai/cudf/pull/19183)) [@galipremsagar](https://github.com/galipremsagar) -- Fix streaming `GroupBy` on non-trivial keys ([#19181](https://github.com/rapidsai/cudf/pull/19181)) [@rjzamora](https://github.com/rjzamora) -- Fix bitmask in from_arrow_host for sliced stringview type ([#19174](https://github.com/rapidsai/cudf/pull/19174)) [@davidwendt](https://github.com/davidwendt) -- Fixed group_by mean with missing values and multiple partitions ([#19165](https://github.com/rapidsai/cudf/pull/19165)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add fallback to ``HStack`` lowering in cudf-polars ([#19163](https://github.com/rapidsai/cudf/pull/19163)) [@rjzamora](https://github.com/rjzamora) -- Fix ``Literal`` partitioning in cudf-polars ([#19160](https://github.com/rapidsai/cudf/pull/19160)) [@rjzamora](https://github.com/rjzamora) -- Fix `from_array_interface` for empty arrays ([#19144](https://github.com/rapidsai/cudf/pull/19144)) [@Matt711](https://github.com/Matt711) -- Adding GH_TOKEN pass-through to summarize job ([#19143](https://github.com/rapidsai/cudf/pull/19143)) [@msarahan](https://github.com/msarahan) -- Fix hash collision in Union([MapFunction]) ([#19124](https://github.com/rapidsai/cudf/pull/19124)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix bug in ``group_by().n_unique()`` in streaming cudf-polars ([#19108](https://github.com/rapidsai/cudf/pull/19108)) [@rjzamora](https://github.com/rjzamora) -- Parse (non-MultiIndex) label-based keys to structured data ([#19103](https://github.com/rapidsai/cudf/pull/19103)) [@mroeschke](https://github.com/mroeschke) -- Fix cudf_polars spilling ([#19101](https://github.com/rapidsai/cudf/pull/19101)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix libcudf strings case logic to set null-row size to zero ([#19095](https://github.com/rapidsai/cudf/pull/19095)) [@davidwendt](https://github.com/davidwendt) -- Temporarily revert "Refactor JNI error handling ([#18983)" (#19076](https://github.com/rapidsai/cudf/pull/18983)" (#19076)) [@abellina](https://github.com/abellina) -- Temporary workaround for incorrect ``SplitScan`` results in cuDF-Polars ([#19071](https://github.com/rapidsai/cudf/pull/19071)) [@rjzamora](https://github.com/rjzamora) -- Use default memory resource for JSON_QUOTE_NORMALIZATION gtests ([#19057](https://github.com/rapidsai/cudf/pull/19057)) [@davidwendt](https://github.com/davidwendt) -- Added null-probability to polynomial benchmarks and fixed transform call-sites ([#18972](https://github.com/rapidsai/cudf/pull/18972)) [@lamarrr](https://github.com/lamarrr) -- Fix flaky custreamz test ([#18961](https://github.com/rapidsai/cudf/pull/18961)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix tdigest percentile correctness for low row-counts ([#18952](https://github.com/rapidsai/cudf/pull/18952)) [@mythrocks](https://github.com/mythrocks) -- Enable `skip_rows` in the chunked parquet reader. ([#18130](https://github.com/rapidsai/cudf/pull/18130)) [@mhaseeb123](https://github.com/mhaseeb123) +- Increase alignment requirement for parquet bloom filter to 256 ([#19595](https://github.com/NVIDIA/cudf/pull/19595)) [@mhaseeb123](https://github.com/mhaseeb123) +- Revert "Add primitive row dispatch support for semi/anti join and cudf::contains" ([#19503](https://github.com/NVIDIA/cudf/pull/19503)) [@PointKernel](https://github.com/PointKernel) +- Allow `np.dtype('object')` for cases that are valid ([#19478](https://github.com/NVIDIA/cudf/pull/19478)) [@galipremsagar](https://github.com/galipremsagar) +- Add conda dependency on nvidia-ml-py. ([#19454](https://github.com/NVIDIA/cudf/pull/19454)) [@bdice](https://github.com/bdice) +- Mark `cudf.pandas` notebook repr test as flaky ([#19441](https://github.com/NVIDIA/cudf/pull/19441)) [@Matt711](https://github.com/Matt711) +- Fix pytest to properly expose a bug ([#19433](https://github.com/NVIDIA/cudf/pull/19433)) [@galipremsagar](https://github.com/galipremsagar) +- Switch from `thrust::sort` to `cub::DeviceRadixSort` in Parquet chunked reader ([#19414](https://github.com/NVIDIA/cudf/pull/19414)) [@ttnghia](https://github.com/ttnghia) +- Use numba-cuda>=0.15.2,<0.16 ([#19413](https://github.com/NVIDIA/cudf/pull/19413)) [@bdice](https://github.com/bdice) +- Update String Transform Examples ([#19407](https://github.com/NVIDIA/cudf/pull/19407)) [@lamarrr](https://github.com/lamarrr) +- [BUG] Make floor division and modulo by 0 match CPU polars ([#19406](https://github.com/NVIDIA/cudf/pull/19406)) [@Matt711](https://github.com/Matt711) +- Handle empty input in cudf::strings::extract APIs ([#19398](https://github.com/NVIDIA/cudf/pull/19398)) [@davidwendt](https://github.com/davidwendt) +- Fix jitify error on exit from FILTER_TEST ([#19395](https://github.com/NVIDIA/cudf/pull/19395)) [@davidwendt](https://github.com/davidwendt) +- Update cudf.pandas tests to silence deprecation warnings ([#19377](https://github.com/NVIDIA/cudf/pull/19377)) [@Matt711](https://github.com/Matt711) +- Replace sprintf with snprintf in libcudf parquet tests ([#19371](https://github.com/NVIDIA/cudf/pull/19371)) [@davidwendt](https://github.com/davidwendt) +- Make DateOffset respect timezone ([#19366](https://github.com/NVIDIA/cudf/pull/19366)) [@Matt711](https://github.com/Matt711) +- Fix flaky tests in `cudf.pandas` ([#19345](https://github.com/NVIDIA/cudf/pull/19345)) [@TomAugspurger](https://github.com/TomAugspurger) +- Update protocol choices for ucxx in PDSH benchmark ([#19343](https://github.com/NVIDIA/cudf/pull/19343)) [@TomAugspurger](https://github.com/TomAugspurger) +- Remove passing pandas tests from xfail list ([#19341](https://github.com/NVIDIA/cudf/pull/19341)) [@Matt711](https://github.com/Matt711) +- Fix Union-Slice bug ([#19336](https://github.com/NVIDIA/cudf/pull/19336)) [@Matt711](https://github.com/Matt711) +- Fix bit shift overflow in segmented_offset_bitmask_binop utility ([#19329](https://github.com/NVIDIA/cudf/pull/19329)) [@davidwendt](https://github.com/davidwendt) +- Fix job filters for `pandas-tests` ([#19322](https://github.com/NVIDIA/cudf/pull/19322)) [@galipremsagar](https://github.com/galipremsagar) +- Fix compile warning in interop_stringview.cpp ([#19320](https://github.com/NVIDIA/cudf/pull/19320)) [@davidwendt](https://github.com/davidwendt) +- Fix a use-after-free issue in TDigest aggregation code. ([#19311](https://github.com/NVIDIA/cudf/pull/19311)) [@nvdbaranec](https://github.com/nvdbaranec) +- Always represent datetime aware data as UTC in strftime ([#19304](https://github.com/NVIDIA/cudf/pull/19304)) [@mroeschke](https://github.com/mroeschke) +- Do not pass cupy objects objects to numba kernels directly ([#19283](https://github.com/NVIDIA/cudf/pull/19283)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Correct docstring for `DataFrame.apply` to match code ([#19262](https://github.com/NVIDIA/cudf/pull/19262)) [@dagardner-nv](https://github.com/dagardner-nv) +- Cast `n_unique` aggregation result to match polars ([#19256](https://github.com/NVIDIA/cudf/pull/19256)) [@Matt711](https://github.com/Matt711) +- Fix Handling of Complex Types in AST ([#19248](https://github.com/NVIDIA/cudf/pull/19248)) [@lamarrr](https://github.com/lamarrr) +- Add missing include ([#19239](https://github.com/NVIDIA/cudf/pull/19239)) [@vyasr](https://github.com/vyasr) +- Raised `MixedTypeErrors` for condition that lead to mixed types ([#19232](https://github.com/NVIDIA/cudf/pull/19232)) [@galipremsagar](https://github.com/galipremsagar) +- Fix errors in the nvCOMP adapter ([#19221](https://github.com/NVIDIA/cudf/pull/19221)) [@vuule](https://github.com/vuule) +- Remove nvToolsExt usage ([#19209](https://github.com/NVIDIA/cudf/pull/19209)) [@vyasr](https://github.com/vyasr) +- Fix a pair of bugs in get_decompression_scratch() size. ([#19207](https://github.com/NVIDIA/cudf/pull/19207)) [@nvdbaranec](https://github.com/nvdbaranec) +- Allow `is_list_like` to return correct values by disabling it ([#19188](https://github.com/NVIDIA/cudf/pull/19188)) [@galipremsagar](https://github.com/galipremsagar) +- Fix slicing after `Join` and `GroupBy` in streaming cudf-polars ([#19187](https://github.com/NVIDIA/cudf/pull/19187)) [@rjzamora](https://github.com/rjzamora) +- Fix `binops` type preservation for some dtypes ([#19183](https://github.com/NVIDIA/cudf/pull/19183)) [@galipremsagar](https://github.com/galipremsagar) +- Fix streaming `GroupBy` on non-trivial keys ([#19181](https://github.com/NVIDIA/cudf/pull/19181)) [@rjzamora](https://github.com/rjzamora) +- Fix bitmask in from_arrow_host for sliced stringview type ([#19174](https://github.com/NVIDIA/cudf/pull/19174)) [@davidwendt](https://github.com/davidwendt) +- Fixed group_by mean with missing values and multiple partitions ([#19165](https://github.com/NVIDIA/cudf/pull/19165)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add fallback to ``HStack`` lowering in cudf-polars ([#19163](https://github.com/NVIDIA/cudf/pull/19163)) [@rjzamora](https://github.com/rjzamora) +- Fix ``Literal`` partitioning in cudf-polars ([#19160](https://github.com/NVIDIA/cudf/pull/19160)) [@rjzamora](https://github.com/rjzamora) +- Fix `from_array_interface` for empty arrays ([#19144](https://github.com/NVIDIA/cudf/pull/19144)) [@Matt711](https://github.com/Matt711) +- Adding GH_TOKEN pass-through to summarize job ([#19143](https://github.com/NVIDIA/cudf/pull/19143)) [@msarahan](https://github.com/msarahan) +- Fix hash collision in Union([MapFunction]) ([#19124](https://github.com/NVIDIA/cudf/pull/19124)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix bug in ``group_by().n_unique()`` in streaming cudf-polars ([#19108](https://github.com/NVIDIA/cudf/pull/19108)) [@rjzamora](https://github.com/rjzamora) +- Parse (non-MultiIndex) label-based keys to structured data ([#19103](https://github.com/NVIDIA/cudf/pull/19103)) [@mroeschke](https://github.com/mroeschke) +- Fix cudf_polars spilling ([#19101](https://github.com/NVIDIA/cudf/pull/19101)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix libcudf strings case logic to set null-row size to zero ([#19095](https://github.com/NVIDIA/cudf/pull/19095)) [@davidwendt](https://github.com/davidwendt) +- Temporarily revert "Refactor JNI error handling ([#18983)" (#19076](https://github.com/NVIDIA/cudf/pull/18983)" (#19076)) [@abellina](https://github.com/abellina) +- Temporary workaround for incorrect ``SplitScan`` results in cuDF-Polars ([#19071](https://github.com/NVIDIA/cudf/pull/19071)) [@rjzamora](https://github.com/rjzamora) +- Use default memory resource for JSON_QUOTE_NORMALIZATION gtests ([#19057](https://github.com/NVIDIA/cudf/pull/19057)) [@davidwendt](https://github.com/davidwendt) +- Added null-probability to polynomial benchmarks and fixed transform call-sites ([#18972](https://github.com/NVIDIA/cudf/pull/18972)) [@lamarrr](https://github.com/lamarrr) +- Fix flaky custreamz test ([#18961](https://github.com/NVIDIA/cudf/pull/18961)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix tdigest percentile correctness for low row-counts ([#18952](https://github.com/NVIDIA/cudf/pull/18952)) [@mythrocks](https://github.com/mythrocks) +- Enable `skip_rows` in the chunked parquet reader. ([#18130](https://github.com/NVIDIA/cudf/pull/18130)) [@mhaseeb123](https://github.com/mhaseeb123) ## 📖 Documentation -- Update conda environment file for CUDA 12.9 compatibility ([#19376](https://github.com/rapidsai/cudf/pull/19376)) [@a-hirota](https://github.com/a-hirota) -- Update recommended gcc version in contibuting guide ([#19365](https://github.com/rapidsai/cudf/pull/19365)) [@davidwendt](https://github.com/davidwendt) -- Autodoc DateOffset ([#19297](https://github.com/rapidsai/cudf/pull/19297)) [@wence-](https://github.com/wence-) -- Fix cudf::column_device_view::element() doxygen ([#19296](https://github.com/rapidsai/cudf/pull/19296)) [@davidwendt](https://github.com/davidwendt) -- Document aggregations for cudf::reduce in doxygen ([#19264](https://github.com/rapidsai/cudf/pull/19264)) [@davidwendt](https://github.com/davidwendt) -- add docs on CI workflow inputs ([#19234](https://github.com/rapidsai/cudf/pull/19234)) [@jameslamb](https://github.com/jameslamb) -- Update README and CONTRIBUTING to reflect new CUDA requirements ([#19138](https://github.com/rapidsai/cudf/pull/19138)) [@PointKernel](https://github.com/PointKernel) -- Remove the extra index URL for CUDA 12 ([#19128](https://github.com/rapidsai/cudf/pull/19128)) [@vyasr](https://github.com/vyasr) -- Improve WordPieceVocabulary.tokenize documentation ([#19098](https://github.com/rapidsai/cudf/pull/19098)) [@davidwendt](https://github.com/davidwendt) -- Add some basic streaming engine documentation ([#19088](https://github.com/rapidsai/cudf/pull/19088)) [@wence-](https://github.com/wence-) -- Update the contributing guide to include pylibcudf in the build command ([#19011](https://github.com/rapidsai/cudf/pull/19011)) [@Matt711](https://github.com/Matt711) -- Fix pylibcudf docs for some strings APIs ([#19004](https://github.com/rapidsai/cudf/pull/19004)) [@davidwendt](https://github.com/davidwendt) -- Update cuDF Python library design with BaseIndex and pylibcudf updates ([#18903](https://github.com/rapidsai/cudf/pull/18903)) [@mroeschke](https://github.com/mroeschke) +- Update conda environment file for CUDA 12.9 compatibility ([#19376](https://github.com/NVIDIA/cudf/pull/19376)) [@a-hirota](https://github.com/a-hirota) +- Update recommended gcc version in contibuting guide ([#19365](https://github.com/NVIDIA/cudf/pull/19365)) [@davidwendt](https://github.com/davidwendt) +- Autodoc DateOffset ([#19297](https://github.com/NVIDIA/cudf/pull/19297)) [@wence-](https://github.com/wence-) +- Fix cudf::column_device_view::element() doxygen ([#19296](https://github.com/NVIDIA/cudf/pull/19296)) [@davidwendt](https://github.com/davidwendt) +- Document aggregations for cudf::reduce in doxygen ([#19264](https://github.com/NVIDIA/cudf/pull/19264)) [@davidwendt](https://github.com/davidwendt) +- add docs on CI workflow inputs ([#19234](https://github.com/NVIDIA/cudf/pull/19234)) [@jameslamb](https://github.com/jameslamb) +- Update README and CONTRIBUTING to reflect new CUDA requirements ([#19138](https://github.com/NVIDIA/cudf/pull/19138)) [@PointKernel](https://github.com/PointKernel) +- Remove the extra index URL for CUDA 12 ([#19128](https://github.com/NVIDIA/cudf/pull/19128)) [@vyasr](https://github.com/vyasr) +- Improve WordPieceVocabulary.tokenize documentation ([#19098](https://github.com/NVIDIA/cudf/pull/19098)) [@davidwendt](https://github.com/davidwendt) +- Add some basic streaming engine documentation ([#19088](https://github.com/NVIDIA/cudf/pull/19088)) [@wence-](https://github.com/wence-) +- Update the contributing guide to include pylibcudf in the build command ([#19011](https://github.com/NVIDIA/cudf/pull/19011)) [@Matt711](https://github.com/Matt711) +- Fix pylibcudf docs for some strings APIs ([#19004](https://github.com/NVIDIA/cudf/pull/19004)) [@davidwendt](https://github.com/davidwendt) +- Update cuDF Python library design with BaseIndex and pylibcudf updates ([#18903](https://github.com/NVIDIA/cudf/pull/18903)) [@mroeschke](https://github.com/mroeschke) ## 🚀 New Features -- Avoid using UVM on systems without a traditional memory resource ([#19444](https://github.com/rapidsai/cudf/pull/19444)) [@Matt711](https://github.com/Matt711) -- Add parquet-sampling configuration options ([#19423](https://github.com/rapidsai/cudf/pull/19423)) [@rjzamora](https://github.com/rjzamora) -- Add new JSON reader interface accepting string column input to pylibcudf ([#19400](https://github.com/rapidsai/cudf/pull/19400)) [@shrshi](https://github.com/shrshi) -- Add a parquet reader utility to update output null masks ([#19370](https://github.com/rapidsai/cudf/pull/19370)) [@mhaseeb123](https://github.com/mhaseeb123) -- Build and ship `shim.cu` file as LTOIR ([#19368](https://github.com/rapidsai/cudf/pull/19368)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add cudf::strings::find_instance API ([#19326](https://github.com/rapidsai/cudf/pull/19326)) [@davidwendt](https://github.com/davidwendt) -- Add single-file streaming ``Sink`` support ([#19317](https://github.com/rapidsai/cudf/pull/19317)) [@rjzamora](https://github.com/rjzamora) -- Support null_count expression ([#19314](https://github.com/rapidsai/cudf/pull/19314)) [@Matt711](https://github.com/Matt711) -- Materialize tables in the experimental Parquet reader ([#19308](https://github.com/rapidsai/cudf/pull/19308)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add new cudf::top_k API ([#19303](https://github.com/rapidsai/cudf/pull/19303)) [@davidwendt](https://github.com/davidwendt) -- Add cudf::strings::split_part API ([#19289](https://github.com/rapidsai/cudf/pull/19289)) [@davidwendt](https://github.com/davidwendt) -- Support output_dtype in cudf::reduce for nunique aggregation ([#19265](https://github.com/rapidsai/cudf/pull/19265)) [@davidwendt](https://github.com/davidwendt) -- Add ``post_traversal`` API to cudf-polars ([#19258](https://github.com/rapidsai/cudf/pull/19258)) [@rjzamora](https://github.com/rjzamora) -- Deprecate `DataFrame.apply_rows` ([#19218](https://github.com/rapidsai/cudf/pull/19218)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Require `numba-cuda>=0.16.0` ([#19213](https://github.com/rapidsai/cudf/pull/19213)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add a mode to co-process decompression and compression on host and device ([#19203](https://github.com/rapidsai/cudf/pull/19203)) [@vuule](https://github.com/vuule) -- Return valid for all-nulls in reduce() with nunique include-nulls aggregation ([#19196](https://github.com/rapidsai/cudf/pull/19196)) [@davidwendt](https://github.com/davidwendt) -- Refactor JNI error handling ([#19149](https://github.com/rapidsai/cudf/pull/19149)) [@ttnghia](https://github.com/ttnghia) -- Add support for horizontal string concatenation `pl.concat_str` ([#19142](https://github.com/rapidsai/cudf/pull/19142)) [@Matt711](https://github.com/Matt711) -- Add PDS-DS Query 1 ([#19131](https://github.com/rapidsai/cudf/pull/19131)) [@Matt711](https://github.com/Matt711) -- Support `cudf-polars` `str.reverse` ([#19117](https://github.com/rapidsai/cudf/pull/19117)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `cudf-polars` `str.pad_end` and `str.pad_start` ([#19116](https://github.com/rapidsai/cudf/pull/19116)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `cudf-polars` `str.head` and `str.tail` ([#19115](https://github.com/rapidsai/cudf/pull/19115)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `cudf-polars` `str.to_titlecase` ([#19114](https://github.com/rapidsai/cudf/pull/19114)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add `cudf/io/codec.hpp` to expose compression/decompression APIs ([#19113](https://github.com/rapidsai/cudf/pull/19113)) [@ttnghia](https://github.com/ttnghia) -- Support converting decimals to/from pylibcudf scalars ([#19106](https://github.com/rapidsai/cudf/pull/19106)) [@Matt711](https://github.com/Matt711) -- Support resource-constrained sort-merge inner join operation through left table partitioning ([#19102](https://github.com/rapidsai/cudf/pull/19102)) [@shrshi](https://github.com/shrshi) -- Filter Parquet row groups using row bounds ([#19082](https://github.com/rapidsai/cudf/pull/19082)) [@mhaseeb123](https://github.com/mhaseeb123) -- Implement UDF Filters ([#19070](https://github.com/rapidsai/cudf/pull/19070)) [@lamarrr](https://github.com/lamarrr) -- Move the remaining libcudf pieces to C++20 ([#19065](https://github.com/rapidsai/cudf/pull/19065)) [@vuule](https://github.com/vuule) -- Allow using a stream per thread at runtime ([#19051](https://github.com/rapidsai/cudf/pull/19051)) [@vyasr](https://github.com/vyasr) -- Remove stacktrace retrieval code ([#19048](https://github.com/rapidsai/cudf/pull/19048)) [@ttnghia](https://github.com/ttnghia) -- Compile libcudf using C++20 Standard ([#19045](https://github.com/rapidsai/cudf/pull/19045)) [@vuule](https://github.com/vuule) -- String Transform Examples: Added Branching, Public API Versions, and Sampling ([#19038](https://github.com/rapidsai/cudf/pull/19038)) [@lamarrr](https://github.com/lamarrr) -- Refactor JNI error handling ([#18983](https://github.com/rapidsai/cudf/pull/18983)) [@ttnghia](https://github.com/ttnghia) -- Add basic ``Sink`` support for streaming cudf-polars executor ([#18963](https://github.com/rapidsai/cudf/pull/18963)) [@rjzamora](https://github.com/rjzamora) -- Fix debug-build Failure in JIT Tests ([#18939](https://github.com/rapidsai/cudf/pull/18939)) [@lamarrr](https://github.com/lamarrr) -- Add from_arrow factory methods for Scalar and DataType ([#18938](https://github.com/rapidsai/cudf/pull/18938)) [@Matt711](https://github.com/Matt711) -- Add pylibcudf.Column.from_arrow factory method ([#18937](https://github.com/rapidsai/cudf/pull/18937)) [@Matt711](https://github.com/Matt711) -- Add pylibcudf.Table.from_arrow factory method ([#18936](https://github.com/rapidsai/cudf/pull/18936)) [@Matt711](https://github.com/Matt711) -- Update nvCOMP adapter ([#18931](https://github.com/rapidsai/cudf/pull/18931)) [@vuule](https://github.com/vuule) -- Create a pylibcudf Column from a iterable of python strings ([#18916](https://github.com/rapidsai/cudf/pull/18916)) [@Matt711](https://github.com/Matt711) -- Add CLI argument to enable OOM protection in PDS-H ([#18914](https://github.com/rapidsai/cudf/pull/18914)) [@pentschev](https://github.com/pentschev) -- Implement data page pruning using Parquet page index stats ([#18873](https://github.com/rapidsai/cudf/pull/18873)) [@mhaseeb123](https://github.com/mhaseeb123) -- Null-handling for Transforms ([#18845](https://github.com/rapidsai/cudf/pull/18845)) [@lamarrr](https://github.com/lamarrr) -- Implement row group pruning with dictionaries in experimental PQ reader ([#18836](https://github.com/rapidsai/cudf/pull/18836)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add support for parquet scan + count operation ([#18463](https://github.com/rapidsai/cudf/pull/18463)) [@Matt711](https://github.com/Matt711) -- Manage strings with NRT ([#18453](https://github.com/rapidsai/cudf/pull/18453)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Avoid using UVM on systems without a traditional memory resource ([#19444](https://github.com/NVIDIA/cudf/pull/19444)) [@Matt711](https://github.com/Matt711) +- Add parquet-sampling configuration options ([#19423](https://github.com/NVIDIA/cudf/pull/19423)) [@rjzamora](https://github.com/rjzamora) +- Add new JSON reader interface accepting string column input to pylibcudf ([#19400](https://github.com/NVIDIA/cudf/pull/19400)) [@shrshi](https://github.com/shrshi) +- Add a parquet reader utility to update output null masks ([#19370](https://github.com/NVIDIA/cudf/pull/19370)) [@mhaseeb123](https://github.com/mhaseeb123) +- Build and ship `shim.cu` file as LTOIR ([#19368](https://github.com/NVIDIA/cudf/pull/19368)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add cudf::strings::find_instance API ([#19326](https://github.com/NVIDIA/cudf/pull/19326)) [@davidwendt](https://github.com/davidwendt) +- Add single-file streaming ``Sink`` support ([#19317](https://github.com/NVIDIA/cudf/pull/19317)) [@rjzamora](https://github.com/rjzamora) +- Support null_count expression ([#19314](https://github.com/NVIDIA/cudf/pull/19314)) [@Matt711](https://github.com/Matt711) +- Materialize tables in the experimental Parquet reader ([#19308](https://github.com/NVIDIA/cudf/pull/19308)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add new cudf::top_k API ([#19303](https://github.com/NVIDIA/cudf/pull/19303)) [@davidwendt](https://github.com/davidwendt) +- Add cudf::strings::split_part API ([#19289](https://github.com/NVIDIA/cudf/pull/19289)) [@davidwendt](https://github.com/davidwendt) +- Support output_dtype in cudf::reduce for nunique aggregation ([#19265](https://github.com/NVIDIA/cudf/pull/19265)) [@davidwendt](https://github.com/davidwendt) +- Add ``post_traversal`` API to cudf-polars ([#19258](https://github.com/NVIDIA/cudf/pull/19258)) [@rjzamora](https://github.com/rjzamora) +- Deprecate `DataFrame.apply_rows` ([#19218](https://github.com/NVIDIA/cudf/pull/19218)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Require `numba-cuda>=0.16.0` ([#19213](https://github.com/NVIDIA/cudf/pull/19213)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add a mode to co-process decompression and compression on host and device ([#19203](https://github.com/NVIDIA/cudf/pull/19203)) [@vuule](https://github.com/vuule) +- Return valid for all-nulls in reduce() with nunique include-nulls aggregation ([#19196](https://github.com/NVIDIA/cudf/pull/19196)) [@davidwendt](https://github.com/davidwendt) +- Refactor JNI error handling ([#19149](https://github.com/NVIDIA/cudf/pull/19149)) [@ttnghia](https://github.com/ttnghia) +- Add support for horizontal string concatenation `pl.concat_str` ([#19142](https://github.com/NVIDIA/cudf/pull/19142)) [@Matt711](https://github.com/Matt711) +- Add PDS-DS Query 1 ([#19131](https://github.com/NVIDIA/cudf/pull/19131)) [@Matt711](https://github.com/Matt711) +- Support `cudf-polars` `str.reverse` ([#19117](https://github.com/NVIDIA/cudf/pull/19117)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `cudf-polars` `str.pad_end` and `str.pad_start` ([#19116](https://github.com/NVIDIA/cudf/pull/19116)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `cudf-polars` `str.head` and `str.tail` ([#19115](https://github.com/NVIDIA/cudf/pull/19115)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `cudf-polars` `str.to_titlecase` ([#19114](https://github.com/NVIDIA/cudf/pull/19114)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add `cudf/io/codec.hpp` to expose compression/decompression APIs ([#19113](https://github.com/NVIDIA/cudf/pull/19113)) [@ttnghia](https://github.com/ttnghia) +- Support converting decimals to/from pylibcudf scalars ([#19106](https://github.com/NVIDIA/cudf/pull/19106)) [@Matt711](https://github.com/Matt711) +- Support resource-constrained sort-merge inner join operation through left table partitioning ([#19102](https://github.com/NVIDIA/cudf/pull/19102)) [@shrshi](https://github.com/shrshi) +- Filter Parquet row groups using row bounds ([#19082](https://github.com/NVIDIA/cudf/pull/19082)) [@mhaseeb123](https://github.com/mhaseeb123) +- Implement UDF Filters ([#19070](https://github.com/NVIDIA/cudf/pull/19070)) [@lamarrr](https://github.com/lamarrr) +- Move the remaining libcudf pieces to C++20 ([#19065](https://github.com/NVIDIA/cudf/pull/19065)) [@vuule](https://github.com/vuule) +- Allow using a stream per thread at runtime ([#19051](https://github.com/NVIDIA/cudf/pull/19051)) [@vyasr](https://github.com/vyasr) +- Remove stacktrace retrieval code ([#19048](https://github.com/NVIDIA/cudf/pull/19048)) [@ttnghia](https://github.com/ttnghia) +- Compile libcudf using C++20 Standard ([#19045](https://github.com/NVIDIA/cudf/pull/19045)) [@vuule](https://github.com/vuule) +- String Transform Examples: Added Branching, Public API Versions, and Sampling ([#19038](https://github.com/NVIDIA/cudf/pull/19038)) [@lamarrr](https://github.com/lamarrr) +- Refactor JNI error handling ([#18983](https://github.com/NVIDIA/cudf/pull/18983)) [@ttnghia](https://github.com/ttnghia) +- Add basic ``Sink`` support for streaming cudf-polars executor ([#18963](https://github.com/NVIDIA/cudf/pull/18963)) [@rjzamora](https://github.com/rjzamora) +- Fix debug-build Failure in JIT Tests ([#18939](https://github.com/NVIDIA/cudf/pull/18939)) [@lamarrr](https://github.com/lamarrr) +- Add from_arrow factory methods for Scalar and DataType ([#18938](https://github.com/NVIDIA/cudf/pull/18938)) [@Matt711](https://github.com/Matt711) +- Add pylibcudf.Column.from_arrow factory method ([#18937](https://github.com/NVIDIA/cudf/pull/18937)) [@Matt711](https://github.com/Matt711) +- Add pylibcudf.Table.from_arrow factory method ([#18936](https://github.com/NVIDIA/cudf/pull/18936)) [@Matt711](https://github.com/Matt711) +- Update nvCOMP adapter ([#18931](https://github.com/NVIDIA/cudf/pull/18931)) [@vuule](https://github.com/vuule) +- Create a pylibcudf Column from a iterable of python strings ([#18916](https://github.com/NVIDIA/cudf/pull/18916)) [@Matt711](https://github.com/Matt711) +- Add CLI argument to enable OOM protection in PDS-H ([#18914](https://github.com/NVIDIA/cudf/pull/18914)) [@pentschev](https://github.com/pentschev) +- Implement data page pruning using Parquet page index stats ([#18873](https://github.com/NVIDIA/cudf/pull/18873)) [@mhaseeb123](https://github.com/mhaseeb123) +- Null-handling for Transforms ([#18845](https://github.com/NVIDIA/cudf/pull/18845)) [@lamarrr](https://github.com/lamarrr) +- Implement row group pruning with dictionaries in experimental PQ reader ([#18836](https://github.com/NVIDIA/cudf/pull/18836)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add support for parquet scan + count operation ([#18463](https://github.com/NVIDIA/cudf/pull/18463)) [@Matt711](https://github.com/Matt711) +- Manage strings with NRT ([#18453](https://github.com/NVIDIA/cudf/pull/18453)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🛠️ Improvements -- Disable codecov comments ([#19472](https://github.com/rapidsai/cudf/pull/19472)) [@bdice](https://github.com/bdice) -- [FEA] Remove CUDA JIT-Compatibility Checks & CCCL WARs ([#19470](https://github.com/rapidsai/cudf/pull/19470)) [@lamarrr](https://github.com/lamarrr) -- Use libnvcomp conda package ([#19439](https://github.com/rapidsai/cudf/pull/19439)) [@bdice](https://github.com/bdice) -- JNI Set RMM_LOG_LEVEL and RMM_LOG_ACTIVE_LEVEL to allow setting log level at compile time ([#19435](https://github.com/rapidsai/cudf/pull/19435)) [@abellina](https://github.com/abellina) -- Use numba-cuda >=0.14.0,<0.15.0 ([#19425](https://github.com/rapidsai/cudf/pull/19425)) [@bdice](https://github.com/bdice) -- fix(docker): use versioned `-latest` tag for all `rapidsai` images ([#19412](https://github.com/rapidsai/cudf/pull/19412)) [@gforsyth](https://github.com/gforsyth) -- Add `bounds_policy` to `pylibcudf.lists.segmented_gather` ([#19411](https://github.com/rapidsai/cudf/pull/19411)) [@TomAugspurger](https://github.com/TomAugspurger) -- Require `nvidia-ml-py` in cudf-polars and adjust default `default_blocksize` ([#19410](https://github.com/rapidsai/cudf/pull/19410)) [@rjzamora](https://github.com/rjzamora) -- More pytest fixtures and avoid GPU params in cuDF classic tests ([#19404](https://github.com/rapidsai/cudf/pull/19404)) [@mroeschke](https://github.com/mroeschke) -- More pytest fixtures and avoid GPU params in cuDF classic tests ([#19402](https://github.com/rapidsai/cudf/pull/19402)) [@mroeschke](https://github.com/mroeschke) -- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19401](https://github.com/rapidsai/cudf/pull/19401)) [@mroeschke](https://github.com/mroeschke) -- Support range syntax and improve validation message when running PDS-H/PDS-DS ([#19399](https://github.com/rapidsai/cudf/pull/19399)) [@Matt711](https://github.com/Matt711) -- Drop cuda 11 usages ([#19386](https://github.com/rapidsai/cudf/pull/19386)) [@galipremsagar](https://github.com/galipremsagar) -- Remove CUDA 11 Workarounds ([#19385](https://github.com/rapidsai/cudf/pull/19385)) [@vuule](https://github.com/vuule) -- Further reduce runtime of cuDF classic IO tests ([#19382](https://github.com/rapidsai/cudf/pull/19382)) [@mroeschke](https://github.com/mroeschke) -- remove cuspatial references, avoid triggering tests on clang-format config changes ([#19380](https://github.com/rapidsai/cudf/pull/19380)) [@jameslamb](https://github.com/jameslamb) -- Add __repr__ to plc.aggregation.Aggregation ([#19379](https://github.com/rapidsai/cudf/pull/19379)) [@Matt711](https://github.com/Matt711) -- Raise on unsupported boolean functions in a groupby context ([#19378](https://github.com/rapidsai/cudf/pull/19378)) [@Matt711](https://github.com/Matt711) -- Configure cudf-polars options through environment variables ([#19369](https://github.com/rapidsai/cudf/pull/19369)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add primitive row dispatch support for semi/anti join and cudf::contains ([#19361](https://github.com/rapidsai/cudf/pull/19361)) [@tgujar](https://github.com/tgujar) -- Refactor hybrid scan reader tests to a separate executable ([#19359](https://github.com/rapidsai/cudf/pull/19359)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add pylibcudf.Column.as_struct_column for cudf_polars ([#19357](https://github.com/rapidsai/cudf/pull/19357)) [@mroeschke](https://github.com/mroeschke) -- Improve error message for `assert_column_eq` in pylibcudf tests ([#19356](https://github.com/rapidsai/cudf/pull/19356)) [@TomAugspurger](https://github.com/TomAugspurger) -- Update the minimum version pinning for polars to 1.28 ([#19352](https://github.com/rapidsai/cudf/pull/19352)) [@Matt711](https://github.com/Matt711) -- Add a `cudf::set_null_masks_safe` API to safely handle intra word aliasing in bulk null mask set ([#19349](https://github.com/rapidsai/cudf/pull/19349)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove profiling ranges on non-public sort-merge join functions ([#19347](https://github.com/rapidsai/cudf/pull/19347)) [@shrshi](https://github.com/shrshi) -- Clean up cudf._lib.strings_udf.pyx ([#19335](https://github.com/rapidsai/cudf/pull/19335)) [@mroeschke](https://github.com/mroeschke) -- Add support for `pandas-2.3.1` ([#19334](https://github.com/rapidsai/cudf/pull/19334)) [@galipremsagar](https://github.com/galipremsagar) -- Allow comparison binop to datetime.date ([#19333](https://github.com/rapidsai/cudf/pull/19333)) [@mroeschke](https://github.com/mroeschke) -- Re-enable std/var reductions for libcudf debug builds ([#19331](https://github.com/rapidsai/cudf/pull/19331)) [@davidwendt](https://github.com/davidwendt) -- Optimize object listing in pandas-tests diff CI ([#19328](https://github.com/rapidsai/cudf/pull/19328)) [@TomAugspurger](https://github.com/TomAugspurger) -- Allow setting `StreamingExecutor.target_partition_size` with an environment variable ([#19316](https://github.com/rapidsai/cudf/pull/19316)) [@TomAugspurger](https://github.com/TomAugspurger) -- Remove unnecessary compute for integer windows ([#19315](https://github.com/rapidsai/cudf/pull/19315)) [@wence-](https://github.com/wence-) -- Update cudf.pandas test skips for pandas==2.3.1 ([#19313](https://github.com/rapidsai/cudf/pull/19313)) [@TomAugspurger](https://github.com/TomAugspurger) -- Support Expr.str.json_decode in cudf_polars ([#19307](https://github.com/rapidsai/cudf/pull/19307)) [@mroeschke](https://github.com/mroeschke) -- Move the Parquet `reader_impl` class declaration out of the `parquet::detail::reader` ([#19305](https://github.com/rapidsai/cudf/pull/19305)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix null mask assignment in aggregators and cleanup with C++20 ([#19302](https://github.com/rapidsai/cudf/pull/19302)) [@PointKernel](https://github.com/PointKernel) -- [pre-commit.ci] pre-commit autoupdate ([#19301](https://github.com/rapidsai/cudf/pull/19301)) [@pre-commit-ci[bot]](https://github.com/pre-commit-ci[bot]) -- Deprecate cudf::round for float types ([#19298](https://github.com/rapidsai/cudf/pull/19298)) [@davidwendt](https://github.com/davidwendt) -- Fixed type annotation for 'state' in make_recursive ([#19294](https://github.com/rapidsai/cudf/pull/19294)) [@TomAugspurger](https://github.com/TomAugspurger) -- Support Expr.str.splitn/split_exact in cudf_polars ([#19290](https://github.com/rapidsai/cudf/pull/19290)) [@mroeschke](https://github.com/mroeschke) -- Improve high-multiplicity joins benchmark ([#19287](https://github.com/rapidsai/cudf/pull/19287)) [@shrshi](https://github.com/shrshi) -- Add data types axis to joins benchmarks ([#19281](https://github.com/rapidsai/cudf/pull/19281)) [@shrshi](https://github.com/shrshi) -- Support Expr.str.strip_prefix/suffix in cudf_polars ([#19278](https://github.com/rapidsai/cudf/pull/19278)) [@mroeschke](https://github.com/mroeschke) -- Support Expr.str.json_path_match/len_bytes/len_chars in cudf_polars ([#19277](https://github.com/rapidsai/cudf/pull/19277)) [@mroeschke](https://github.com/mroeschke) -- Introduce classes for collecting source statistics ([#19276](https://github.com/rapidsai/cudf/pull/19276)) [@rjzamora](https://github.com/rjzamora) -- Support Expr.str.find & Expr.str.join for non string data in cudf_polars ([#19275](https://github.com/rapidsai/cudf/pull/19275)) [@mroeschke](https://github.com/mroeschke) -- Move shuffle method defaulting to config options creation ([#19274](https://github.com/rapidsai/cudf/pull/19274)) [@wence-](https://github.com/wence-) -- Rename "cardinality_factor" configuration to "unique_fraction" ([#19273](https://github.com/rapidsai/cudf/pull/19273)) [@rjzamora](https://github.com/rjzamora) -- Serialize `ConfigOptions` in pdsh benchmark output ([#19272](https://github.com/rapidsai/cudf/pull/19272)) [@TomAugspurger](https://github.com/TomAugspurger) -- Support `Expr.str.extract/extract_groups` in cudf_polars ([#19271](https://github.com/rapidsai/cudf/pull/19271)) [@mroeschke](https://github.com/mroeschke) -- Fix includes for segmented-reduce source files ([#19266](https://github.com/rapidsai/cudf/pull/19266)) [@davidwendt](https://github.com/davidwendt) -- Change default cudf-polars executor to "streaming" ([#19263](https://github.com/rapidsai/cudf/pull/19263)) [@TomAugspurger](https://github.com/TomAugspurger) -- Update snapshot repo to central.soantype.com ([#19259](https://github.com/rapidsai/cudf/pull/19259)) [@pxLi](https://github.com/pxLi) -- Raise `NotImplementedError` for `LazyFrame.profile` with the streaming exeuctor ([#19257](https://github.com/rapidsai/cudf/pull/19257)) [@TomAugspurger](https://github.com/TomAugspurger) -- Move ast expression function definitions to .cpp files ([#19250](https://github.com/rapidsai/cudf/pull/19250)) [@davidwendt](https://github.com/davidwendt) -- Enable chunked reading of PQ sources with `>2B` rows ([#19245](https://github.com/rapidsai/cudf/pull/19245)) [@mhaseeb123](https://github.com/mhaseeb123) -- Support `str.count_matches` and `str.contains_any` expressions in cudf_polars ([#19235](https://github.com/rapidsai/cudf/pull/19235)) [@mroeschke](https://github.com/mroeschke) -- Remove cudautils.py ([#19233](https://github.com/rapidsai/cudf/pull/19233)) [@mroeschke](https://github.com/mroeschke) -- Use CUDA 12.9 in Conda, Devcontainers, Spark, GHA, etc. ([#19231](https://github.com/rapidsai/cudf/pull/19231)) [@jakirkham](https://github.com/jakirkham) -- Leverage new pylibcudf grouped_range_rolling_window for cuDF classic rolling(window: timedelta) ([#19230](https://github.com/rapidsai/cudf/pull/19230)) [@mroeschke](https://github.com/mroeschke) -- Add nvtx annotations for task-based shuffle ([#19229](https://github.com/rapidsai/cudf/pull/19229)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add annotations and docstrings to indexing_utils.py ([#19228](https://github.com/rapidsai/cudf/pull/19228)) [@mroeschke](https://github.com/mroeschke) -- Use cub radix sort directly for all fixed-width-types in cudf::sorted_order ([#19227](https://github.com/rapidsai/cudf/pull/19227)) [@davidwendt](https://github.com/davidwendt) -- Move get_mask_offset_word utility to null_mask.cuh ([#19226](https://github.com/rapidsai/cudf/pull/19226)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf-polars PolarsDtype typing issues ([#19225](https://github.com/rapidsai/cudf/pull/19225)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add test for deserializing cudf_polars class instances ([#19224](https://github.com/rapidsai/cudf/pull/19224)) [@TomAugspurger](https://github.com/TomAugspurger) -- Make pyarrow an optional dependency of pylibcudf ([#19223](https://github.com/rapidsai/cudf/pull/19223)) [@mroeschke](https://github.com/mroeschke) -- Remove NumPy usage in cudf_polars ([#19222](https://github.com/rapidsai/cudf/pull/19222)) [@mroeschke](https://github.com/mroeschke) -- Remove pyarrow from cudf_polars tests ([#19219](https://github.com/rapidsai/cudf/pull/19219)) [@mroeschke](https://github.com/mroeschke) -- Pin Polars to <1.32 ([#19217](https://github.com/rapidsai/cudf/pull/19217)) [@Matt711](https://github.com/Matt711) -- Remove nvidia and dask channels ([#19216](https://github.com/rapidsai/cudf/pull/19216)) [@vyasr](https://github.com/vyasr) -- Refactor Transform Utilities ([#19212](https://github.com/rapidsai/cudf/pull/19212)) [@lamarrr](https://github.com/lamarrr) -- Refactor `grid_1d` class ([#19211](https://github.com/rapidsai/cudf/pull/19211)) [@lamarrr](https://github.com/lamarrr) -- Use radix sort for all fixed-width-types in cudf::sort ([#19208](https://github.com/rapidsai/cudf/pull/19208)) [@davidwendt](https://github.com/davidwendt) -- Fix mypy notes / warnings in cudf ([#19206](https://github.com/rapidsai/cudf/pull/19206)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add `pandas-2.3.0` support ([#19202](https://github.com/rapidsai/cudf/pull/19202)) [@galipremsagar](https://github.com/galipremsagar) -- Avoid `pylibcudf.interop.to_arrow` in `DataFrame.to_polars` in cudf_polars ([#19198](https://github.com/rapidsai/cudf/pull/19198)) [@mroeschke](https://github.com/mroeschke) -- Fix cudf-polars label ([#19197](https://github.com/rapidsai/cudf/pull/19197)) [@vyasr](https://github.com/vyasr) -- Record scale factor in experimental PDS-H benchmark ([#19195](https://github.com/rapidsai/cudf/pull/19195)) [@rjzamora](https://github.com/rjzamora) -- Require dtype argument to cudf_polars `Column` container ([#19193](https://github.com/rapidsai/cudf/pull/19193)) [@mroeschke](https://github.com/mroeschke) -- Modify cuGraph, cudf_pandas third party test data to avoid cuGraph bug ([#19189](https://github.com/rapidsai/cudf/pull/19189)) [@mroeschke](https://github.com/mroeschke) -- Avoid ConfigOptions in IR nodes ([#19186](https://github.com/rapidsai/cudf/pull/19186)) [@TomAugspurger](https://github.com/TomAugspurger) -- Use numba-cuda >=0.14.0,<0.15.0 to get pynvjitlink by default. ([#19182](https://github.com/rapidsai/cudf/pull/19182)) [@bdice](https://github.com/bdice) -- Use cuda::std:: traits and utilities for AST operators ([#19179](https://github.com/rapidsai/cudf/pull/19179)) [@PointKernel](https://github.com/PointKernel) -- Reenable predicate pushdown in streaming cudf-polars ([#19178](https://github.com/rapidsai/cudf/pull/19178)) [@TomAugspurger](https://github.com/TomAugspurger) -- remove more references to cubinlinker and ptxcompiler ([#19177](https://github.com/rapidsai/cudf/pull/19177)) [@jameslamb](https://github.com/jameslamb) -- Update coverage reporting for cudf-polars ([#19175](https://github.com/rapidsai/cudf/pull/19175)) [@TomAugspurger](https://github.com/TomAugspurger) -- Implement rich_repr for expressions ([#19173](https://github.com/rapidsai/cudf/pull/19173)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add script to generate javadoc with JDK17 ([#19170](https://github.com/rapidsai/cudf/pull/19170)) [@YanxuanLiu](https://github.com/YanxuanLiu) -- Make pylibcudf default stream choice consistent with libcudf ([#19167](https://github.com/rapidsai/cudf/pull/19167)) [@vyasr](https://github.com/vyasr) -- Part 2/2: Refactor PQ reader preprocessing utilities for reuse in hybrid scan ([#19166](https://github.com/rapidsai/cudf/pull/19166)) [@mhaseeb123](https://github.com/mhaseeb123) -- Leverage new pylibcudf grouped_range_rolling_window for cuDF classic `rolling(window: int)` ([#19162](https://github.com/rapidsai/cudf/pull/19162)) [@mroeschke](https://github.com/mroeschke) -- Support setting `max_rows_per_partition` and report total time in pdsh benchmarks ([#19158](https://github.com/rapidsai/cudf/pull/19158)) [@Matt711](https://github.com/Matt711) -- Define more StringColumn methods for StringMethods accessor ([#19157](https://github.com/rapidsai/cudf/pull/19157)) [@mroeschke](https://github.com/mroeschke) -- Optimize parquet reader's stats based row group filtering ([#19156](https://github.com/rapidsai/cudf/pull/19156)) [@mhaseeb123](https://github.com/mhaseeb123) -- Support polars Datetime with timezone types in cudf_polars ([#19155](https://github.com/rapidsai/cudf/pull/19155)) [@mroeschke](https://github.com/mroeschke) -- Configurable blocksize mode for streaming executor in unit tests ([#19146](https://github.com/rapidsai/cudf/pull/19146)) [@TomAugspurger](https://github.com/TomAugspurger) -- Optimizations for tdigest generation. ([#19140](https://github.com/rapidsai/cudf/pull/19140)) [@nvdbaranec](https://github.com/nvdbaranec) -- Remove CUDA 11 from dependencies.yaml ([#19139](https://github.com/rapidsai/cudf/pull/19139)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Use radix sort for float/double types ([#19137](https://github.com/rapidsai/cudf/pull/19137)) [@davidwendt](https://github.com/davidwendt) -- Support radix sort for timestamp and duration types ([#19136](https://github.com/rapidsai/cudf/pull/19136)) [@davidwendt](https://github.com/davidwendt) -- Used TypeDict for CachingVisitor.state ([#19135](https://github.com/rapidsai/cudf/pull/19135)) [@TomAugspurger](https://github.com/TomAugspurger) -- Move Accessor implementation to their own directory ([#19134](https://github.com/rapidsai/cudf/pull/19134)) [@mroeschke](https://github.com/mroeschke) -- Add benchmarks for sorting float and timestamp ([#19133](https://github.com/rapidsai/cudf/pull/19133)) [@davidwendt](https://github.com/davidwendt) -- Enable using page mask in `decompress_page_data` in Parquet reader ([#19132](https://github.com/rapidsai/cudf/pull/19132)) [@mhaseeb123](https://github.com/mhaseeb123) -- refactor(shellcheck): fix all shellcheck warnings/errors ([#19129](https://github.com/rapidsai/cudf/pull/19129)) [@gforsyth](https://github.com/gforsyth) -- Remove pytest pin ([#19127](https://github.com/rapidsai/cudf/pull/19127)) [@vyasr](https://github.com/vyasr) -- Move pdsh utility functions/classes to a seperate module ([#19126](https://github.com/rapidsai/cudf/pull/19126)) [@Matt711](https://github.com/Matt711) -- Use pylibcudf.Column.from_cuda_array_interface in as_column ([#19123](https://github.com/rapidsai/cudf/pull/19123)) [@mroeschke](https://github.com/mroeschke) -- Add validate arg to polars pdsh benchmarks ([#19121](https://github.com/rapidsai/cudf/pull/19121)) [@Matt711](https://github.com/Matt711) -- Share Index.values with base implementaiton ([#19112](https://github.com/rapidsai/cudf/pull/19112)) [@mroeschke](https://github.com/mroeschke) -- Use __len__ instead of len(obj.some_attribute) ([#19111](https://github.com/rapidsai/cudf/pull/19111)) [@mroeschke](https://github.com/mroeschke) -- Consistently handle ascending/na_position conversions to pylibcudf ([#19110](https://github.com/rapidsai/cudf/pull/19110)) [@mroeschke](https://github.com/mroeschke) -- Raise EmptyDataError in pandas-compat mode for empty read_csv ([#19109](https://github.com/rapidsai/cudf/pull/19109)) [@mroeschke](https://github.com/mroeschke) -- Use cooperative-groups for warp-parallel kernels in nvtext ([#19107](https://github.com/rapidsai/cudf/pull/19107)) [@davidwendt](https://github.com/davidwendt) -- Quick fixes of `modernize-use-constraints` rule ([#19105](https://github.com/rapidsai/cudf/pull/19105)) [@vuule](https://github.com/vuule) -- Avoid O(n) lookup when creating cuDF Python mixins ([#19104](https://github.com/rapidsai/cudf/pull/19104)) [@mroeschke](https://github.com/mroeschke) -- Update cudf to accommodate breaking changes in cuCollections ([#19093](https://github.com/rapidsai/cudf/pull/19093)) [@PointKernel](https://github.com/PointKernel) -- Remove `hostdevice_vector::element` due to unnecessary synchronization ([#19092](https://github.com/rapidsai/cudf/pull/19092)) [@JigaoLuo](https://github.com/JigaoLuo) -- Support passing DataType to Column container in `cudf_polars` ([#19091](https://github.com/rapidsai/cudf/pull/19091)) [@mroeschke](https://github.com/mroeschke) -- Add strings zfill overload to accept widths column ([#19090](https://github.com/rapidsai/cudf/pull/19090)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-25.06 to branch-25.08 ([#19087](https://github.com/rapidsai/cudf/pull/19087)) [@Matt711](https://github.com/Matt711) -- Optimize tokenization for dask task graphs in cudf-polars ([#19083](https://github.com/rapidsai/cudf/pull/19083)) [@TomAugspurger](https://github.com/TomAugspurger) -- Multi-column null sanitization for struct columns ([#19080](https://github.com/rapidsai/cudf/pull/19080)) [@shrshi](https://github.com/shrshi) -- Support `polars.Expr.value_counts` in `cudf_polars` ([#19079](https://github.com/rapidsai/cudf/pull/19079)) [@mroeschke](https://github.com/mroeschke) -- Support `polars.struct` expression in `cudf_polars` ([#19075](https://github.com/rapidsai/cudf/pull/19075)) [@mroeschke](https://github.com/mroeschke) -- Improve pdsh query docs ([#19073](https://github.com/rapidsai/cudf/pull/19073)) [@Matt711](https://github.com/Matt711) -- Update mypy configuration to check against polars ([#19072](https://github.com/rapidsai/cudf/pull/19072)) [@TomAugspurger](https://github.com/TomAugspurger) -- [cudf-polars] Update rapidsmpf import paths ([#19068](https://github.com/rapidsai/cudf/pull/19068)) [@madsbk](https://github.com/madsbk) -- Fix clang-tidy `modernize-use-integer-sign-comparison` rule ([#19066](https://github.com/rapidsai/cudf/pull/19066)) [@vuule](https://github.com/vuule) -- [cudf-polars] Use RapidsMPF's config options ([#19059](https://github.com/rapidsai/cudf/pull/19059)) [@madsbk](https://github.com/madsbk) -- Unskip narwhals tests for cudf-polars run ([#19056](https://github.com/rapidsai/cudf/pull/19056)) [@Matt711](https://github.com/Matt711) -- Remove unnecessary synchronization (miss-sync) during Parquet reading (Part 1: device_scalar) ([#19055](https://github.com/rapidsai/cudf/pull/19055)) [@JigaoLuo](https://github.com/JigaoLuo) -- Part 1/2: Refactor PQ reader chunking utilities for reuse in hybrid scan ([#19054](https://github.com/rapidsai/cudf/pull/19054)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add support for StructFunction expressions in cudf_polars ([#19052](https://github.com/rapidsai/cudf/pull/19052)) [@mroeschke](https://github.com/mroeschke) -- Swap cuda::std::distance for thrust::distance ([#19050](https://github.com/rapidsai/cudf/pull/19050)) [@vyasr](https://github.com/vyasr) -- Rename `parquet_chunked_writer` to `chunked_parquet_writer` for consistency with the reader ([#19047](https://github.com/rapidsai/cudf/pull/19047)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add pylibcudf.Scalar.to_py to avoid scalar conversion to host via pyarrow ([#19043](https://github.com/rapidsai/cudf/pull/19043)) [@mroeschke](https://github.com/mroeschke) -- Fix and expand `to_parquet` tests of the `skip_compression` option ([#19042](https://github.com/rapidsai/cudf/pull/19042)) [@vuule](https://github.com/vuule) -- Remove CUDA 11 devcontainers and update CI scripts ([#19040](https://github.com/rapidsai/cudf/pull/19040)) [@bdice](https://github.com/bdice) -- refactor(rattler): remove cuda 11 branching ([#19039](https://github.com/rapidsai/cudf/pull/19039)) [@gforsyth](https://github.com/gforsyth) -- Use thrust::tabulate_output_iterator ([#19037](https://github.com/rapidsai/cudf/pull/19037)) [@bdice](https://github.com/bdice) -- Remove skip_rows workaround for chunked Parquet reader in cudf-polars ([#19036](https://github.com/rapidsai/cudf/pull/19036)) [@Matt711](https://github.com/Matt711) -- Prefer chaining pylibcudf IO options in cudf-polars ([#19022](https://github.com/rapidsai/cudf/pull/19022)) [@Matt711](https://github.com/Matt711) -- `batched_memset` to use a `host_span` arg instead of `std::vector` ([#19020](https://github.com/rapidsai/cudf/pull/19020)) [@mhaseeb123](https://github.com/mhaseeb123) -- Import from collections.abc for consistent typing/runing access ([#19019](https://github.com/rapidsai/cudf/pull/19019)) [@mroeschke](https://github.com/mroeschke) -- Avoid using cudf module for type annotations ([#19018](https://github.com/rapidsai/cudf/pull/19018)) [@mroeschke](https://github.com/mroeschke) -- Mark pandas unit test test_eval_no_support_column_name as xpassing ([#19016](https://github.com/rapidsai/cudf/pull/19016)) [@mroeschke](https://github.com/mroeschke) -- Improving Parquet decode throughput for struct type columns ([#19014](https://github.com/rapidsai/cudf/pull/19014)) [@shrshi](https://github.com/shrshi) -- Unify Frame._split and DataFrame.scatter_by_map/partition_by_hash implementations ([#19013](https://github.com/rapidsai/cudf/pull/19013)) [@mroeschke](https://github.com/mroeschke) -- Move IndexedFrame.memory_usage docstrings to DataFrame/Series, make RangeIndex methods consistent with base class ([#19010](https://github.com/rapidsai/cudf/pull/19010)) [@mroeschke](https://github.com/mroeschke) -- Share DataFrame/Series.(de)seralize methods, implement to_dlpack directly on Frame ([#19008](https://github.com/rapidsai/cudf/pull/19008)) [@mroeschke](https://github.com/mroeschke) -- Pin narhwals to 1.41 ([#19007](https://github.com/rapidsai/cudf/pull/19007)) [@Matt711](https://github.com/Matt711) -- Add year range check to cudf::strings::is_timestamp ([#19006](https://github.com/rapidsai/cudf/pull/19006)) [@davidwendt](https://github.com/davidwendt) -- Add cudf::strings::contains_multiple to pylibcudf ([#19003](https://github.com/rapidsai/cudf/pull/19003)) [@davidwendt](https://github.com/davidwendt) -- Avoid unnecessary partition step in streaming join ([#19002](https://github.com/rapidsai/cudf/pull/19002)) [@rjzamora](https://github.com/rjzamora) -- Part 2/n: Use cooperative groups in PQ decoders ([#18978](https://github.com/rapidsai/cudf/pull/18978)) [@mhaseeb123](https://github.com/mhaseeb123) -- Move libcudf copying benchmarks to nvbench ([#18976](https://github.com/rapidsai/cudf/pull/18976)) [@davidwendt](https://github.com/davidwendt) -- Add lag/lead/bitwise/row_number aggregations to pylibcudf ([#18975](https://github.com/rapidsai/cudf/pull/18975)) [@mroeschke](https://github.com/mroeschke) -- Switch to importing rather than cimporting datetime ([#18974](https://github.com/rapidsai/cudf/pull/18974)) [@vyasr](https://github.com/vyasr) -- stop uploading packages to downloads.rapids.ai ([#18973](https://github.com/rapidsai/cudf/pull/18973)) [@jameslamb](https://github.com/jameslamb) -- Trace `IR.do_evaluate` in cudf_polars ([#18970](https://github.com/rapidsai/cudf/pull/18970)) [@TomAugspurger](https://github.com/TomAugspurger) -- xfail more pandas unit tests that fail with cudf.pandas before execution instead of xfailing after execution ([#18965](https://github.com/rapidsai/cudf/pull/18965)) [@mroeschke](https://github.com/mroeschke) -- Remove test checks that depend on the compression engine ([#18960](https://github.com/rapidsai/cudf/pull/18960)) [@vuule](https://github.com/vuule) -- Use cooperative-groups for warp-parallel kernels in strings functions ([#18959](https://github.com/rapidsai/cudf/pull/18959)) [@davidwendt](https://github.com/davidwendt) -- fetch code before running pull request labeler ([#18958](https://github.com/rapidsai/cudf/pull/18958)) [@jameslamb](https://github.com/jameslamb) -- Use cooperative groups in parquet decoder kernels ([#18954](https://github.com/rapidsai/cudf/pull/18954)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add a DataType container in cudf_polars ([#18953](https://github.com/rapidsai/cudf/pull/18953)) [@mroeschke](https://github.com/mroeschke) -- add 'rapids-init-pip' to test_cudf_polars_polars_tests.sh ([#18951](https://github.com/rapidsai/cudf/pull/18951)) [@jameslamb](https://github.com/jameslamb) -- parameterized ucx / ucxx ([#18949](https://github.com/rapidsai/cudf/pull/18949)) [@quasiben](https://github.com/quasiben) -- Rework cudf::sorted_order implementation for faster compile ([#18948](https://github.com/rapidsai/cudf/pull/18948)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated Series methods, isclose ([#18947](https://github.com/rapidsai/cudf/pull/18947)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated groupby.collect ([#18946](https://github.com/rapidsai/cudf/pull/18946)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated get_dummies(cats=, ...) ([#18944](https://github.com/rapidsai/cudf/pull/18944)) [@mroeschke](https://github.com/mroeschke) -- Add .python_typecode and .typestr attributes to DataType ([#18941](https://github.com/rapidsai/cudf/pull/18941)) [@Matt711](https://github.com/Matt711) -- Remove deprecated APIs ([#18933](https://github.com/rapidsai/cudf/pull/18933)) [@vuule](https://github.com/vuule) -- Remove cudf.Scalar ([#18927](https://github.com/rapidsai/cudf/pull/18927)) [@mroeschke](https://github.com/mroeschke) -- Add #pragma once to prevent redundant includes and speed up compilation ([#18925](https://github.com/rapidsai/cudf/pull/18925)) [@PointKernel](https://github.com/PointKernel) -- Bump polars version to <1.31 ([#18920](https://github.com/rapidsai/cudf/pull/18920)) [@Matt711](https://github.com/Matt711) -- Apply primitive row operators into hash join ([#18896](https://github.com/rapidsai/cudf/pull/18896)) [@PointKernel](https://github.com/PointKernel) -- Branch 25.08 merge branch 25.06 ([#18895](https://github.com/rapidsai/cudf/pull/18895)) [@vyasr](https://github.com/vyasr) -- Remove deprecated `cudf::io::host_buffer` ([#18881](https://github.com/rapidsai/cudf/pull/18881)) [@Matt711](https://github.com/Matt711) -- Fix decompression scratch size in AUTO mode ([#18878](https://github.com/rapidsai/cudf/pull/18878)) [@vuule](https://github.com/vuule) -- Apply linter suggestions to cuIO code ([#18876](https://github.com/rapidsai/cudf/pull/18876)) [@vuule](https://github.com/vuule) -- xfail pandas unit tests that fail with cudf.pandas ([#18872](https://github.com/rapidsai/cudf/pull/18872)) [@mroeschke](https://github.com/mroeschke) -- Branch 25.08 merge branch 25.06 ([#18855](https://github.com/rapidsai/cudf/pull/18855)) [@vyasr](https://github.com/vyasr) -- Add support for extended dtypes in `cudf.pandas` ([#18832](https://github.com/rapidsai/cudf/pull/18832)) [@galipremsagar](https://github.com/galipremsagar) -- Auto merge fix for branch-25.08 ([#18824](https://github.com/rapidsai/cudf/pull/18824)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-25.06 to branch-25.08 ([#18817](https://github.com/rapidsai/cudf/pull/18817)) [@Matt711](https://github.com/Matt711) -- Forward-merge branch-25.06 to branch-25.08 ([#18756](https://github.com/rapidsai/cudf/pull/18756)) [@Matt711](https://github.com/Matt711) -- Fix auto merge conflict for branch-25.08 ([#18733](https://github.com/rapidsai/cudf/pull/18733)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-25.06 to branch-25.08 ([#18698](https://github.com/rapidsai/cudf/pull/18698)) [@Matt711](https://github.com/Matt711) -- Fix merge conflict for auto-merger 25.06 to 25.08 ([#18693](https://github.com/rapidsai/cudf/pull/18693)) [@davidwendt](https://github.com/davidwendt) -- Fix merge conflict: branch-25.06 into branch-25.08 ([#18668](https://github.com/rapidsai/cudf/pull/18668)) [@davidwendt](https://github.com/davidwendt) -- Make cuda12 as JNI default ([#18651](https://github.com/rapidsai/cudf/pull/18651)) [@pxLi](https://github.com/pxLi) -- Forward-merge branch-25.06 into branch-25.08 ([#18647](https://github.com/rapidsai/cudf/pull/18647)) [@bdice](https://github.com/bdice) -- Fix merge branch-25.06 into branch-25.08 ([#18622](https://github.com/rapidsai/cudf/pull/18622)) [@davidwendt](https://github.com/davidwendt) -- Store polars Series instead of pyarrow Array in cudf_polars LiteralColumn expr ([#18564](https://github.com/rapidsai/cudf/pull/18564)) [@mroeschke](https://github.com/mroeschke) -- Refactor strings split/record with whitespace logic ([#18560](https://github.com/rapidsai/cudf/pull/18560)) [@davidwendt](https://github.com/davidwendt) -- Refactor hash join with multiset ([#18021](https://github.com/rapidsai/cudf/pull/18021)) [@PointKernel](https://github.com/PointKernel) +- Disable codecov comments ([#19472](https://github.com/NVIDIA/cudf/pull/19472)) [@bdice](https://github.com/bdice) +- [FEA] Remove CUDA JIT-Compatibility Checks & CCCL WARs ([#19470](https://github.com/NVIDIA/cudf/pull/19470)) [@lamarrr](https://github.com/lamarrr) +- Use libnvcomp conda package ([#19439](https://github.com/NVIDIA/cudf/pull/19439)) [@bdice](https://github.com/bdice) +- JNI Set RMM_LOG_LEVEL and RMM_LOG_ACTIVE_LEVEL to allow setting log level at compile time ([#19435](https://github.com/NVIDIA/cudf/pull/19435)) [@abellina](https://github.com/abellina) +- Use numba-cuda >=0.14.0,<0.15.0 ([#19425](https://github.com/NVIDIA/cudf/pull/19425)) [@bdice](https://github.com/bdice) +- fix(docker): use versioned `-latest` tag for all `rapidsai` images ([#19412](https://github.com/NVIDIA/cudf/pull/19412)) [@gforsyth](https://github.com/gforsyth) +- Add `bounds_policy` to `pylibcudf.lists.segmented_gather` ([#19411](https://github.com/NVIDIA/cudf/pull/19411)) [@TomAugspurger](https://github.com/TomAugspurger) +- Require `nvidia-ml-py` in cudf-polars and adjust default `default_blocksize` ([#19410](https://github.com/NVIDIA/cudf/pull/19410)) [@rjzamora](https://github.com/rjzamora) +- More pytest fixtures and avoid GPU params in cuDF classic tests ([#19404](https://github.com/NVIDIA/cudf/pull/19404)) [@mroeschke](https://github.com/mroeschke) +- More pytest fixtures and avoid GPU params in cuDF classic tests ([#19402](https://github.com/NVIDIA/cudf/pull/19402)) [@mroeschke](https://github.com/mroeschke) +- Use more pytest fixtures and avoid GPU parameterization in cuDF classic tests ([#19401](https://github.com/NVIDIA/cudf/pull/19401)) [@mroeschke](https://github.com/mroeschke) +- Support range syntax and improve validation message when running PDS-H/PDS-DS ([#19399](https://github.com/NVIDIA/cudf/pull/19399)) [@Matt711](https://github.com/Matt711) +- Drop cuda 11 usages ([#19386](https://github.com/NVIDIA/cudf/pull/19386)) [@galipremsagar](https://github.com/galipremsagar) +- Remove CUDA 11 Workarounds ([#19385](https://github.com/NVIDIA/cudf/pull/19385)) [@vuule](https://github.com/vuule) +- Further reduce runtime of cuDF classic IO tests ([#19382](https://github.com/NVIDIA/cudf/pull/19382)) [@mroeschke](https://github.com/mroeschke) +- remove cuspatial references, avoid triggering tests on clang-format config changes ([#19380](https://github.com/NVIDIA/cudf/pull/19380)) [@jameslamb](https://github.com/jameslamb) +- Add __repr__ to plc.aggregation.Aggregation ([#19379](https://github.com/NVIDIA/cudf/pull/19379)) [@Matt711](https://github.com/Matt711) +- Raise on unsupported boolean functions in a groupby context ([#19378](https://github.com/NVIDIA/cudf/pull/19378)) [@Matt711](https://github.com/Matt711) +- Configure cudf-polars options through environment variables ([#19369](https://github.com/NVIDIA/cudf/pull/19369)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add primitive row dispatch support for semi/anti join and cudf::contains ([#19361](https://github.com/NVIDIA/cudf/pull/19361)) [@tgujar](https://github.com/tgujar) +- Refactor hybrid scan reader tests to a separate executable ([#19359](https://github.com/NVIDIA/cudf/pull/19359)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add pylibcudf.Column.as_struct_column for cudf_polars ([#19357](https://github.com/NVIDIA/cudf/pull/19357)) [@mroeschke](https://github.com/mroeschke) +- Improve error message for `assert_column_eq` in pylibcudf tests ([#19356](https://github.com/NVIDIA/cudf/pull/19356)) [@TomAugspurger](https://github.com/TomAugspurger) +- Update the minimum version pinning for polars to 1.28 ([#19352](https://github.com/NVIDIA/cudf/pull/19352)) [@Matt711](https://github.com/Matt711) +- Add a `cudf::set_null_masks_safe` API to safely handle intra word aliasing in bulk null mask set ([#19349](https://github.com/NVIDIA/cudf/pull/19349)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove profiling ranges on non-public sort-merge join functions ([#19347](https://github.com/NVIDIA/cudf/pull/19347)) [@shrshi](https://github.com/shrshi) +- Clean up cudf._lib.strings_udf.pyx ([#19335](https://github.com/NVIDIA/cudf/pull/19335)) [@mroeschke](https://github.com/mroeschke) +- Add support for `pandas-2.3.1` ([#19334](https://github.com/NVIDIA/cudf/pull/19334)) [@galipremsagar](https://github.com/galipremsagar) +- Allow comparison binop to datetime.date ([#19333](https://github.com/NVIDIA/cudf/pull/19333)) [@mroeschke](https://github.com/mroeschke) +- Re-enable std/var reductions for libcudf debug builds ([#19331](https://github.com/NVIDIA/cudf/pull/19331)) [@davidwendt](https://github.com/davidwendt) +- Optimize object listing in pandas-tests diff CI ([#19328](https://github.com/NVIDIA/cudf/pull/19328)) [@TomAugspurger](https://github.com/TomAugspurger) +- Allow setting `StreamingExecutor.target_partition_size` with an environment variable ([#19316](https://github.com/NVIDIA/cudf/pull/19316)) [@TomAugspurger](https://github.com/TomAugspurger) +- Remove unnecessary compute for integer windows ([#19315](https://github.com/NVIDIA/cudf/pull/19315)) [@wence-](https://github.com/wence-) +- Update cudf.pandas test skips for pandas==2.3.1 ([#19313](https://github.com/NVIDIA/cudf/pull/19313)) [@TomAugspurger](https://github.com/TomAugspurger) +- Support Expr.str.json_decode in cudf_polars ([#19307](https://github.com/NVIDIA/cudf/pull/19307)) [@mroeschke](https://github.com/mroeschke) +- Move the Parquet `reader_impl` class declaration out of the `parquet::detail::reader` ([#19305](https://github.com/NVIDIA/cudf/pull/19305)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix null mask assignment in aggregators and cleanup with C++20 ([#19302](https://github.com/NVIDIA/cudf/pull/19302)) [@PointKernel](https://github.com/PointKernel) +- [pre-commit.ci] pre-commit autoupdate ([#19301](https://github.com/NVIDIA/cudf/pull/19301)) [@pre-commit-ci[bot]](https://github.com/pre-commit-ci[bot]) +- Deprecate cudf::round for float types ([#19298](https://github.com/NVIDIA/cudf/pull/19298)) [@davidwendt](https://github.com/davidwendt) +- Fixed type annotation for 'state' in make_recursive ([#19294](https://github.com/NVIDIA/cudf/pull/19294)) [@TomAugspurger](https://github.com/TomAugspurger) +- Support Expr.str.splitn/split_exact in cudf_polars ([#19290](https://github.com/NVIDIA/cudf/pull/19290)) [@mroeschke](https://github.com/mroeschke) +- Improve high-multiplicity joins benchmark ([#19287](https://github.com/NVIDIA/cudf/pull/19287)) [@shrshi](https://github.com/shrshi) +- Add data types axis to joins benchmarks ([#19281](https://github.com/NVIDIA/cudf/pull/19281)) [@shrshi](https://github.com/shrshi) +- Support Expr.str.strip_prefix/suffix in cudf_polars ([#19278](https://github.com/NVIDIA/cudf/pull/19278)) [@mroeschke](https://github.com/mroeschke) +- Support Expr.str.json_path_match/len_bytes/len_chars in cudf_polars ([#19277](https://github.com/NVIDIA/cudf/pull/19277)) [@mroeschke](https://github.com/mroeschke) +- Introduce classes for collecting source statistics ([#19276](https://github.com/NVIDIA/cudf/pull/19276)) [@rjzamora](https://github.com/rjzamora) +- Support Expr.str.find & Expr.str.join for non string data in cudf_polars ([#19275](https://github.com/NVIDIA/cudf/pull/19275)) [@mroeschke](https://github.com/mroeschke) +- Move shuffle method defaulting to config options creation ([#19274](https://github.com/NVIDIA/cudf/pull/19274)) [@wence-](https://github.com/wence-) +- Rename "cardinality_factor" configuration to "unique_fraction" ([#19273](https://github.com/NVIDIA/cudf/pull/19273)) [@rjzamora](https://github.com/rjzamora) +- Serialize `ConfigOptions` in pdsh benchmark output ([#19272](https://github.com/NVIDIA/cudf/pull/19272)) [@TomAugspurger](https://github.com/TomAugspurger) +- Support `Expr.str.extract/extract_groups` in cudf_polars ([#19271](https://github.com/NVIDIA/cudf/pull/19271)) [@mroeschke](https://github.com/mroeschke) +- Fix includes for segmented-reduce source files ([#19266](https://github.com/NVIDIA/cudf/pull/19266)) [@davidwendt](https://github.com/davidwendt) +- Change default cudf-polars executor to "streaming" ([#19263](https://github.com/NVIDIA/cudf/pull/19263)) [@TomAugspurger](https://github.com/TomAugspurger) +- Update snapshot repo to central.soantype.com ([#19259](https://github.com/NVIDIA/cudf/pull/19259)) [@pxLi](https://github.com/pxLi) +- Raise `NotImplementedError` for `LazyFrame.profile` with the streaming exeuctor ([#19257](https://github.com/NVIDIA/cudf/pull/19257)) [@TomAugspurger](https://github.com/TomAugspurger) +- Move ast expression function definitions to .cpp files ([#19250](https://github.com/NVIDIA/cudf/pull/19250)) [@davidwendt](https://github.com/davidwendt) +- Enable chunked reading of PQ sources with `>2B` rows ([#19245](https://github.com/NVIDIA/cudf/pull/19245)) [@mhaseeb123](https://github.com/mhaseeb123) +- Support `str.count_matches` and `str.contains_any` expressions in cudf_polars ([#19235](https://github.com/NVIDIA/cudf/pull/19235)) [@mroeschke](https://github.com/mroeschke) +- Remove cudautils.py ([#19233](https://github.com/NVIDIA/cudf/pull/19233)) [@mroeschke](https://github.com/mroeschke) +- Use CUDA 12.9 in Conda, Devcontainers, Spark, GHA, etc. ([#19231](https://github.com/NVIDIA/cudf/pull/19231)) [@jakirkham](https://github.com/jakirkham) +- Leverage new pylibcudf grouped_range_rolling_window for cuDF classic rolling(window: timedelta) ([#19230](https://github.com/NVIDIA/cudf/pull/19230)) [@mroeschke](https://github.com/mroeschke) +- Add nvtx annotations for task-based shuffle ([#19229](https://github.com/NVIDIA/cudf/pull/19229)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add annotations and docstrings to indexing_utils.py ([#19228](https://github.com/NVIDIA/cudf/pull/19228)) [@mroeschke](https://github.com/mroeschke) +- Use cub radix sort directly for all fixed-width-types in cudf::sorted_order ([#19227](https://github.com/NVIDIA/cudf/pull/19227)) [@davidwendt](https://github.com/davidwendt) +- Move get_mask_offset_word utility to null_mask.cuh ([#19226](https://github.com/NVIDIA/cudf/pull/19226)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf-polars PolarsDtype typing issues ([#19225](https://github.com/NVIDIA/cudf/pull/19225)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add test for deserializing cudf_polars class instances ([#19224](https://github.com/NVIDIA/cudf/pull/19224)) [@TomAugspurger](https://github.com/TomAugspurger) +- Make pyarrow an optional dependency of pylibcudf ([#19223](https://github.com/NVIDIA/cudf/pull/19223)) [@mroeschke](https://github.com/mroeschke) +- Remove NumPy usage in cudf_polars ([#19222](https://github.com/NVIDIA/cudf/pull/19222)) [@mroeschke](https://github.com/mroeschke) +- Remove pyarrow from cudf_polars tests ([#19219](https://github.com/NVIDIA/cudf/pull/19219)) [@mroeschke](https://github.com/mroeschke) +- Pin Polars to <1.32 ([#19217](https://github.com/NVIDIA/cudf/pull/19217)) [@Matt711](https://github.com/Matt711) +- Remove nvidia and dask channels ([#19216](https://github.com/NVIDIA/cudf/pull/19216)) [@vyasr](https://github.com/vyasr) +- Refactor Transform Utilities ([#19212](https://github.com/NVIDIA/cudf/pull/19212)) [@lamarrr](https://github.com/lamarrr) +- Refactor `grid_1d` class ([#19211](https://github.com/NVIDIA/cudf/pull/19211)) [@lamarrr](https://github.com/lamarrr) +- Use radix sort for all fixed-width-types in cudf::sort ([#19208](https://github.com/NVIDIA/cudf/pull/19208)) [@davidwendt](https://github.com/davidwendt) +- Fix mypy notes / warnings in cudf ([#19206](https://github.com/NVIDIA/cudf/pull/19206)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add `pandas-2.3.0` support ([#19202](https://github.com/NVIDIA/cudf/pull/19202)) [@galipremsagar](https://github.com/galipremsagar) +- Avoid `pylibcudf.interop.to_arrow` in `DataFrame.to_polars` in cudf_polars ([#19198](https://github.com/NVIDIA/cudf/pull/19198)) [@mroeschke](https://github.com/mroeschke) +- Fix cudf-polars label ([#19197](https://github.com/NVIDIA/cudf/pull/19197)) [@vyasr](https://github.com/vyasr) +- Record scale factor in experimental PDS-H benchmark ([#19195](https://github.com/NVIDIA/cudf/pull/19195)) [@rjzamora](https://github.com/rjzamora) +- Require dtype argument to cudf_polars `Column` container ([#19193](https://github.com/NVIDIA/cudf/pull/19193)) [@mroeschke](https://github.com/mroeschke) +- Modify cuGraph, cudf_pandas third party test data to avoid cuGraph bug ([#19189](https://github.com/NVIDIA/cudf/pull/19189)) [@mroeschke](https://github.com/mroeschke) +- Avoid ConfigOptions in IR nodes ([#19186](https://github.com/NVIDIA/cudf/pull/19186)) [@TomAugspurger](https://github.com/TomAugspurger) +- Use numba-cuda >=0.14.0,<0.15.0 to get pynvjitlink by default. ([#19182](https://github.com/NVIDIA/cudf/pull/19182)) [@bdice](https://github.com/bdice) +- Use cuda::std:: traits and utilities for AST operators ([#19179](https://github.com/NVIDIA/cudf/pull/19179)) [@PointKernel](https://github.com/PointKernel) +- Reenable predicate pushdown in streaming cudf-polars ([#19178](https://github.com/NVIDIA/cudf/pull/19178)) [@TomAugspurger](https://github.com/TomAugspurger) +- remove more references to cubinlinker and ptxcompiler ([#19177](https://github.com/NVIDIA/cudf/pull/19177)) [@jameslamb](https://github.com/jameslamb) +- Update coverage reporting for cudf-polars ([#19175](https://github.com/NVIDIA/cudf/pull/19175)) [@TomAugspurger](https://github.com/TomAugspurger) +- Implement rich_repr for expressions ([#19173](https://github.com/NVIDIA/cudf/pull/19173)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add script to generate javadoc with JDK17 ([#19170](https://github.com/NVIDIA/cudf/pull/19170)) [@YanxuanLiu](https://github.com/YanxuanLiu) +- Make pylibcudf default stream choice consistent with libcudf ([#19167](https://github.com/NVIDIA/cudf/pull/19167)) [@vyasr](https://github.com/vyasr) +- Part 2/2: Refactor PQ reader preprocessing utilities for reuse in hybrid scan ([#19166](https://github.com/NVIDIA/cudf/pull/19166)) [@mhaseeb123](https://github.com/mhaseeb123) +- Leverage new pylibcudf grouped_range_rolling_window for cuDF classic `rolling(window: int)` ([#19162](https://github.com/NVIDIA/cudf/pull/19162)) [@mroeschke](https://github.com/mroeschke) +- Support setting `max_rows_per_partition` and report total time in pdsh benchmarks ([#19158](https://github.com/NVIDIA/cudf/pull/19158)) [@Matt711](https://github.com/Matt711) +- Define more StringColumn methods for StringMethods accessor ([#19157](https://github.com/NVIDIA/cudf/pull/19157)) [@mroeschke](https://github.com/mroeschke) +- Optimize parquet reader's stats based row group filtering ([#19156](https://github.com/NVIDIA/cudf/pull/19156)) [@mhaseeb123](https://github.com/mhaseeb123) +- Support polars Datetime with timezone types in cudf_polars ([#19155](https://github.com/NVIDIA/cudf/pull/19155)) [@mroeschke](https://github.com/mroeschke) +- Configurable blocksize mode for streaming executor in unit tests ([#19146](https://github.com/NVIDIA/cudf/pull/19146)) [@TomAugspurger](https://github.com/TomAugspurger) +- Optimizations for tdigest generation. ([#19140](https://github.com/NVIDIA/cudf/pull/19140)) [@nvdbaranec](https://github.com/nvdbaranec) +- Remove CUDA 11 from dependencies.yaml ([#19139](https://github.com/NVIDIA/cudf/pull/19139)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Use radix sort for float/double types ([#19137](https://github.com/NVIDIA/cudf/pull/19137)) [@davidwendt](https://github.com/davidwendt) +- Support radix sort for timestamp and duration types ([#19136](https://github.com/NVIDIA/cudf/pull/19136)) [@davidwendt](https://github.com/davidwendt) +- Used TypeDict for CachingVisitor.state ([#19135](https://github.com/NVIDIA/cudf/pull/19135)) [@TomAugspurger](https://github.com/TomAugspurger) +- Move Accessor implementation to their own directory ([#19134](https://github.com/NVIDIA/cudf/pull/19134)) [@mroeschke](https://github.com/mroeschke) +- Add benchmarks for sorting float and timestamp ([#19133](https://github.com/NVIDIA/cudf/pull/19133)) [@davidwendt](https://github.com/davidwendt) +- Enable using page mask in `decompress_page_data` in Parquet reader ([#19132](https://github.com/NVIDIA/cudf/pull/19132)) [@mhaseeb123](https://github.com/mhaseeb123) +- refactor(shellcheck): fix all shellcheck warnings/errors ([#19129](https://github.com/NVIDIA/cudf/pull/19129)) [@gforsyth](https://github.com/gforsyth) +- Remove pytest pin ([#19127](https://github.com/NVIDIA/cudf/pull/19127)) [@vyasr](https://github.com/vyasr) +- Move pdsh utility functions/classes to a seperate module ([#19126](https://github.com/NVIDIA/cudf/pull/19126)) [@Matt711](https://github.com/Matt711) +- Use pylibcudf.Column.from_cuda_array_interface in as_column ([#19123](https://github.com/NVIDIA/cudf/pull/19123)) [@mroeschke](https://github.com/mroeschke) +- Add validate arg to polars pdsh benchmarks ([#19121](https://github.com/NVIDIA/cudf/pull/19121)) [@Matt711](https://github.com/Matt711) +- Share Index.values with base implementaiton ([#19112](https://github.com/NVIDIA/cudf/pull/19112)) [@mroeschke](https://github.com/mroeschke) +- Use __len__ instead of len(obj.some_attribute) ([#19111](https://github.com/NVIDIA/cudf/pull/19111)) [@mroeschke](https://github.com/mroeschke) +- Consistently handle ascending/na_position conversions to pylibcudf ([#19110](https://github.com/NVIDIA/cudf/pull/19110)) [@mroeschke](https://github.com/mroeschke) +- Raise EmptyDataError in pandas-compat mode for empty read_csv ([#19109](https://github.com/NVIDIA/cudf/pull/19109)) [@mroeschke](https://github.com/mroeschke) +- Use cooperative-groups for warp-parallel kernels in nvtext ([#19107](https://github.com/NVIDIA/cudf/pull/19107)) [@davidwendt](https://github.com/davidwendt) +- Quick fixes of `modernize-use-constraints` rule ([#19105](https://github.com/NVIDIA/cudf/pull/19105)) [@vuule](https://github.com/vuule) +- Avoid O(n) lookup when creating cuDF Python mixins ([#19104](https://github.com/NVIDIA/cudf/pull/19104)) [@mroeschke](https://github.com/mroeschke) +- Update cudf to accommodate breaking changes in cuCollections ([#19093](https://github.com/NVIDIA/cudf/pull/19093)) [@PointKernel](https://github.com/PointKernel) +- Remove `hostdevice_vector::element` due to unnecessary synchronization ([#19092](https://github.com/NVIDIA/cudf/pull/19092)) [@JigaoLuo](https://github.com/JigaoLuo) +- Support passing DataType to Column container in `cudf_polars` ([#19091](https://github.com/NVIDIA/cudf/pull/19091)) [@mroeschke](https://github.com/mroeschke) +- Add strings zfill overload to accept widths column ([#19090](https://github.com/NVIDIA/cudf/pull/19090)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-25.06 to branch-25.08 ([#19087](https://github.com/NVIDIA/cudf/pull/19087)) [@Matt711](https://github.com/Matt711) +- Optimize tokenization for dask task graphs in cudf-polars ([#19083](https://github.com/NVIDIA/cudf/pull/19083)) [@TomAugspurger](https://github.com/TomAugspurger) +- Multi-column null sanitization for struct columns ([#19080](https://github.com/NVIDIA/cudf/pull/19080)) [@shrshi](https://github.com/shrshi) +- Support `polars.Expr.value_counts` in `cudf_polars` ([#19079](https://github.com/NVIDIA/cudf/pull/19079)) [@mroeschke](https://github.com/mroeschke) +- Support `polars.struct` expression in `cudf_polars` ([#19075](https://github.com/NVIDIA/cudf/pull/19075)) [@mroeschke](https://github.com/mroeschke) +- Improve pdsh query docs ([#19073](https://github.com/NVIDIA/cudf/pull/19073)) [@Matt711](https://github.com/Matt711) +- Update mypy configuration to check against polars ([#19072](https://github.com/NVIDIA/cudf/pull/19072)) [@TomAugspurger](https://github.com/TomAugspurger) +- [cudf-polars] Update rapidsmpf import paths ([#19068](https://github.com/NVIDIA/cudf/pull/19068)) [@madsbk](https://github.com/madsbk) +- Fix clang-tidy `modernize-use-integer-sign-comparison` rule ([#19066](https://github.com/NVIDIA/cudf/pull/19066)) [@vuule](https://github.com/vuule) +- [cudf-polars] Use RapidsMPF's config options ([#19059](https://github.com/NVIDIA/cudf/pull/19059)) [@madsbk](https://github.com/madsbk) +- Unskip narwhals tests for cudf-polars run ([#19056](https://github.com/NVIDIA/cudf/pull/19056)) [@Matt711](https://github.com/Matt711) +- Remove unnecessary synchronization (miss-sync) during Parquet reading (Part 1: device_scalar) ([#19055](https://github.com/NVIDIA/cudf/pull/19055)) [@JigaoLuo](https://github.com/JigaoLuo) +- Part 1/2: Refactor PQ reader chunking utilities for reuse in hybrid scan ([#19054](https://github.com/NVIDIA/cudf/pull/19054)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add support for StructFunction expressions in cudf_polars ([#19052](https://github.com/NVIDIA/cudf/pull/19052)) [@mroeschke](https://github.com/mroeschke) +- Swap cuda::std::distance for thrust::distance ([#19050](https://github.com/NVIDIA/cudf/pull/19050)) [@vyasr](https://github.com/vyasr) +- Rename `parquet_chunked_writer` to `chunked_parquet_writer` for consistency with the reader ([#19047](https://github.com/NVIDIA/cudf/pull/19047)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add pylibcudf.Scalar.to_py to avoid scalar conversion to host via pyarrow ([#19043](https://github.com/NVIDIA/cudf/pull/19043)) [@mroeschke](https://github.com/mroeschke) +- Fix and expand `to_parquet` tests of the `skip_compression` option ([#19042](https://github.com/NVIDIA/cudf/pull/19042)) [@vuule](https://github.com/vuule) +- Remove CUDA 11 devcontainers and update CI scripts ([#19040](https://github.com/NVIDIA/cudf/pull/19040)) [@bdice](https://github.com/bdice) +- refactor(rattler): remove cuda 11 branching ([#19039](https://github.com/NVIDIA/cudf/pull/19039)) [@gforsyth](https://github.com/gforsyth) +- Use thrust::tabulate_output_iterator ([#19037](https://github.com/NVIDIA/cudf/pull/19037)) [@bdice](https://github.com/bdice) +- Remove skip_rows workaround for chunked Parquet reader in cudf-polars ([#19036](https://github.com/NVIDIA/cudf/pull/19036)) [@Matt711](https://github.com/Matt711) +- Prefer chaining pylibcudf IO options in cudf-polars ([#19022](https://github.com/NVIDIA/cudf/pull/19022)) [@Matt711](https://github.com/Matt711) +- `batched_memset` to use a `host_span` arg instead of `std::vector` ([#19020](https://github.com/NVIDIA/cudf/pull/19020)) [@mhaseeb123](https://github.com/mhaseeb123) +- Import from collections.abc for consistent typing/runing access ([#19019](https://github.com/NVIDIA/cudf/pull/19019)) [@mroeschke](https://github.com/mroeschke) +- Avoid using cudf module for type annotations ([#19018](https://github.com/NVIDIA/cudf/pull/19018)) [@mroeschke](https://github.com/mroeschke) +- Mark pandas unit test test_eval_no_support_column_name as xpassing ([#19016](https://github.com/NVIDIA/cudf/pull/19016)) [@mroeschke](https://github.com/mroeschke) +- Improving Parquet decode throughput for struct type columns ([#19014](https://github.com/NVIDIA/cudf/pull/19014)) [@shrshi](https://github.com/shrshi) +- Unify Frame._split and DataFrame.scatter_by_map/partition_by_hash implementations ([#19013](https://github.com/NVIDIA/cudf/pull/19013)) [@mroeschke](https://github.com/mroeschke) +- Move IndexedFrame.memory_usage docstrings to DataFrame/Series, make RangeIndex methods consistent with base class ([#19010](https://github.com/NVIDIA/cudf/pull/19010)) [@mroeschke](https://github.com/mroeschke) +- Share DataFrame/Series.(de)seralize methods, implement to_dlpack directly on Frame ([#19008](https://github.com/NVIDIA/cudf/pull/19008)) [@mroeschke](https://github.com/mroeschke) +- Pin narhwals to 1.41 ([#19007](https://github.com/NVIDIA/cudf/pull/19007)) [@Matt711](https://github.com/Matt711) +- Add year range check to cudf::strings::is_timestamp ([#19006](https://github.com/NVIDIA/cudf/pull/19006)) [@davidwendt](https://github.com/davidwendt) +- Add cudf::strings::contains_multiple to pylibcudf ([#19003](https://github.com/NVIDIA/cudf/pull/19003)) [@davidwendt](https://github.com/davidwendt) +- Avoid unnecessary partition step in streaming join ([#19002](https://github.com/NVIDIA/cudf/pull/19002)) [@rjzamora](https://github.com/rjzamora) +- Part 2/n: Use cooperative groups in PQ decoders ([#18978](https://github.com/NVIDIA/cudf/pull/18978)) [@mhaseeb123](https://github.com/mhaseeb123) +- Move libcudf copying benchmarks to nvbench ([#18976](https://github.com/NVIDIA/cudf/pull/18976)) [@davidwendt](https://github.com/davidwendt) +- Add lag/lead/bitwise/row_number aggregations to pylibcudf ([#18975](https://github.com/NVIDIA/cudf/pull/18975)) [@mroeschke](https://github.com/mroeschke) +- Switch to importing rather than cimporting datetime ([#18974](https://github.com/NVIDIA/cudf/pull/18974)) [@vyasr](https://github.com/vyasr) +- stop uploading packages to downloads.rapids.ai ([#18973](https://github.com/NVIDIA/cudf/pull/18973)) [@jameslamb](https://github.com/jameslamb) +- Trace `IR.do_evaluate` in cudf_polars ([#18970](https://github.com/NVIDIA/cudf/pull/18970)) [@TomAugspurger](https://github.com/TomAugspurger) +- xfail more pandas unit tests that fail with cudf.pandas before execution instead of xfailing after execution ([#18965](https://github.com/NVIDIA/cudf/pull/18965)) [@mroeschke](https://github.com/mroeschke) +- Remove test checks that depend on the compression engine ([#18960](https://github.com/NVIDIA/cudf/pull/18960)) [@vuule](https://github.com/vuule) +- Use cooperative-groups for warp-parallel kernels in strings functions ([#18959](https://github.com/NVIDIA/cudf/pull/18959)) [@davidwendt](https://github.com/davidwendt) +- fetch code before running pull request labeler ([#18958](https://github.com/NVIDIA/cudf/pull/18958)) [@jameslamb](https://github.com/jameslamb) +- Use cooperative groups in parquet decoder kernels ([#18954](https://github.com/NVIDIA/cudf/pull/18954)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add a DataType container in cudf_polars ([#18953](https://github.com/NVIDIA/cudf/pull/18953)) [@mroeschke](https://github.com/mroeschke) +- add 'rapids-init-pip' to test_cudf_polars_polars_tests.sh ([#18951](https://github.com/NVIDIA/cudf/pull/18951)) [@jameslamb](https://github.com/jameslamb) +- parameterized ucx / ucxx ([#18949](https://github.com/NVIDIA/cudf/pull/18949)) [@quasiben](https://github.com/quasiben) +- Rework cudf::sorted_order implementation for faster compile ([#18948](https://github.com/NVIDIA/cudf/pull/18948)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated Series methods, isclose ([#18947](https://github.com/NVIDIA/cudf/pull/18947)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated groupby.collect ([#18946](https://github.com/NVIDIA/cudf/pull/18946)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated get_dummies(cats=, ...) ([#18944](https://github.com/NVIDIA/cudf/pull/18944)) [@mroeschke](https://github.com/mroeschke) +- Add .python_typecode and .typestr attributes to DataType ([#18941](https://github.com/NVIDIA/cudf/pull/18941)) [@Matt711](https://github.com/Matt711) +- Remove deprecated APIs ([#18933](https://github.com/NVIDIA/cudf/pull/18933)) [@vuule](https://github.com/vuule) +- Remove cudf.Scalar ([#18927](https://github.com/NVIDIA/cudf/pull/18927)) [@mroeschke](https://github.com/mroeschke) +- Add #pragma once to prevent redundant includes and speed up compilation ([#18925](https://github.com/NVIDIA/cudf/pull/18925)) [@PointKernel](https://github.com/PointKernel) +- Bump polars version to <1.31 ([#18920](https://github.com/NVIDIA/cudf/pull/18920)) [@Matt711](https://github.com/Matt711) +- Apply primitive row operators into hash join ([#18896](https://github.com/NVIDIA/cudf/pull/18896)) [@PointKernel](https://github.com/PointKernel) +- Branch 25.08 merge branch 25.06 ([#18895](https://github.com/NVIDIA/cudf/pull/18895)) [@vyasr](https://github.com/vyasr) +- Remove deprecated `cudf::io::host_buffer` ([#18881](https://github.com/NVIDIA/cudf/pull/18881)) [@Matt711](https://github.com/Matt711) +- Fix decompression scratch size in AUTO mode ([#18878](https://github.com/NVIDIA/cudf/pull/18878)) [@vuule](https://github.com/vuule) +- Apply linter suggestions to cuIO code ([#18876](https://github.com/NVIDIA/cudf/pull/18876)) [@vuule](https://github.com/vuule) +- xfail pandas unit tests that fail with cudf.pandas ([#18872](https://github.com/NVIDIA/cudf/pull/18872)) [@mroeschke](https://github.com/mroeschke) +- Branch 25.08 merge branch 25.06 ([#18855](https://github.com/NVIDIA/cudf/pull/18855)) [@vyasr](https://github.com/vyasr) +- Add support for extended dtypes in `cudf.pandas` ([#18832](https://github.com/NVIDIA/cudf/pull/18832)) [@galipremsagar](https://github.com/galipremsagar) +- Auto merge fix for branch-25.08 ([#18824](https://github.com/NVIDIA/cudf/pull/18824)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-25.06 to branch-25.08 ([#18817](https://github.com/NVIDIA/cudf/pull/18817)) [@Matt711](https://github.com/Matt711) +- Forward-merge branch-25.06 to branch-25.08 ([#18756](https://github.com/NVIDIA/cudf/pull/18756)) [@Matt711](https://github.com/Matt711) +- Fix auto merge conflict for branch-25.08 ([#18733](https://github.com/NVIDIA/cudf/pull/18733)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-25.06 to branch-25.08 ([#18698](https://github.com/NVIDIA/cudf/pull/18698)) [@Matt711](https://github.com/Matt711) +- Fix merge conflict for auto-merger 25.06 to 25.08 ([#18693](https://github.com/NVIDIA/cudf/pull/18693)) [@davidwendt](https://github.com/davidwendt) +- Fix merge conflict: branch-25.06 into branch-25.08 ([#18668](https://github.com/NVIDIA/cudf/pull/18668)) [@davidwendt](https://github.com/davidwendt) +- Make cuda12 as JNI default ([#18651](https://github.com/NVIDIA/cudf/pull/18651)) [@pxLi](https://github.com/pxLi) +- Forward-merge branch-25.06 into branch-25.08 ([#18647](https://github.com/NVIDIA/cudf/pull/18647)) [@bdice](https://github.com/bdice) +- Fix merge branch-25.06 into branch-25.08 ([#18622](https://github.com/NVIDIA/cudf/pull/18622)) [@davidwendt](https://github.com/davidwendt) +- Store polars Series instead of pyarrow Array in cudf_polars LiteralColumn expr ([#18564](https://github.com/NVIDIA/cudf/pull/18564)) [@mroeschke](https://github.com/mroeschke) +- Refactor strings split/record with whitespace logic ([#18560](https://github.com/NVIDIA/cudf/pull/18560)) [@davidwendt](https://github.com/davidwendt) +- Refactor hash join with multiset ([#18021](https://github.com/NVIDIA/cudf/pull/18021)) [@PointKernel](https://github.com/PointKernel) # cudf 25.06.00 (5 Jun 2025) ## 🚨 Breaking Changes -- Remove cudf.BaseIndex ([#18751](https://github.com/rapidsai/cudf/pull/18751)) [@mroeschke](https://github.com/mroeschke) -- Implement `BIT_COUNT` unary operation ([#18589](https://github.com/rapidsai/cudf/pull/18589)) [@ttnghia](https://github.com/ttnghia) -- Expose column chunk metadata in `read_parquet_metadata()` ([#18579](https://github.com/rapidsai/cudf/pull/18579)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix overflow for `MERGE_M2` groupby aggregation ([#18546](https://github.com/rapidsai/cudf/pull/18546)) [@ttnghia](https://github.com/ttnghia) -- Deduplicate parquet physical type enums ([#18526](https://github.com/rapidsai/cudf/pull/18526)) [@mhaseeb123](https://github.com/mhaseeb123) -- Implemented String Output & User-data Support for Transforms ([#18490](https://github.com/rapidsai/cudf/pull/18490)) [@lamarrr](https://github.com/lamarrr) -- Promote Parquet type enums to enum classes ([#18441](https://github.com/rapidsai/cudf/pull/18441)) [@mhaseeb123](https://github.com/mhaseeb123) -- Move parquet schema types and structs to public headers ([#18424](https://github.com/rapidsai/cudf/pull/18424)) [@mhaseeb123](https://github.com/mhaseeb123) -- Start removal of vector factories with `_sync` suffix by deprecating them and adding versions without the suffix ([#18414](https://github.com/rapidsai/cudf/pull/18414)) [@vuule](https://github.com/vuule) -- Skip decoding of pages marked as pruned in PQ reader ([#18347](https://github.com/rapidsai/cudf/pull/18347)) [@mhaseeb123](https://github.com/mhaseeb123) -- Deprecate nvtext subword tokenizer ([#18334](https://github.com/rapidsai/cudf/pull/18334)) [@davidwendt](https://github.com/davidwendt) -- Add standard data ingestion pipelines to pylibcudf for ndarrays ([#18311](https://github.com/rapidsai/cudf/pull/18311)) [@Matt711](https://github.com/Matt711) -- Remove extranous modules from top level cudf namespace ([#18287](https://github.com/rapidsai/cudf/pull/18287)) [@mroeschke](https://github.com/mroeschke) -- Add Keep Option Parameter to Distinct ([#18237](https://github.com/rapidsai/cudf/pull/18237)) [@warrickhe](https://github.com/warrickhe) -- Update to CCCL 2.8.x with no CCCL patches ([#18235](https://github.com/rapidsai/cudf/pull/18235)) [@bdice](https://github.com/bdice) +- Remove cudf.BaseIndex ([#18751](https://github.com/NVIDIA/cudf/pull/18751)) [@mroeschke](https://github.com/mroeschke) +- Implement `BIT_COUNT` unary operation ([#18589](https://github.com/NVIDIA/cudf/pull/18589)) [@ttnghia](https://github.com/ttnghia) +- Expose column chunk metadata in `read_parquet_metadata()` ([#18579](https://github.com/NVIDIA/cudf/pull/18579)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix overflow for `MERGE_M2` groupby aggregation ([#18546](https://github.com/NVIDIA/cudf/pull/18546)) [@ttnghia](https://github.com/ttnghia) +- Deduplicate parquet physical type enums ([#18526](https://github.com/NVIDIA/cudf/pull/18526)) [@mhaseeb123](https://github.com/mhaseeb123) +- Implemented String Output & User-data Support for Transforms ([#18490](https://github.com/NVIDIA/cudf/pull/18490)) [@lamarrr](https://github.com/lamarrr) +- Promote Parquet type enums to enum classes ([#18441](https://github.com/NVIDIA/cudf/pull/18441)) [@mhaseeb123](https://github.com/mhaseeb123) +- Move parquet schema types and structs to public headers ([#18424](https://github.com/NVIDIA/cudf/pull/18424)) [@mhaseeb123](https://github.com/mhaseeb123) +- Start removal of vector factories with `_sync` suffix by deprecating them and adding versions without the suffix ([#18414](https://github.com/NVIDIA/cudf/pull/18414)) [@vuule](https://github.com/vuule) +- Skip decoding of pages marked as pruned in PQ reader ([#18347](https://github.com/NVIDIA/cudf/pull/18347)) [@mhaseeb123](https://github.com/mhaseeb123) +- Deprecate nvtext subword tokenizer ([#18334](https://github.com/NVIDIA/cudf/pull/18334)) [@davidwendt](https://github.com/davidwendt) +- Add standard data ingestion pipelines to pylibcudf for ndarrays ([#18311](https://github.com/NVIDIA/cudf/pull/18311)) [@Matt711](https://github.com/Matt711) +- Remove extranous modules from top level cudf namespace ([#18287](https://github.com/NVIDIA/cudf/pull/18287)) [@mroeschke](https://github.com/mroeschke) +- Add Keep Option Parameter to Distinct ([#18237](https://github.com/NVIDIA/cudf/pull/18237)) [@warrickhe](https://github.com/warrickhe) +- Update to CCCL 2.8.x with no CCCL patches ([#18235](https://github.com/NVIDIA/cudf/pull/18235)) [@bdice](https://github.com/bdice) ## 🐛 Bug Fixes -- Disable pytest benchmark for Narwhals CI job ([#19074](https://github.com/rapidsai/cudf/pull/19074)) [@Matt711](https://github.com/Matt711) -- Avoid undefined behaviour in rolling_store_output_functor ([#19069](https://github.com/rapidsai/cudf/pull/19069)) [@wence-](https://github.com/wence-) -- Filter out pkg_resources UserWarning to make nightly CI pass ([#19058](https://github.com/rapidsai/cudf/pull/19058)) [@Matt711](https://github.com/Matt711) -- Pin deltalake to <1.0.0 ([#19017](https://github.com/rapidsai/cudf/pull/19017)) [@Matt711](https://github.com/Matt711) -- [BUG] Incorrectly getting the caller's frame when searching for locals and globals in cudf.pandas ([#18979](https://github.com/rapidsai/cudf/pull/18979)) [@Matt711](https://github.com/Matt711) -- Ensure gc fixture is used in custreamz test ([#18915](https://github.com/rapidsai/cudf/pull/18915)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix a potential segfault in PQ reader's number of rows per source calculation ([#18906](https://github.com/rapidsai/cudf/pull/18906)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix Dataframe `getitem` when `MultiIndex` columns exist ([#18880](https://github.com/rapidsai/cudf/pull/18880)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure __eq__/__ne__ between Columns in public objects don't return bool ([#18875](https://github.com/rapidsai/cudf/pull/18875)) [@mroeschke](https://github.com/mroeschke) -- Fix fencepost error in `Repartition` task generation ([#18854](https://github.com/rapidsai/cudf/pull/18854)) [@wence-](https://github.com/wence-) -- Fix cudf_polars pl.col(...).len() always excluding null values ([#18849](https://github.com/rapidsai/cudf/pull/18849)) [@mroeschke](https://github.com/mroeschke) -- Throw a descriptive exception in Parquet reader when trying to read files with more than two billion rows ([#18835](https://github.com/rapidsai/cudf/pull/18835)) [@mhaseeb123](https://github.com/mhaseeb123) -- Skip a decompression test ([#18825](https://github.com/rapidsai/cudf/pull/18825)) [@vuule](https://github.com/vuule) -- Update strings benchmarks to use alloc_size column/table function ([#18822](https://github.com/rapidsai/cudf/pull/18822)) [@davidwendt](https://github.com/davidwendt) -- Fix host decompression of empty DEFLATE data ([#18805](https://github.com/rapidsai/cudf/pull/18805)) [@vuule](https://github.com/vuule) -- Avoid going OOM in `test_row_limit_exceed_raises` by using dummy array ([#18802](https://github.com/rapidsai/cudf/pull/18802)) [@Matt711](https://github.com/Matt711) -- Fix host decompression of empty Snappy data ([#18800](https://github.com/rapidsai/cudf/pull/18800)) [@vuule](https://github.com/vuule) -- Skip test that fails due to polars issue ([#18787](https://github.com/rapidsai/cudf/pull/18787)) [@wence-](https://github.com/wence-) -- Ensure scalar dtype is always set in from_py ([#18780](https://github.com/rapidsai/cudf/pull/18780)) [@vyasr](https://github.com/vyasr) -- Fix reading of Snappy compressed Avro files ([#18774](https://github.com/rapidsai/cudf/pull/18774)) [@vuule](https://github.com/vuule) -- Fix missing semicolon in label_bins.cu ([#18765](https://github.com/rapidsai/cudf/pull/18765)) [@evanramos-nvidia](https://github.com/evanramos-nvidia) -- Fix noexcept annotations on strings_column_view ([#18763](https://github.com/rapidsai/cudf/pull/18763)) [@wence-](https://github.com/wence-) -- Fix integer overflows in pylibcudf `from_column_view_of_arbitrary` ([#18758](https://github.com/rapidsai/cudf/pull/18758)) [@wence-](https://github.com/wence-) -- Fix overflow case and clean up some logic ([#18734](https://github.com/rapidsai/cudf/pull/18734)) [@vyasr](https://github.com/vyasr) -- Link to `nvtx3::nvtx3-cpp` instead of `nvToolsExt` ([#18730](https://github.com/rapidsai/cudf/pull/18730)) [@jakirkham](https://github.com/jakirkham) -- Revise `DaskIntegration` protocol to align with `rapidsmpf` ([#18720](https://github.com/rapidsai/cudf/pull/18720)) [@rjzamora](https://github.com/rjzamora) -- Fix `skip_compression` option in the Parquet writer with host compression ([#18714](https://github.com/rapidsai/cudf/pull/18714)) [@vuule](https://github.com/vuule) -- Add missing header ([#18671](https://github.com/rapidsai/cudf/pull/18671)) [@vyasr](https://github.com/vyasr) -- Revert "Set flag to always use unsafe atomic storage" ([#18657](https://github.com/rapidsai/cudf/pull/18657)) [@PointKernel](https://github.com/PointKernel) -- Fix optional operator* called on a disengaged value in clamp.cu ([#18655](https://github.com/rapidsai/cudf/pull/18655)) [@davidwendt](https://github.com/davidwendt) -- Add missing header to host_memory.cpp ([#18649](https://github.com/rapidsai/cudf/pull/18649)) [@alliepiper](https://github.com/alliepiper) -- Fix device compression when writing Parquet files without using nvCOMP ([#18644](https://github.com/rapidsai/cudf/pull/18644)) [@vuule](https://github.com/vuule) -- Add CUDA_ARCHITECTURES setting to cpp-linters script ([#18637](https://github.com/rapidsai/cudf/pull/18637)) [@davidwendt](https://github.com/davidwendt) -- Pin to cython<3.1 ([#18617](https://github.com/rapidsai/cudf/pull/18617)) [@wence-](https://github.com/wence-) -- Fix `DataFrame.memory_usage` output order ([#18595](https://github.com/rapidsai/cudf/pull/18595)) [@mroeschke](https://github.com/mroeschke) -- Set flag to always use unsafe atomic storage ([#18590](https://github.com/rapidsai/cudf/pull/18590)) [@PointKernel](https://github.com/PointKernel) -- Update KvikIO S3 endpoint usage ([#18565](https://github.com/rapidsai/cudf/pull/18565)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Skip cuml third-party integration tests that may segfault ([#18561](https://github.com/rapidsai/cudf/pull/18561)) [@Matt711](https://github.com/Matt711) -- Allow .iloc with cuDF objects as column indexers ([#18558](https://github.com/rapidsai/cudf/pull/18558)) [@mroeschke](https://github.com/mroeschke) -- Fix overflow for `MERGE_M2` groupby aggregation ([#18546](https://github.com/rapidsai/cudf/pull/18546)) [@ttnghia](https://github.com/ttnghia) -- Add back cudf root ([#18544](https://github.com/rapidsai/cudf/pull/18544)) [@vyasr](https://github.com/vyasr) -- Change default memory resource for 'distributed' cudf-polars ([#18531](https://github.com/rapidsai/cudf/pull/18531)) [@rjzamora](https://github.com/rjzamora) -- Fix copy-on-write buffer separation and cleanup ([#18530](https://github.com/rapidsai/cudf/pull/18530)) [@galipremsagar](https://github.com/galipremsagar) -- Fix cpp examples cmake to use the rapids_config.cmake ([#18501](https://github.com/rapidsai/cudf/pull/18501)) [@davidwendt](https://github.com/davidwendt) -- Rename rapidsmp to rapidsmpf ([#18493](https://github.com/rapidsai/cudf/pull/18493)) [@rjzamora](https://github.com/rjzamora) -- Fix compilation with the C++20 standard ([#18486](https://github.com/rapidsai/cudf/pull/18486)) [@vuule](https://github.com/vuule) -- Fix an error when reading some compressed Parquet V2 files ([#18478](https://github.com/rapidsai/cudf/pull/18478)) [@vuule](https://github.com/vuule) -- Support title-case characters in strings capitalize() and title() APIs ([#18457](https://github.com/rapidsai/cudf/pull/18457)) [@davidwendt](https://github.com/davidwendt) -- Ensure DataFrame column label operations reset label_dtype ([#18452](https://github.com/rapidsai/cudf/pull/18452)) [@mroeschke](https://github.com/mroeschke) -- Fix a segfault when reading a Parquet file with unsupported compression type ([#18451](https://github.com/rapidsai/cudf/pull/18451)) [@vuule](https://github.com/vuule) -- Fix logger macros ([#18444](https://github.com/rapidsai/cudf/pull/18444)) [@vyasr](https://github.com/vyasr) -- Fix auto-detection of compression type in host-side decompression ([#18440](https://github.com/rapidsai/cudf/pull/18440)) [@shrshi](https://github.com/shrshi) -- Use delete not free to release data allocated with new ([#18412](https://github.com/rapidsai/cudf/pull/18412)) [@wence-](https://github.com/wence-) -- Fix synchronization issues in host compression and decompression ([#18395](https://github.com/rapidsai/cudf/pull/18395)) [@vuule](https://github.com/vuule) -- Update Dask array-conversion handling ([#18382](https://github.com/rapidsai/cudf/pull/18382)) [@rjzamora](https://github.com/rjzamora) -- Fixed indexing on empty DataFrame with no columns ([#18381](https://github.com/rapidsai/cudf/pull/18381)) [@TomAugspurger](https://github.com/TomAugspurger) -- Deterministic hashing for DataFrameScan nodes in cudf-polars multi-partition executor ([#18351](https://github.com/rapidsai/cudf/pull/18351)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix index of right table in unary operators in AST, in Joins ([#18333](https://github.com/rapidsai/cudf/pull/18333)) [@karthikeyann](https://github.com/karthikeyann) -- Add offsetalator to contiguous-split ([#18312](https://github.com/rapidsai/cudf/pull/18312)) [@davidwendt](https://github.com/davidwendt) -- Support large strings in nvtext vocabulary-tokenizer ([#18283](https://github.com/rapidsai/cudf/pull/18283)) [@davidwendt](https://github.com/davidwendt) -- Handle empty aggregations in multi-partition cudf.polars group_by ([#18277](https://github.com/rapidsai/cudf/pull/18277)) [@TomAugspurger](https://github.com/TomAugspurger) +- Disable pytest benchmark for Narwhals CI job ([#19074](https://github.com/NVIDIA/cudf/pull/19074)) [@Matt711](https://github.com/Matt711) +- Avoid undefined behaviour in rolling_store_output_functor ([#19069](https://github.com/NVIDIA/cudf/pull/19069)) [@wence-](https://github.com/wence-) +- Filter out pkg_resources UserWarning to make nightly CI pass ([#19058](https://github.com/NVIDIA/cudf/pull/19058)) [@Matt711](https://github.com/Matt711) +- Pin deltalake to <1.0.0 ([#19017](https://github.com/NVIDIA/cudf/pull/19017)) [@Matt711](https://github.com/Matt711) +- [BUG] Incorrectly getting the caller's frame when searching for locals and globals in cudf.pandas ([#18979](https://github.com/NVIDIA/cudf/pull/18979)) [@Matt711](https://github.com/Matt711) +- Ensure gc fixture is used in custreamz test ([#18915](https://github.com/NVIDIA/cudf/pull/18915)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix a potential segfault in PQ reader's number of rows per source calculation ([#18906](https://github.com/NVIDIA/cudf/pull/18906)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix Dataframe `getitem` when `MultiIndex` columns exist ([#18880](https://github.com/NVIDIA/cudf/pull/18880)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure __eq__/__ne__ between Columns in public objects don't return bool ([#18875](https://github.com/NVIDIA/cudf/pull/18875)) [@mroeschke](https://github.com/mroeschke) +- Fix fencepost error in `Repartition` task generation ([#18854](https://github.com/NVIDIA/cudf/pull/18854)) [@wence-](https://github.com/wence-) +- Fix cudf_polars pl.col(...).len() always excluding null values ([#18849](https://github.com/NVIDIA/cudf/pull/18849)) [@mroeschke](https://github.com/mroeschke) +- Throw a descriptive exception in Parquet reader when trying to read files with more than two billion rows ([#18835](https://github.com/NVIDIA/cudf/pull/18835)) [@mhaseeb123](https://github.com/mhaseeb123) +- Skip a decompression test ([#18825](https://github.com/NVIDIA/cudf/pull/18825)) [@vuule](https://github.com/vuule) +- Update strings benchmarks to use alloc_size column/table function ([#18822](https://github.com/NVIDIA/cudf/pull/18822)) [@davidwendt](https://github.com/davidwendt) +- Fix host decompression of empty DEFLATE data ([#18805](https://github.com/NVIDIA/cudf/pull/18805)) [@vuule](https://github.com/vuule) +- Avoid going OOM in `test_row_limit_exceed_raises` by using dummy array ([#18802](https://github.com/NVIDIA/cudf/pull/18802)) [@Matt711](https://github.com/Matt711) +- Fix host decompression of empty Snappy data ([#18800](https://github.com/NVIDIA/cudf/pull/18800)) [@vuule](https://github.com/vuule) +- Skip test that fails due to polars issue ([#18787](https://github.com/NVIDIA/cudf/pull/18787)) [@wence-](https://github.com/wence-) +- Ensure scalar dtype is always set in from_py ([#18780](https://github.com/NVIDIA/cudf/pull/18780)) [@vyasr](https://github.com/vyasr) +- Fix reading of Snappy compressed Avro files ([#18774](https://github.com/NVIDIA/cudf/pull/18774)) [@vuule](https://github.com/vuule) +- Fix missing semicolon in label_bins.cu ([#18765](https://github.com/NVIDIA/cudf/pull/18765)) [@evanramos-nvidia](https://github.com/evanramos-nvidia) +- Fix noexcept annotations on strings_column_view ([#18763](https://github.com/NVIDIA/cudf/pull/18763)) [@wence-](https://github.com/wence-) +- Fix integer overflows in pylibcudf `from_column_view_of_arbitrary` ([#18758](https://github.com/NVIDIA/cudf/pull/18758)) [@wence-](https://github.com/wence-) +- Fix overflow case and clean up some logic ([#18734](https://github.com/NVIDIA/cudf/pull/18734)) [@vyasr](https://github.com/vyasr) +- Link to `nvtx3::nvtx3-cpp` instead of `nvToolsExt` ([#18730](https://github.com/NVIDIA/cudf/pull/18730)) [@jakirkham](https://github.com/jakirkham) +- Revise `DaskIntegration` protocol to align with `rapidsmpf` ([#18720](https://github.com/NVIDIA/cudf/pull/18720)) [@rjzamora](https://github.com/rjzamora) +- Fix `skip_compression` option in the Parquet writer with host compression ([#18714](https://github.com/NVIDIA/cudf/pull/18714)) [@vuule](https://github.com/vuule) +- Add missing header ([#18671](https://github.com/NVIDIA/cudf/pull/18671)) [@vyasr](https://github.com/vyasr) +- Revert "Set flag to always use unsafe atomic storage" ([#18657](https://github.com/NVIDIA/cudf/pull/18657)) [@PointKernel](https://github.com/PointKernel) +- Fix optional operator* called on a disengaged value in clamp.cu ([#18655](https://github.com/NVIDIA/cudf/pull/18655)) [@davidwendt](https://github.com/davidwendt) +- Add missing header to host_memory.cpp ([#18649](https://github.com/NVIDIA/cudf/pull/18649)) [@alliepiper](https://github.com/alliepiper) +- Fix device compression when writing Parquet files without using nvCOMP ([#18644](https://github.com/NVIDIA/cudf/pull/18644)) [@vuule](https://github.com/vuule) +- Add CUDA_ARCHITECTURES setting to cpp-linters script ([#18637](https://github.com/NVIDIA/cudf/pull/18637)) [@davidwendt](https://github.com/davidwendt) +- Pin to cython<3.1 ([#18617](https://github.com/NVIDIA/cudf/pull/18617)) [@wence-](https://github.com/wence-) +- Fix `DataFrame.memory_usage` output order ([#18595](https://github.com/NVIDIA/cudf/pull/18595)) [@mroeschke](https://github.com/mroeschke) +- Set flag to always use unsafe atomic storage ([#18590](https://github.com/NVIDIA/cudf/pull/18590)) [@PointKernel](https://github.com/PointKernel) +- Update KvikIO S3 endpoint usage ([#18565](https://github.com/NVIDIA/cudf/pull/18565)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Skip cuml third-party integration tests that may segfault ([#18561](https://github.com/NVIDIA/cudf/pull/18561)) [@Matt711](https://github.com/Matt711) +- Allow .iloc with cuDF objects as column indexers ([#18558](https://github.com/NVIDIA/cudf/pull/18558)) [@mroeschke](https://github.com/mroeschke) +- Fix overflow for `MERGE_M2` groupby aggregation ([#18546](https://github.com/NVIDIA/cudf/pull/18546)) [@ttnghia](https://github.com/ttnghia) +- Add back cudf root ([#18544](https://github.com/NVIDIA/cudf/pull/18544)) [@vyasr](https://github.com/vyasr) +- Change default memory resource for 'distributed' cudf-polars ([#18531](https://github.com/NVIDIA/cudf/pull/18531)) [@rjzamora](https://github.com/rjzamora) +- Fix copy-on-write buffer separation and cleanup ([#18530](https://github.com/NVIDIA/cudf/pull/18530)) [@galipremsagar](https://github.com/galipremsagar) +- Fix cpp examples cmake to use the rapids_config.cmake ([#18501](https://github.com/NVIDIA/cudf/pull/18501)) [@davidwendt](https://github.com/davidwendt) +- Rename rapidsmp to rapidsmpf ([#18493](https://github.com/NVIDIA/cudf/pull/18493)) [@rjzamora](https://github.com/rjzamora) +- Fix compilation with the C++20 standard ([#18486](https://github.com/NVIDIA/cudf/pull/18486)) [@vuule](https://github.com/vuule) +- Fix an error when reading some compressed Parquet V2 files ([#18478](https://github.com/NVIDIA/cudf/pull/18478)) [@vuule](https://github.com/vuule) +- Support title-case characters in strings capitalize() and title() APIs ([#18457](https://github.com/NVIDIA/cudf/pull/18457)) [@davidwendt](https://github.com/davidwendt) +- Ensure DataFrame column label operations reset label_dtype ([#18452](https://github.com/NVIDIA/cudf/pull/18452)) [@mroeschke](https://github.com/mroeschke) +- Fix a segfault when reading a Parquet file with unsupported compression type ([#18451](https://github.com/NVIDIA/cudf/pull/18451)) [@vuule](https://github.com/vuule) +- Fix logger macros ([#18444](https://github.com/NVIDIA/cudf/pull/18444)) [@vyasr](https://github.com/vyasr) +- Fix auto-detection of compression type in host-side decompression ([#18440](https://github.com/NVIDIA/cudf/pull/18440)) [@shrshi](https://github.com/shrshi) +- Use delete not free to release data allocated with new ([#18412](https://github.com/NVIDIA/cudf/pull/18412)) [@wence-](https://github.com/wence-) +- Fix synchronization issues in host compression and decompression ([#18395](https://github.com/NVIDIA/cudf/pull/18395)) [@vuule](https://github.com/vuule) +- Update Dask array-conversion handling ([#18382](https://github.com/NVIDIA/cudf/pull/18382)) [@rjzamora](https://github.com/rjzamora) +- Fixed indexing on empty DataFrame with no columns ([#18381](https://github.com/NVIDIA/cudf/pull/18381)) [@TomAugspurger](https://github.com/TomAugspurger) +- Deterministic hashing for DataFrameScan nodes in cudf-polars multi-partition executor ([#18351](https://github.com/NVIDIA/cudf/pull/18351)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix index of right table in unary operators in AST, in Joins ([#18333](https://github.com/NVIDIA/cudf/pull/18333)) [@karthikeyann](https://github.com/karthikeyann) +- Add offsetalator to contiguous-split ([#18312](https://github.com/NVIDIA/cudf/pull/18312)) [@davidwendt](https://github.com/davidwendt) +- Support large strings in nvtext vocabulary-tokenizer ([#18283](https://github.com/NVIDIA/cudf/pull/18283)) [@davidwendt](https://github.com/davidwendt) +- Handle empty aggregations in multi-partition cudf.polars group_by ([#18277](https://github.com/NVIDIA/cudf/pull/18277)) [@TomAugspurger](https://github.com/TomAugspurger) ## 📖 Documentation -- Docs for streaming executor options ([#18934](https://github.com/rapidsai/cudf/pull/18934)) [@quasiben](https://github.com/quasiben) -- Fix some duplicate toctree issues and improve groupby docs ([#18580](https://github.com/rapidsai/cudf/pull/18580)) [@vyasr](https://github.com/vyasr) -- [DOC] Running libcudf benchmarks and comparing output results ([#18548](https://github.com/rapidsai/cudf/pull/18548)) [@Matt711](https://github.com/Matt711) -- Fix doxygen usage of the contraction for it is ([#18517](https://github.com/rapidsai/cudf/pull/18517)) [@davidwendt](https://github.com/davidwendt) -- Clarify [@brief tag as description/title on documentation guide ([#18515](https://github.com/rapidsai/cudf/pull/18515)) @davidwendt](https://github.com/brief tag as description/title on documentation guide ([#18515](https://github.com/rapidsai/cudf/pull/18515)) @davidwendt) -- [DOC] Improve clarity in parquet APIs set_row_groups and set_columns parquet ([#18466](https://github.com/rapidsai/cudf/pull/18466)) [@Matt711](https://github.com/Matt711) -- Add a usage page to cudf-polars documentation ([#18460](https://github.com/rapidsai/cudf/pull/18460)) [@Matt711](https://github.com/Matt711) -- [DOC] Fix typo in CONTRIBUTING.md on build type tests ([#18456](https://github.com/rapidsai/cudf/pull/18456)) [@JigaoLuo](https://github.com/JigaoLuo) -- improve docs related to documentation contribution ([#18418](https://github.com/rapidsai/cudf/pull/18418)) [@ncclementi](https://github.com/ncclementi) -- Add restart kernel note in cudf pandas docs ([#18374](https://github.com/rapidsai/cudf/pull/18374)) [@ncclementi](https://github.com/ncclementi) +- Docs for streaming executor options ([#18934](https://github.com/NVIDIA/cudf/pull/18934)) [@quasiben](https://github.com/quasiben) +- Fix some duplicate toctree issues and improve groupby docs ([#18580](https://github.com/NVIDIA/cudf/pull/18580)) [@vyasr](https://github.com/vyasr) +- [DOC] Running libcudf benchmarks and comparing output results ([#18548](https://github.com/NVIDIA/cudf/pull/18548)) [@Matt711](https://github.com/Matt711) +- Fix doxygen usage of the contraction for it is ([#18517](https://github.com/NVIDIA/cudf/pull/18517)) [@davidwendt](https://github.com/davidwendt) +- Clarify [@brief tag as description/title on documentation guide ([#18515](https://github.com/NVIDIA/cudf/pull/18515)) @davidwendt](https://github.com/brief tag as description/title on documentation guide ([#18515](https://github.com/NVIDIA/cudf/pull/18515)) @davidwendt) +- [DOC] Improve clarity in parquet APIs set_row_groups and set_columns parquet ([#18466](https://github.com/NVIDIA/cudf/pull/18466)) [@Matt711](https://github.com/Matt711) +- Add a usage page to cudf-polars documentation ([#18460](https://github.com/NVIDIA/cudf/pull/18460)) [@Matt711](https://github.com/Matt711) +- [DOC] Fix typo in CONTRIBUTING.md on build type tests ([#18456](https://github.com/NVIDIA/cudf/pull/18456)) [@JigaoLuo](https://github.com/JigaoLuo) +- improve docs related to documentation contribution ([#18418](https://github.com/NVIDIA/cudf/pull/18418)) [@ncclementi](https://github.com/ncclementi) +- Add restart kernel note in cudf pandas docs ([#18374](https://github.com/NVIDIA/cudf/pull/18374)) [@ncclementi](https://github.com/ncclementi) ## 🚀 New Features -- Add CLI argument to enable RMM async memory resource in PDS-H ([#18899](https://github.com/rapidsai/cudf/pull/18899)) [@pentschev](https://github.com/pentschev) -- Scan a headerless CSV file with column names provided ([#18816](https://github.com/rapidsai/cudf/pull/18816)) [@Matt711](https://github.com/Matt711) -- Add fast paths for `DataFrame.to_cupy` ([#18801](https://github.com/rapidsai/cudf/pull/18801)) [@Matt711](https://github.com/Matt711) -- Require `numba-cuda>=0.11.0` ([#18770](https://github.com/rapidsai/cudf/pull/18770)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Create a pylibcudf Column from a python iterable ([#18768](https://github.com/rapidsai/cudf/pull/18768)) [@Matt711](https://github.com/Matt711) -- Support `ConditianalJoin` via broadcasting in cudf-polars streaming engine ([#18723](https://github.com/rapidsai/cudf/pull/18723)) [@rjzamora](https://github.com/rjzamora) -- Experimental PQ reader utility to calculate total rows in input row groups ([#18716](https://github.com/rapidsai/cudf/pull/18716)) [@mhaseeb123](https://github.com/mhaseeb123) -- Extend `explain_query` to support printing the logical plan (pre lowered plan) ([#18708](https://github.com/rapidsai/cudf/pull/18708)) [@Matt711](https://github.com/Matt711) -- Reuse `libcudf` dependencies for Java JNI build when they are available ([#18682](https://github.com/rapidsai/cudf/pull/18682)) [@ttnghia](https://github.com/ttnghia) -- Add alloc_size member function to cudf::column and cudf::table ([#18639](https://github.com/rapidsai/cudf/pull/18639)) [@davidwendt](https://github.com/davidwendt) -- Print the physical cudf-polars plan in ``pdsh.py`` ([#18635](https://github.com/rapidsai/cudf/pull/18635)) [@rjzamora](https://github.com/rjzamora) -- String Transform Examples ([#18616](https://github.com/rapidsai/cudf/pull/18616)) [@lamarrr](https://github.com/lamarrr) -- Add streaming support for ``group_by -> n_unique`` to cudf-polars ([#18606](https://github.com/rapidsai/cudf/pull/18606)) [@rjzamora](https://github.com/rjzamora) -- Export cudf compiler flags and definitions ([#18604](https://github.com/rapidsai/cudf/pull/18604)) [@ttnghia](https://github.com/ttnghia) -- Implement `BIT_COUNT` unary operation ([#18589](https://github.com/rapidsai/cudf/pull/18589)) [@ttnghia](https://github.com/ttnghia) -- Expose column chunk metadata in `read_parquet_metadata()` ([#18579](https://github.com/rapidsai/cudf/pull/18579)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add APIs to check ORC and Parquet compression support at runtime ([#18578](https://github.com/rapidsai/cudf/pull/18578)) [@vuule](https://github.com/vuule) -- Add `Distinct` support to the cudf-polars streaming executor ([#18576](https://github.com/rapidsai/cudf/pull/18576)) [@rjzamora](https://github.com/rjzamora) -- Add support for large list host Arrow data conversion ([#18562](https://github.com/rapidsai/cudf/pull/18562)) [@vyasr](https://github.com/vyasr) -- Implement `BITWISE_AGG` aggregations (bitwise `AND`, `OR` and `XOR`) for sort-based groupby and reduction ([#18551](https://github.com/rapidsai/cudf/pull/18551)) [@ttnghia](https://github.com/ttnghia) -- Implement row group pruning with bloom filters in experimental PQ reader ([#18545](https://github.com/rapidsai/cudf/pull/18545)) [@mhaseeb123](https://github.com/mhaseeb123) -- Implement row group pruning with stats in experimental PQ reader ([#18543](https://github.com/rapidsai/cudf/pull/18543)) [@mhaseeb123](https://github.com/mhaseeb123) -- [JNI] Expose row-wise sha1 api ([#18540](https://github.com/rapidsai/cudf/pull/18540)) [@warrickhe](https://github.com/warrickhe) -- Add ``Sort`` + ``head/tail`` support to streaming cudf-polars executor ([#18538](https://github.com/rapidsai/cudf/pull/18538)) [@rjzamora](https://github.com/rjzamora) -- Add multi-partition MapFunction support to cudf-polars ([#18523](https://github.com/rapidsai/cudf/pull/18523)) [@rjzamora](https://github.com/rjzamora) -- Adds support for writing raw UTF-8 characters (without escaping) in the JSON writer ([#18508](https://github.com/rapidsai/cudf/pull/18508)) [@Matt711](https://github.com/Matt711) -- Support reading from device buffers in the pylibcudf IO APIs ([#18496](https://github.com/rapidsai/cudf/pull/18496)) [@Matt711](https://github.com/Matt711) -- Support multi-partition ``Select`` operations with aggregations ([#18492](https://github.com/rapidsai/cudf/pull/18492)) [@rjzamora](https://github.com/rjzamora) -- Implemented String Output & User-data Support for Transforms ([#18490](https://github.com/rapidsai/cudf/pull/18490)) [@lamarrr](https://github.com/lamarrr) -- Add a utility to bulk set multiple null masks ([#18489](https://github.com/rapidsai/cudf/pull/18489)) [@mhaseeb123](https://github.com/mhaseeb123) -- High level interface for experimental PQ reader and implementation of metadata APIs ([#18480](https://github.com/rapidsai/cudf/pull/18480)) [@mhaseeb123](https://github.com/mhaseeb123) -- Added `pylibcudf.utilities.is_ptds_enabled` ([#18467](https://github.com/rapidsai/cudf/pull/18467)) [@TomAugspurger](https://github.com/TomAugspurger) -- Add a public API for copying a table_view to device array ([#18450](https://github.com/rapidsai/cudf/pull/18450)) [@Matt711](https://github.com/Matt711) -- Support `cudf-polars` `cast_time_unit` ([#18442](https://github.com/rapidsai/cudf/pull/18442)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support creating a pylibcudf Column from a host array ([#18425](https://github.com/rapidsai/cudf/pull/18425)) [@Matt711](https://github.com/Matt711) -- Move parquet schema types and structs to public headers ([#18424](https://github.com/rapidsai/cudf/pull/18424)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add optional dtype argument to `Scalar.from_any` ([#18415](https://github.com/rapidsai/cudf/pull/18415)) [@Matt711](https://github.com/Matt711) -- Expose `cudf::chunked_pack` in pylibcudf ([#18411](https://github.com/rapidsai/cudf/pull/18411)) [@wence-](https://github.com/wence-) -- Add support for long string columns in cudf::contiguous_split ([#18393](https://github.com/rapidsai/cudf/pull/18393)) [@nvdbaranec](https://github.com/nvdbaranec) -- Implemented String Input support for Transforms and Removed `jit::column_device_view` ([#18378](https://github.com/rapidsai/cudf/pull/18378)) [@lamarrr](https://github.com/lamarrr) -- Automatically dispatch between host and device decompression/compression based on the number of buffers ([#18363](https://github.com/rapidsai/cudf/pull/18363)) [@vuule](https://github.com/vuule) -- Expose join hash table load factor ([#18361](https://github.com/rapidsai/cudf/pull/18361)) [@PointKernel](https://github.com/PointKernel) -- Skip decoding of pages marked as pruned in PQ reader ([#18347](https://github.com/rapidsai/cudf/pull/18347)) [@mhaseeb123](https://github.com/mhaseeb123) -- Sort-based inner join for high-multiplicity tables ([#18318](https://github.com/rapidsai/cudf/pull/18318)) [@shrshi](https://github.com/shrshi) -- Support constructing pylibcudf Columns and Tables from views into arbitrary objects ([#18314](https://github.com/rapidsai/cudf/pull/18314)) [@vyasr](https://github.com/vyasr) -- Add standard data ingestion pipelines to pylibcudf for ndarrays ([#18311](https://github.com/rapidsai/cudf/pull/18311)) [@Matt711](https://github.com/Matt711) -- Support `cudf-polars` `isoyear` and `week` (`isoweek`) ([#18265](https://github.com/rapidsai/cudf/pull/18265)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add Keep Option Parameter to Distinct ([#18237](https://github.com/rapidsai/cudf/pull/18237)) [@warrickhe](https://github.com/warrickhe) -- Add rapidsmp shuffle support to cudf-polars ([#18231](https://github.com/rapidsai/cudf/pull/18231)) [@rjzamora](https://github.com/rjzamora) -- Support `cudf-polars` `strftime` ([#18181](https://github.com/rapidsai/cudf/pull/18181)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add benchmark for join operations with low build table cardinality ([#18105](https://github.com/rapidsai/cudf/pull/18105)) [@shrshi](https://github.com/shrshi) -- Add nvtext substring deduplication APIs (Part 2) ([#18104](https://github.com/rapidsai/cudf/pull/18104)) [@davidwendt](https://github.com/davidwendt) -- Support `include_file_paths` in cudf polars ([#18057](https://github.com/rapidsai/cudf/pull/18057)) [@Matt711](https://github.com/Matt711) -- Add support for the Arrow device capsule interfaces ([#15370](https://github.com/rapidsai/cudf/pull/15370)) [@vyasr](https://github.com/vyasr) +- Add CLI argument to enable RMM async memory resource in PDS-H ([#18899](https://github.com/NVIDIA/cudf/pull/18899)) [@pentschev](https://github.com/pentschev) +- Scan a headerless CSV file with column names provided ([#18816](https://github.com/NVIDIA/cudf/pull/18816)) [@Matt711](https://github.com/Matt711) +- Add fast paths for `DataFrame.to_cupy` ([#18801](https://github.com/NVIDIA/cudf/pull/18801)) [@Matt711](https://github.com/Matt711) +- Require `numba-cuda>=0.11.0` ([#18770](https://github.com/NVIDIA/cudf/pull/18770)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Create a pylibcudf Column from a python iterable ([#18768](https://github.com/NVIDIA/cudf/pull/18768)) [@Matt711](https://github.com/Matt711) +- Support `ConditianalJoin` via broadcasting in cudf-polars streaming engine ([#18723](https://github.com/NVIDIA/cudf/pull/18723)) [@rjzamora](https://github.com/rjzamora) +- Experimental PQ reader utility to calculate total rows in input row groups ([#18716](https://github.com/NVIDIA/cudf/pull/18716)) [@mhaseeb123](https://github.com/mhaseeb123) +- Extend `explain_query` to support printing the logical plan (pre lowered plan) ([#18708](https://github.com/NVIDIA/cudf/pull/18708)) [@Matt711](https://github.com/Matt711) +- Reuse `libcudf` dependencies for Java JNI build when they are available ([#18682](https://github.com/NVIDIA/cudf/pull/18682)) [@ttnghia](https://github.com/ttnghia) +- Add alloc_size member function to cudf::column and cudf::table ([#18639](https://github.com/NVIDIA/cudf/pull/18639)) [@davidwendt](https://github.com/davidwendt) +- Print the physical cudf-polars plan in ``pdsh.py`` ([#18635](https://github.com/NVIDIA/cudf/pull/18635)) [@rjzamora](https://github.com/rjzamora) +- String Transform Examples ([#18616](https://github.com/NVIDIA/cudf/pull/18616)) [@lamarrr](https://github.com/lamarrr) +- Add streaming support for ``group_by -> n_unique`` to cudf-polars ([#18606](https://github.com/NVIDIA/cudf/pull/18606)) [@rjzamora](https://github.com/rjzamora) +- Export cudf compiler flags and definitions ([#18604](https://github.com/NVIDIA/cudf/pull/18604)) [@ttnghia](https://github.com/ttnghia) +- Implement `BIT_COUNT` unary operation ([#18589](https://github.com/NVIDIA/cudf/pull/18589)) [@ttnghia](https://github.com/ttnghia) +- Expose column chunk metadata in `read_parquet_metadata()` ([#18579](https://github.com/NVIDIA/cudf/pull/18579)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add APIs to check ORC and Parquet compression support at runtime ([#18578](https://github.com/NVIDIA/cudf/pull/18578)) [@vuule](https://github.com/vuule) +- Add `Distinct` support to the cudf-polars streaming executor ([#18576](https://github.com/NVIDIA/cudf/pull/18576)) [@rjzamora](https://github.com/rjzamora) +- Add support for large list host Arrow data conversion ([#18562](https://github.com/NVIDIA/cudf/pull/18562)) [@vyasr](https://github.com/vyasr) +- Implement `BITWISE_AGG` aggregations (bitwise `AND`, `OR` and `XOR`) for sort-based groupby and reduction ([#18551](https://github.com/NVIDIA/cudf/pull/18551)) [@ttnghia](https://github.com/ttnghia) +- Implement row group pruning with bloom filters in experimental PQ reader ([#18545](https://github.com/NVIDIA/cudf/pull/18545)) [@mhaseeb123](https://github.com/mhaseeb123) +- Implement row group pruning with stats in experimental PQ reader ([#18543](https://github.com/NVIDIA/cudf/pull/18543)) [@mhaseeb123](https://github.com/mhaseeb123) +- [JNI] Expose row-wise sha1 api ([#18540](https://github.com/NVIDIA/cudf/pull/18540)) [@warrickhe](https://github.com/warrickhe) +- Add ``Sort`` + ``head/tail`` support to streaming cudf-polars executor ([#18538](https://github.com/NVIDIA/cudf/pull/18538)) [@rjzamora](https://github.com/rjzamora) +- Add multi-partition MapFunction support to cudf-polars ([#18523](https://github.com/NVIDIA/cudf/pull/18523)) [@rjzamora](https://github.com/rjzamora) +- Adds support for writing raw UTF-8 characters (without escaping) in the JSON writer ([#18508](https://github.com/NVIDIA/cudf/pull/18508)) [@Matt711](https://github.com/Matt711) +- Support reading from device buffers in the pylibcudf IO APIs ([#18496](https://github.com/NVIDIA/cudf/pull/18496)) [@Matt711](https://github.com/Matt711) +- Support multi-partition ``Select`` operations with aggregations ([#18492](https://github.com/NVIDIA/cudf/pull/18492)) [@rjzamora](https://github.com/rjzamora) +- Implemented String Output & User-data Support for Transforms ([#18490](https://github.com/NVIDIA/cudf/pull/18490)) [@lamarrr](https://github.com/lamarrr) +- Add a utility to bulk set multiple null masks ([#18489](https://github.com/NVIDIA/cudf/pull/18489)) [@mhaseeb123](https://github.com/mhaseeb123) +- High level interface for experimental PQ reader and implementation of metadata APIs ([#18480](https://github.com/NVIDIA/cudf/pull/18480)) [@mhaseeb123](https://github.com/mhaseeb123) +- Added `pylibcudf.utilities.is_ptds_enabled` ([#18467](https://github.com/NVIDIA/cudf/pull/18467)) [@TomAugspurger](https://github.com/TomAugspurger) +- Add a public API for copying a table_view to device array ([#18450](https://github.com/NVIDIA/cudf/pull/18450)) [@Matt711](https://github.com/Matt711) +- Support `cudf-polars` `cast_time_unit` ([#18442](https://github.com/NVIDIA/cudf/pull/18442)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support creating a pylibcudf Column from a host array ([#18425](https://github.com/NVIDIA/cudf/pull/18425)) [@Matt711](https://github.com/Matt711) +- Move parquet schema types and structs to public headers ([#18424](https://github.com/NVIDIA/cudf/pull/18424)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add optional dtype argument to `Scalar.from_any` ([#18415](https://github.com/NVIDIA/cudf/pull/18415)) [@Matt711](https://github.com/Matt711) +- Expose `cudf::chunked_pack` in pylibcudf ([#18411](https://github.com/NVIDIA/cudf/pull/18411)) [@wence-](https://github.com/wence-) +- Add support for long string columns in cudf::contiguous_split ([#18393](https://github.com/NVIDIA/cudf/pull/18393)) [@nvdbaranec](https://github.com/nvdbaranec) +- Implemented String Input support for Transforms and Removed `jit::column_device_view` ([#18378](https://github.com/NVIDIA/cudf/pull/18378)) [@lamarrr](https://github.com/lamarrr) +- Automatically dispatch between host and device decompression/compression based on the number of buffers ([#18363](https://github.com/NVIDIA/cudf/pull/18363)) [@vuule](https://github.com/vuule) +- Expose join hash table load factor ([#18361](https://github.com/NVIDIA/cudf/pull/18361)) [@PointKernel](https://github.com/PointKernel) +- Skip decoding of pages marked as pruned in PQ reader ([#18347](https://github.com/NVIDIA/cudf/pull/18347)) [@mhaseeb123](https://github.com/mhaseeb123) +- Sort-based inner join for high-multiplicity tables ([#18318](https://github.com/NVIDIA/cudf/pull/18318)) [@shrshi](https://github.com/shrshi) +- Support constructing pylibcudf Columns and Tables from views into arbitrary objects ([#18314](https://github.com/NVIDIA/cudf/pull/18314)) [@vyasr](https://github.com/vyasr) +- Add standard data ingestion pipelines to pylibcudf for ndarrays ([#18311](https://github.com/NVIDIA/cudf/pull/18311)) [@Matt711](https://github.com/Matt711) +- Support `cudf-polars` `isoyear` and `week` (`isoweek`) ([#18265](https://github.com/NVIDIA/cudf/pull/18265)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add Keep Option Parameter to Distinct ([#18237](https://github.com/NVIDIA/cudf/pull/18237)) [@warrickhe](https://github.com/warrickhe) +- Add rapidsmp shuffle support to cudf-polars ([#18231](https://github.com/NVIDIA/cudf/pull/18231)) [@rjzamora](https://github.com/rjzamora) +- Support `cudf-polars` `strftime` ([#18181](https://github.com/NVIDIA/cudf/pull/18181)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add benchmark for join operations with low build table cardinality ([#18105](https://github.com/NVIDIA/cudf/pull/18105)) [@shrshi](https://github.com/shrshi) +- Add nvtext substring deduplication APIs (Part 2) ([#18104](https://github.com/NVIDIA/cudf/pull/18104)) [@davidwendt](https://github.com/davidwendt) +- Support `include_file_paths` in cudf polars ([#18057](https://github.com/NVIDIA/cudf/pull/18057)) [@Matt711](https://github.com/Matt711) +- Add support for the Arrow device capsule interfaces ([#15370](https://github.com/NVIDIA/cudf/pull/15370)) [@vyasr](https://github.com/vyasr) ## 🛠️ Improvements -- use 'rapids-init-pip' in wheel CI, other CI changes ([#18902](https://github.com/rapidsai/cudf/pull/18902)) [@jameslamb](https://github.com/jameslamb) -- Avoid RecursionError in custreamz test ([#18887](https://github.com/rapidsai/cudf/pull/18887)) [@TomAugspurger](https://github.com/TomAugspurger) -- Update NumPy dependency in cudf.pandas-catboost integration test ([#18870](https://github.com/rapidsai/cudf/pull/18870)) [@Matt711](https://github.com/Matt711) -- CPU only execution for PDSH ([#18869](https://github.com/rapidsai/cudf/pull/18869)) [@quasiben](https://github.com/quasiben) -- Remove more top level cudf imports in core ([#18862](https://github.com/rapidsai/cudf/pull/18862)) [@mroeschke](https://github.com/mroeschke) -- Remove top level cudf imports in core ([#18857](https://github.com/rapidsai/cudf/pull/18857)) [@mroeschke](https://github.com/mroeschke) -- Add CUDF_INSTALL_DIR for JAVA build script ([#18852](https://github.com/rapidsai/cudf/pull/18852)) [@pxLi](https://github.com/pxLi) -- Call the correct `from_pandas` in `hdf` reader ([#18850](https://github.com/rapidsai/cudf/pull/18850)) [@galipremsagar](https://github.com/galipremsagar) -- Update `__all__` in `cudf_polars/dsl/ir.py` ([#18848](https://github.com/rapidsai/cudf/pull/18848)) [@Matt711](https://github.com/Matt711) -- Upload examples conda package ([#18847](https://github.com/rapidsai/cudf/pull/18847)) [@vyasr](https://github.com/vyasr) -- Add retries to prevent failures in occasionally slow CI runs ([#18843](https://github.com/rapidsai/cudf/pull/18843)) [@galipremsagar](https://github.com/galipremsagar) -- Finish CUDA 12.9 migration and use branch-25.06 workflows ([#18839](https://github.com/rapidsai/cudf/pull/18839)) [@bdice](https://github.com/bdice) -- Remove toplevel `import cudf` from window/tools/join directories ([#18833](https://github.com/rapidsai/cudf/pull/18833)) [@mroeschke](https://github.com/mroeschke) -- Remove toplevel `import cudf` from cudf/io files ([#18829](https://github.com/rapidsai/cudf/pull/18829)) [@mroeschke](https://github.com/mroeschke) -- Update pdsh benchmark script to support explain-only ([#18826](https://github.com/rapidsai/cudf/pull/18826)) [@TomAugspurger](https://github.com/TomAugspurger) -- Refactor UDF utils and add a hook to enable NRT when necessary ([#18823](https://github.com/rapidsai/cudf/pull/18823)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix memory access error in nvtext::edit_distance ([#18821](https://github.com/rapidsai/cudf/pull/18821)) [@davidwendt](https://github.com/davidwendt) -- Update to clang 20 ([#18818](https://github.com/rapidsai/cudf/pull/18818)) [@bdice](https://github.com/bdice) -- Reduce more data sizes of Python tests ([#18814](https://github.com/rapidsai/cudf/pull/18814)) [@mroeschke](https://github.com/mroeschke) -- Mark DataFrame.dtypes as an _external_only_api ([#18809](https://github.com/rapidsai/cudf/pull/18809)) [@mroeschke](https://github.com/mroeschke) -- Change calls to thrust::swap to cuda::std::swap ([#18808](https://github.com/rapidsai/cudf/pull/18808)) [@davidwendt](https://github.com/davidwendt) -- Move implemented BaseIndex methods over to Index ([#18807](https://github.com/rapidsai/cudf/pull/18807)) [@mroeschke](https://github.com/mroeschke) -- Improve pandas version fetching script ([#18793](https://github.com/rapidsai/cudf/pull/18793)) [@galipremsagar](https://github.com/galipremsagar) -- Change cudf::sort googlebench benchmarks to nvbench ([#18786](https://github.com/rapidsai/cudf/pull/18786)) [@davidwendt](https://github.com/davidwendt) -- Only warn in cudf.pandas if rmm mode explicitly set and rmm already configured ([#18785](https://github.com/rapidsai/cudf/pull/18785)) [@jcrist](https://github.com/jcrist) -- Quote head_rev in conda recipes ([#18784](https://github.com/rapidsai/cudf/pull/18784)) [@bdice](https://github.com/bdice) -- Move RangeIndex implementation below Index ([#18777](https://github.com/rapidsai/cudf/pull/18777)) [@mroeschke](https://github.com/mroeschke) -- Remove unecessary _Ravelled class ([#18771](https://github.com/rapidsai/cudf/pull/18771)) [@Matt711](https://github.com/Matt711) -- Remove pytest-rerunfailures ([#18766](https://github.com/rapidsai/cudf/pull/18766)) [@mroeschke](https://github.com/mroeschke) -- Replace from_arrow with direct calls Column/Table constructors in pylibcudf and cudf-polars tests ([#18762](https://github.com/rapidsai/cudf/pull/18762)) [@Matt711](https://github.com/Matt711) -- CUDA 12.9 use updated compression flags ([#18755](https://github.com/rapidsai/cudf/pull/18755)) [@robertmaynard](https://github.com/robertmaynard) -- fix(rattler): add `librmm` to host for `libcudf` to fix overlinking error ([#18754](https://github.com/rapidsai/cudf/pull/18754)) [@gforsyth](https://github.com/gforsyth) -- Remove the file name from the output in cudf-polars' explain APIs ([#18752](https://github.com/rapidsai/cudf/pull/18752)) [@Matt711](https://github.com/Matt711) -- Remove cudf.BaseIndex ([#18751](https://github.com/rapidsai/cudf/pull/18751)) [@mroeschke](https://github.com/mroeschke) -- Support creating a pylibcudf Column from a general ndarray ([#18744](https://github.com/rapidsai/cudf/pull/18744)) [@Matt711](https://github.com/Matt711) -- Improve lowering of ``Distinct`` IR nodes for high-cardinality data ([#18725](https://github.com/rapidsai/cudf/pull/18725)) [@rjzamora](https://github.com/rjzamora) -- Simplify Numba-CUDA MVC logic ([#18724](https://github.com/rapidsai/cudf/pull/18724)) [@bdice](https://github.com/bdice) -- Test with CUDA 12.9.0 ([#18721](https://github.com/rapidsai/cudf/pull/18721)) [@bdice](https://github.com/bdice) -- Add more `cudf.Series` microbenchmarks ([#18718](https://github.com/rapidsai/cudf/pull/18718)) [@Matt711](https://github.com/Matt711) -- Run unit-tests-cudf-pandas on branch-25.06 for nightly tests ([#18717](https://github.com/rapidsai/cudf/pull/18717)) [@davidwendt](https://github.com/davidwendt) -- Move `test_large_unique_categories_repr` to benchmarks ([#18715](https://github.com/rapidsai/cudf/pull/18715)) [@galipremsagar](https://github.com/galipremsagar) -- Allow `pylibcudf.Column` to consume objects exposing `__arrow_c_stream__` ([#18712](https://github.com/rapidsai/cudf/pull/18712)) [@mroeschke](https://github.com/mroeschke) -- Switch from printing to logging ([#18711](https://github.com/rapidsai/cudf/pull/18711)) [@vyasr](https://github.com/vyasr) -- Add Python tests for different compression implementations ([#18710](https://github.com/rapidsai/cudf/pull/18710)) [@vuule](https://github.com/vuule) -- Remove redundant xfails in cuml integration tests ([#18699](https://github.com/rapidsai/cudf/pull/18699)) [@Matt711](https://github.com/Matt711) -- ci: run unit-tests-cudf-pandas on `branch-25.06` workflow ([#18692](https://github.com/rapidsai/cudf/pull/18692)) [@gforsyth](https://github.com/gforsyth) -- Exclude librmm.so from auditwheel ([#18691](https://github.com/rapidsai/cudf/pull/18691)) [@bdice](https://github.com/bdice) -- Add C++ tests for different compression implementations ([#18690](https://github.com/rapidsai/cudf/pull/18690)) [@vuule](https://github.com/vuule) -- Improve runtime of cuDF Python unit tests ([#18689](https://github.com/rapidsai/cudf/pull/18689)) [@mroeschke](https://github.com/mroeschke) -- Require at least numba-cuda `0.10.1` ([#18688](https://github.com/rapidsai/cudf/pull/18688)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add `nvidia-cuda-{nvrtc, nvcc}` as a dependency for cuDF wheels ([#18686](https://github.com/rapidsai/cudf/pull/18686)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support rolling aggregations in in-memory cudf-polars execution ([#18681](https://github.com/rapidsai/cudf/pull/18681)) [@wence-](https://github.com/wence-) -- Replace ``parquet_blocksize`` with ``target_partition_size`` ([#18669](https://github.com/rapidsai/cudf/pull/18669)) [@rjzamora](https://github.com/rjzamora) -- Skip test_large_unique_categories_repr in CI ([#18666](https://github.com/rapidsai/cudf/pull/18666)) [@bdice](https://github.com/bdice) -- Locally import pyarrow.dataset and fsspec for `import cudf` performance ([#18663](https://github.com/rapidsai/cudf/pull/18663)) [@mroeschke](https://github.com/mroeschke) -- Disable `arm64` python tests ([#18662](https://github.com/rapidsai/cudf/pull/18662)) [@galipremsagar](https://github.com/galipremsagar) -- Pin numba-cuda>=0.9.0,!=0.10.0 due to CI hangs on ARM ([#18661](https://github.com/rapidsai/cudf/pull/18661)) [@mroeschke](https://github.com/mroeschke) -- Fix compile warnings in Java JNI ([#18660](https://github.com/rapidsai/cudf/pull/18660)) [@ttnghia](https://github.com/ttnghia) -- Drop ``Empty`` nodes from IR graph ([#18658](https://github.com/rapidsai/cudf/pull/18658)) [@rjzamora](https://github.com/rjzamora) -- Add support for Python 3.13 ([#18648](https://github.com/rapidsai/cudf/pull/18648)) [@gforsyth](https://github.com/gforsyth) -- Cleanup libcudf detail/aggregation.hpp/.cuh ([#18642](https://github.com/rapidsai/cudf/pull/18642)) [@davidwendt](https://github.com/davidwendt) -- Skip all known pytest failures in pandas-tests ([#18641](https://github.com/rapidsai/cudf/pull/18641)) [@galipremsagar](https://github.com/galipremsagar) -- Preserve partitioning after ``Filter`` and ``Projection`` in cudf-polars ([#18638](https://github.com/rapidsai/cudf/pull/18638)) [@rjzamora](https://github.com/rjzamora) -- Support quantile in cudf-polars grouped aggregations ([#18634](https://github.com/rapidsai/cudf/pull/18634)) [@wence-](https://github.com/wence-) -- Deprecate Series.nullmask, Series.nullable, Series.from_categorical, Series.from_masked_array, cudf.isclose ([#18631](https://github.com/rapidsai/cudf/pull/18631)) [@mroeschke](https://github.com/mroeschke) -- Access private objects by importing from module instead of `cudf.core/util` namespace ([#18629](https://github.com/rapidsai/cudf/pull/18629)) [@mroeschke](https://github.com/mroeschke) -- Replace unnecessary cudf::size_of() calls with sizeof() ([#18628](https://github.com/rapidsai/cudf/pull/18628)) [@davidwendt](https://github.com/davidwendt) -- Improve cold cache dropping ([#18626](https://github.com/rapidsai/cudf/pull/18626)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Improve default config values for cudf-polars streaming ([#18623](https://github.com/rapidsai/cudf/pull/18623)) [@rjzamora](https://github.com/rjzamora) -- Add gtest error check for nvtext::wordpiece_tokenize ([#18621](https://github.com/rapidsai/cudf/pull/18621)) [@davidwendt](https://github.com/davidwendt) -- Polars dataframe serialize using chunked pack ([#18614](https://github.com/rapidsai/cudf/pull/18614)) [@madsbk](https://github.com/madsbk) -- xfail all known errors in pandas-test suite ([#18612](https://github.com/rapidsai/cudf/pull/18612)) [@galipremsagar](https://github.com/galipremsagar) -- Add `TemporalBaseColumn` as a parent class to `DatetimeColumn` and `TimedeltaColumn` ([#18611](https://github.com/rapidsai/cudf/pull/18611)) [@mroeschke](https://github.com/mroeschke) -- Update cudf::cast internal function to use sizeof instead of cudf::size_of ([#18607](https://github.com/rapidsai/cudf/pull/18607)) [@davidwendt](https://github.com/davidwendt) -- Move cudf/utils/utils.py methods to appropriate locations ([#18605](https://github.com/rapidsai/cudf/pull/18605)) [@mroeschke](https://github.com/mroeschke) -- pylibcudf.Column: add `device_buffer_size` and register a dask.sizeof function for cudf-polars Column and DataFrame ([#18602](https://github.com/rapidsai/cudf/pull/18602)) [@madsbk](https://github.com/madsbk) -- Use `cached_property` for Datetime and Timedelta column properties ([#18601](https://github.com/rapidsai/cudf/pull/18601)) [@mroeschke](https://github.com/mroeschke) -- Annotate and simplify `from_arrow` ([#18600](https://github.com/rapidsai/cudf/pull/18600)) [@mroeschke](https://github.com/mroeschke) -- Enable reporting peak memory usage for gtests ([#18599](https://github.com/rapidsai/cudf/pull/18599)) [@davidwendt](https://github.com/davidwendt) -- Prune methods from Frame that are specific to subclasses ([#18597](https://github.com/rapidsai/cudf/pull/18597)) [@mroeschke](https://github.com/mroeschke) -- Switch `tensorflow` integration tests to use 12.x ([#18596](https://github.com/rapidsai/cudf/pull/18596)) [@galipremsagar](https://github.com/galipremsagar) -- refactor: use `libnvcomp` from `libkvikio` wheel to unblock Python 3.13 upgrade ([#18593](https://github.com/rapidsai/cudf/pull/18593)) [@gforsyth](https://github.com/gforsyth) -- Add temporary pdsh benchmarks to ``cudf_polars.experimental`` ([#18592](https://github.com/rapidsai/cudf/pull/18592)) [@rjzamora](https://github.com/rjzamora) -- Update `numba-cuda` dependency to `>=0.9.0` ([#18591](https://github.com/rapidsai/cudf/pull/18591)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- use 'certifi' certificates in fetch_pandas_versions script ([#18588](https://github.com/rapidsai/cudf/pull/18588)) [@jameslamb](https://github.com/jameslamb) -- Add nvtext substring duplication APIs (Part 1) ([#18585](https://github.com/rapidsai/cudf/pull/18585)) [@davidwendt](https://github.com/davidwendt) -- Bump polars version to <1.29 ([#18581](https://github.com/rapidsai/cudf/pull/18581)) [@Matt711](https://github.com/Matt711) -- Allow datetime.timedelta objects in pylibcudf.Scalar.from_py ([#18577](https://github.com/rapidsai/cudf/pull/18577)) [@mroeschke](https://github.com/mroeschke) -- Rework strings split_helper utility for better reuse ([#18575](https://github.com/rapidsai/cudf/pull/18575)) [@davidwendt](https://github.com/davidwendt) -- Additional tests strings for strings split APIs ([#18574](https://github.com/rapidsai/cudf/pull/18574)) [@davidwendt](https://github.com/davidwendt) -- Support datetime.datetime objects in pylibcudf.Scalar.from_py ([#18572](https://github.com/rapidsai/cudf/pull/18572)) [@mroeschke](https://github.com/mroeschke) -- Store Python scalars instead of PyArrow Scalars in cudf_polars Literal expr ([#18563](https://github.com/rapidsai/cudf/pull/18563)) [@mroeschke](https://github.com/mroeschke) -- Support `plc.Scalar.from_py(None)` and `plc.Scalar.from_py(int, float type)` ([#18559](https://github.com/rapidsai/cudf/pull/18559)) [@mroeschke](https://github.com/mroeschke) -- Add xfail window function tests for cudf_polars ([#18557](https://github.com/rapidsai/cudf/pull/18557)) [@btepera](https://github.com/btepera) -- Add fast paths to `Series.to_cupy` and `Series.values` ([#18555](https://github.com/rapidsai/cudf/pull/18555)) [@Matt711](https://github.com/Matt711) -- Reduce cudf-polars pyarrow usage ([#18554](https://github.com/rapidsai/cudf/pull/18554)) [@vyasr](https://github.com/vyasr) -- Avoid possible invalid kernel grid error in `cudf::set_null_masks` if no bitmasks to set ([#18553](https://github.com/rapidsai/cudf/pull/18553)) [@mhaseeb123](https://github.com/mhaseeb123) -- Adjust cudf Python groupby test for cuCollections update ([#18550](https://github.com/rapidsai/cudf/pull/18550)) [@mroeschke](https://github.com/mroeschke) -- Refactor scan test I/O logic into shared `make_partitioned_source` helper ([#18542](https://github.com/rapidsai/cudf/pull/18542)) [@Matt711](https://github.com/Matt711) -- Download build artifacts from Github for CI jobs ([#18539](https://github.com/rapidsai/cudf/pull/18539)) [@VenkateshJaya](https://github.com/VenkateshJaya) -- Update hypothesis version ([#18537](https://github.com/rapidsai/cudf/pull/18537)) [@galipremsagar](https://github.com/galipremsagar) -- Make Python testing dependencies more specific to pylibcudf vs cudf ([#18535](https://github.com/rapidsai/cudf/pull/18535)) [@mroeschke](https://github.com/mroeschke) -- Pin hypothesis<6.131.1 due to performance issues ([#18532](https://github.com/rapidsai/cudf/pull/18532)) [@mroeschke](https://github.com/mroeschke) -- Deduplicate parquet physical type enums ([#18526](https://github.com/rapidsai/cudf/pull/18526)) [@mhaseeb123](https://github.com/mhaseeb123) -- Reduce the number of miscellaenous pandas unit tests run with cudf.pandas ([#18524](https://github.com/rapidsai/cudf/pull/18524)) [@mroeschke](https://github.com/mroeschke) -- Improve nvtext::tokenize_with_vocabulary performance ([#18522](https://github.com/rapidsai/cudf/pull/18522)) [@davidwendt](https://github.com/davidwendt) -- Make pylibcudf.Column.from_rmm_buffer a Python staticmethod ([#18521](https://github.com/rapidsai/cudf/pull/18521)) [@mroeschke](https://github.com/mroeschke) -- Add more short circuit checks for .equals ([#18520](https://github.com/rapidsai/cudf/pull/18520)) [@mroeschke](https://github.com/mroeschke) -- Add synchronous task scheduler to cudf-polars ([#18519](https://github.com/rapidsai/cudf/pull/18519)) [@rjzamora](https://github.com/rjzamora) -- Don't fetch dlpack headers when building cuDF Python ([#18518](https://github.com/rapidsai/cudf/pull/18518)) [@mroeschke](https://github.com/mroeschke) -- Refactor polars configuration ([#18516](https://github.com/rapidsai/cudf/pull/18516)) [@TomAugspurger](https://github.com/TomAugspurger) -- Refactor internal strings utility to separate header and definition file ([#18514](https://github.com/rapidsai/cudf/pull/18514)) [@davidwendt](https://github.com/davidwendt) -- Fix `print()` keyword argument in cudf pandas test ([#18513](https://github.com/rapidsai/cudf/pull/18513)) [@trxcllnt](https://github.com/trxcllnt) -- Improve performance of strings split-record on whitespace ([#18510](https://github.com/rapidsai/cudf/pull/18510)) [@davidwendt](https://github.com/davidwendt) -- Use `cuda::std::iter_value_t` instead of thrust iterator traits ([#18509](https://github.com/rapidsai/cudf/pull/18509)) [@miscco](https://github.com/miscco) -- Remove redundant task-graph logic for streaming ``GroupBy`` ([#18507](https://github.com/rapidsai/cudf/pull/18507)) [@rjzamora](https://github.com/rjzamora) -- Replace `GPU_ARCHS` build variable by `CMAKE_CUDA_ARCHITECTURES` ([#18506](https://github.com/rapidsai/cudf/pull/18506)) [@ttnghia](https://github.com/ttnghia) -- Optimize pandas metadata generation to reduce memory pressure ([#18505](https://github.com/rapidsai/cudf/pull/18505)) [@galipremsagar](https://github.com/galipremsagar) -- Replace deprecated host_buffer in favor of host_span in SourceInfo ([#18503](https://github.com/rapidsai/cudf/pull/18503)) [@Matt711](https://github.com/Matt711) -- Add pylibcudf.Column.from_rmm_buffer ([#18502](https://github.com/rapidsai/cudf/pull/18502)) [@mroeschke](https://github.com/mroeschke) -- Replace thrust functors with libcu++ ones ([#18500](https://github.com/rapidsai/cudf/pull/18500)) [@miscco](https://github.com/miscco) -- Rename cudf-polars executors ([#18499](https://github.com/rapidsai/cudf/pull/18499)) [@rjzamora](https://github.com/rjzamora) -- Remove casting functions in pylibcudf utils ([#18497](https://github.com/rapidsai/cudf/pull/18497)) [@Matt711](https://github.com/Matt711) -- Increase wheel size limit. ([#18487](https://github.com/rapidsai/cudf/pull/18487)) [@bdice](https://github.com/bdice) -- Add CategoricalIndex.from_codes ([#18485](https://github.com/rapidsai/cudf/pull/18485)) [@mroeschke](https://github.com/mroeschke) -- Split join header ([#18484](https://github.com/rapidsai/cudf/pull/18484)) [@shrshi](https://github.com/shrshi) -- Fix unspecified behavior involving move semantics and order of evaluation ([#18481](https://github.com/rapidsai/cudf/pull/18481)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Remove need for to_cudf_compatible_scalar ([#18477](https://github.com/rapidsai/cudf/pull/18477)) [@mroeschke](https://github.com/mroeschke) -- Rerun flaky pytests in CI ([#18476](https://github.com/rapidsai/cudf/pull/18476)) [@galipremsagar](https://github.com/galipremsagar) -- Vendor RAPIDS.cmake ([#18473](https://github.com/rapidsai/cudf/pull/18473)) [@bdice](https://github.com/bdice) -- Add ARM conda environments. ([#18470](https://github.com/rapidsai/cudf/pull/18470)) [@bdice](https://github.com/bdice) -- Bump polars version to <1.28 ([#18469](https://github.com/rapidsai/cudf/pull/18469)) [@Matt711](https://github.com/Matt711) -- Add sink support in cudf_polars ([#18468](https://github.com/rapidsai/cudf/pull/18468)) [@mroeschke](https://github.com/mroeschke) -- Enable rapidsmpf spilling in cudf-polars ([#18461](https://github.com/rapidsai/cudf/pull/18461)) [@madsbk](https://github.com/madsbk) -- Promote Parquet type enums to enum classes ([#18441](https://github.com/rapidsai/cudf/pull/18441)) [@mhaseeb123](https://github.com/mhaseeb123) -- Consolidate logic in DataFrame.__init__ for listlike arguments ([#18439](https://github.com/rapidsai/cudf/pull/18439)) [@mroeschke](https://github.com/mroeschke) -- Update compression formats supported in JSON reader ([#18438](https://github.com/rapidsai/cudf/pull/18438)) [@shrshi](https://github.com/shrshi) -- Disabled Jitify Minification ([#18436](https://github.com/rapidsai/cudf/pull/18436)) [@lamarrr](https://github.com/lamarrr) -- Fix printing decimal128 types that are zero ([#18435](https://github.com/rapidsai/cudf/pull/18435)) [@trxcllnt](https://github.com/trxcllnt) -- Replace direct use of nvCOMP and of its adapter with the higher-level decompression API ([#18434](https://github.com/rapidsai/cudf/pull/18434)) [@vuule](https://github.com/vuule) -- Add more `cudf.DataFrame` constructor pytest benchmarks ([#18433](https://github.com/rapidsai/cudf/pull/18433)) [@mroeschke](https://github.com/mroeschke) -- Test against stable tags for narwhals ([#18431](https://github.com/rapidsai/cudf/pull/18431)) [@Matt711](https://github.com/Matt711) -- Refcount-based dropping of cached evaluations in cudf-polars executor ([#18430](https://github.com/rapidsai/cudf/pull/18430)) [@wence-](https://github.com/wence-) -- Replace `Thrust` iterator facilities with libcu++ ones ([#18427](https://github.com/rapidsai/cudf/pull/18427)) [@miscco](https://github.com/miscco) -- Remove numpy requirement when converting 2d cuda array interface objects to pylibcudf Columns ([#18426](https://github.com/rapidsai/cudf/pull/18426)) [@Matt711](https://github.com/Matt711) -- Share more cudf.Column methods for `indices_of`/`isin` ([#18423](https://github.com/rapidsai/cudf/pull/18423)) [@mroeschke](https://github.com/mroeschke) -- Switch the ptr type in gpumemoryview from Py_ssize_t to uintptr_t ([#18419](https://github.com/rapidsai/cudf/pull/18419)) [@Matt711](https://github.com/Matt711) -- Add strings::extract_single API ([#18417](https://github.com/rapidsai/cudf/pull/18417)) [@davidwendt](https://github.com/davidwendt) -- Add to_arrow_host_stringview interop API ([#18416](https://github.com/rapidsai/cudf/pull/18416)) [@davidwendt](https://github.com/davidwendt) -- Start removal of vector factories with `_sync` suffix by deprecating them and adding versions without the suffix ([#18414](https://github.com/rapidsai/cudf/pull/18414)) [@vuule](https://github.com/vuule) -- Allow polars arrow conversion to produce string_view ([#18413](https://github.com/rapidsai/cudf/pull/18413)) [@wence-](https://github.com/wence-) -- Change `dask_cudf.to_parquet` behavior for local filesystems ([#18408](https://github.com/rapidsai/cudf/pull/18408)) [@rjzamora](https://github.com/rjzamora) -- Add rank and label_bin methods to ColumnBase ([#18407](https://github.com/rapidsai/cudf/pull/18407)) [@mroeschke](https://github.com/mroeschke) -- Improve performance of strings::like for long strings ([#18406](https://github.com/rapidsai/cudf/pull/18406)) [@davidwendt](https://github.com/davidwendt) -- Automatic single-partition fallback in cudf-polars ([#18405](https://github.com/rapidsai/cudf/pull/18405)) [@rjzamora](https://github.com/rjzamora) -- Remove `_sync` suffix from hostdevice types ([#18404](https://github.com/rapidsai/cudf/pull/18404)) [@vuule](https://github.com/vuule) -- Use owning Arrow types in C++ to expose data to Python ([#18402](https://github.com/rapidsai/cudf/pull/18402)) [@vyasr](https://github.com/vyasr) -- add static push and pop methods to NvtxRange ([#18401](https://github.com/rapidsai/cudf/pull/18401)) [@zpuller](https://github.com/zpuller) -- Deprecate cudf.Scalar ([#18394](https://github.com/rapidsai/cudf/pull/18394)) [@mroeschke](https://github.com/mroeschke) -- Bump polars version to <1.27 ([#18387](https://github.com/rapidsai/cudf/pull/18387)) [@Matt711](https://github.com/Matt711) -- Branch 25.06 merge 25.04 ([#18380](https://github.com/rapidsai/cudf/pull/18380)) [@Matt711](https://github.com/Matt711) -- Silence warning by setting BUILD_SHARED_LIBS ([#18371](https://github.com/rapidsai/cudf/pull/18371)) [@vyasr](https://github.com/vyasr) -- Rewrite groupby aggregations in cudf-polars to simplify evaluation ([#18369](https://github.com/rapidsai/cudf/pull/18369)) [@wence-](https://github.com/wence-) -- Pass stream through when taking ownership from libcudf ([#18367](https://github.com/rapidsai/cudf/pull/18367)) [@wence-](https://github.com/wence-) -- Expose new grouped_range_rolling API in pylibcudf ([#18365](https://github.com/rapidsai/cudf/pull/18365)) [@wence-](https://github.com/wence-) -- Avoid patching sort algorithms from CCCL ([#18364](https://github.com/rapidsai/cudf/pull/18364)) [@miscco](https://github.com/miscco) -- Deprecate old nvtext::normalize_characters ([#18360](https://github.com/rapidsai/cudf/pull/18360)) [@davidwendt](https://github.com/davidwendt) -- refactor(rattler): enable strict channel priority for builds ([#18358](https://github.com/rapidsai/cudf/pull/18358)) [@gforsyth](https://github.com/gforsyth) -- Optimize `sequences` by introducing `make_offsets_child_column` ([#18357](https://github.com/rapidsai/cudf/pull/18357)) [@ustcfy](https://github.com/ustcfy) -- Decompress all data in a single `decompress_page_data` when reading Parquet input in a single chunk ([#18352](https://github.com/rapidsai/cudf/pull/18352)) [@vuule](https://github.com/vuule) -- Moving wheel builds to specified location and uploading build artifacts to Github ([#18346](https://github.com/rapidsai/cudf/pull/18346)) [@VenkateshJaya](https://github.com/VenkateshJaya) -- Performance improvement for to_lower/to_upper for multi-byte UTF-8 characters ([#18345](https://github.com/rapidsai/cudf/pull/18345)) [@davidwendt](https://github.com/davidwendt) -- Branch 25.06 merge branch 25.04 ([#18344](https://github.com/rapidsai/cudf/pull/18344)) [@vyasr](https://github.com/vyasr) -- Use dask-cuda for cudf-polars experimental testing ([#18343](https://github.com/rapidsai/cudf/pull/18343)) [@rjzamora](https://github.com/rjzamora) -- Deprecate nvtext subword tokenizer ([#18334](https://github.com/rapidsai/cudf/pull/18334)) [@davidwendt](https://github.com/davidwendt) -- Remove cudf.Scalar in as_column ([#18331](https://github.com/rapidsai/cudf/pull/18331)) [@mroeschke](https://github.com/mroeschke) -- Add tests for `cudf.polars` to be able to work on a cpu-only machine ([#18327](https://github.com/rapidsai/cudf/pull/18327)) [@galipremsagar](https://github.com/galipremsagar) -- Allow `cudf.DataFrame.from_pylibcudf` to accept a `pylibcudf.io.TableWithMetadata` ([#18319](https://github.com/rapidsai/cudf/pull/18319)) [@mroeschke](https://github.com/mroeschke) -- Avoid stateful construction in `DataFrame.__init__` ([#18306](https://github.com/rapidsai/cudf/pull/18306)) [@mroeschke](https://github.com/mroeschke) -- Improve the groupby performance for extremely low cardinality ([#18290](https://github.com/rapidsai/cudf/pull/18290)) [@PointKernel](https://github.com/PointKernel) -- Remove extranous modules from top level cudf namespace ([#18287](https://github.com/rapidsai/cudf/pull/18287)) [@mroeschke](https://github.com/mroeschke) -- Require type annotations in cudf.polars ([#18285](https://github.com/rapidsai/cudf/pull/18285)) [@TomAugspurger](https://github.com/TomAugspurger) -- Removing unnecessary StreamSynchronization in reading ([#18279](https://github.com/rapidsai/cudf/pull/18279)) [@JigaoLuo](https://github.com/JigaoLuo) -- Update to CCCL 2.8.x with no CCCL patches ([#18235](https://github.com/rapidsai/cudf/pull/18235)) [@bdice](https://github.com/bdice) -- Reduce register pressure for compute_column_kernel ([#18226](https://github.com/rapidsai/cudf/pull/18226)) [@matal-nvidia](https://github.com/matal-nvidia) -- Use the mapped buffer for all read operations in the memory-mapped source; switch default source to the kvikIO one ([#18204](https://github.com/rapidsai/cudf/pull/18204)) [@vuule](https://github.com/vuule) -- Improve test coverage in the catboost integration tests ([#18126](https://github.com/rapidsai/cudf/pull/18126)) [@Matt711](https://github.com/Matt711) -- Create file sources in parallel ([#18094](https://github.com/rapidsai/cudf/pull/18094)) [@vuule](https://github.com/vuule) -- Enable `stumpy_distributed` tests ([#17969](https://github.com/rapidsai/cudf/pull/17969)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor distinct join to use primitive row operators when proper ([#17726](https://github.com/rapidsai/cudf/pull/17726)) [@PointKernel](https://github.com/PointKernel) -- Update chunked parquet reader benchmarks ([#16543](https://github.com/rapidsai/cudf/pull/16543)) [@sdrp713](https://github.com/sdrp713) +- use 'rapids-init-pip' in wheel CI, other CI changes ([#18902](https://github.com/NVIDIA/cudf/pull/18902)) [@jameslamb](https://github.com/jameslamb) +- Avoid RecursionError in custreamz test ([#18887](https://github.com/NVIDIA/cudf/pull/18887)) [@TomAugspurger](https://github.com/TomAugspurger) +- Update NumPy dependency in cudf.pandas-catboost integration test ([#18870](https://github.com/NVIDIA/cudf/pull/18870)) [@Matt711](https://github.com/Matt711) +- CPU only execution for PDSH ([#18869](https://github.com/NVIDIA/cudf/pull/18869)) [@quasiben](https://github.com/quasiben) +- Remove more top level cudf imports in core ([#18862](https://github.com/NVIDIA/cudf/pull/18862)) [@mroeschke](https://github.com/mroeschke) +- Remove top level cudf imports in core ([#18857](https://github.com/NVIDIA/cudf/pull/18857)) [@mroeschke](https://github.com/mroeschke) +- Add CUDF_INSTALL_DIR for JAVA build script ([#18852](https://github.com/NVIDIA/cudf/pull/18852)) [@pxLi](https://github.com/pxLi) +- Call the correct `from_pandas` in `hdf` reader ([#18850](https://github.com/NVIDIA/cudf/pull/18850)) [@galipremsagar](https://github.com/galipremsagar) +- Update `__all__` in `cudf_polars/dsl/ir.py` ([#18848](https://github.com/NVIDIA/cudf/pull/18848)) [@Matt711](https://github.com/Matt711) +- Upload examples conda package ([#18847](https://github.com/NVIDIA/cudf/pull/18847)) [@vyasr](https://github.com/vyasr) +- Add retries to prevent failures in occasionally slow CI runs ([#18843](https://github.com/NVIDIA/cudf/pull/18843)) [@galipremsagar](https://github.com/galipremsagar) +- Finish CUDA 12.9 migration and use branch-25.06 workflows ([#18839](https://github.com/NVIDIA/cudf/pull/18839)) [@bdice](https://github.com/bdice) +- Remove toplevel `import cudf` from window/tools/join directories ([#18833](https://github.com/NVIDIA/cudf/pull/18833)) [@mroeschke](https://github.com/mroeschke) +- Remove toplevel `import cudf` from cudf/io files ([#18829](https://github.com/NVIDIA/cudf/pull/18829)) [@mroeschke](https://github.com/mroeschke) +- Update pdsh benchmark script to support explain-only ([#18826](https://github.com/NVIDIA/cudf/pull/18826)) [@TomAugspurger](https://github.com/TomAugspurger) +- Refactor UDF utils and add a hook to enable NRT when necessary ([#18823](https://github.com/NVIDIA/cudf/pull/18823)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix memory access error in nvtext::edit_distance ([#18821](https://github.com/NVIDIA/cudf/pull/18821)) [@davidwendt](https://github.com/davidwendt) +- Update to clang 20 ([#18818](https://github.com/NVIDIA/cudf/pull/18818)) [@bdice](https://github.com/bdice) +- Reduce more data sizes of Python tests ([#18814](https://github.com/NVIDIA/cudf/pull/18814)) [@mroeschke](https://github.com/mroeschke) +- Mark DataFrame.dtypes as an _external_only_api ([#18809](https://github.com/NVIDIA/cudf/pull/18809)) [@mroeschke](https://github.com/mroeschke) +- Change calls to thrust::swap to cuda::std::swap ([#18808](https://github.com/NVIDIA/cudf/pull/18808)) [@davidwendt](https://github.com/davidwendt) +- Move implemented BaseIndex methods over to Index ([#18807](https://github.com/NVIDIA/cudf/pull/18807)) [@mroeschke](https://github.com/mroeschke) +- Improve pandas version fetching script ([#18793](https://github.com/NVIDIA/cudf/pull/18793)) [@galipremsagar](https://github.com/galipremsagar) +- Change cudf::sort googlebench benchmarks to nvbench ([#18786](https://github.com/NVIDIA/cudf/pull/18786)) [@davidwendt](https://github.com/davidwendt) +- Only warn in cudf.pandas if rmm mode explicitly set and rmm already configured ([#18785](https://github.com/NVIDIA/cudf/pull/18785)) [@jcrist](https://github.com/jcrist) +- Quote head_rev in conda recipes ([#18784](https://github.com/NVIDIA/cudf/pull/18784)) [@bdice](https://github.com/bdice) +- Move RangeIndex implementation below Index ([#18777](https://github.com/NVIDIA/cudf/pull/18777)) [@mroeschke](https://github.com/mroeschke) +- Remove unecessary _Ravelled class ([#18771](https://github.com/NVIDIA/cudf/pull/18771)) [@Matt711](https://github.com/Matt711) +- Remove pytest-rerunfailures ([#18766](https://github.com/NVIDIA/cudf/pull/18766)) [@mroeschke](https://github.com/mroeschke) +- Replace from_arrow with direct calls Column/Table constructors in pylibcudf and cudf-polars tests ([#18762](https://github.com/NVIDIA/cudf/pull/18762)) [@Matt711](https://github.com/Matt711) +- CUDA 12.9 use updated compression flags ([#18755](https://github.com/NVIDIA/cudf/pull/18755)) [@robertmaynard](https://github.com/robertmaynard) +- fix(rattler): add `librmm` to host for `libcudf` to fix overlinking error ([#18754](https://github.com/NVIDIA/cudf/pull/18754)) [@gforsyth](https://github.com/gforsyth) +- Remove the file name from the output in cudf-polars' explain APIs ([#18752](https://github.com/NVIDIA/cudf/pull/18752)) [@Matt711](https://github.com/Matt711) +- Remove cudf.BaseIndex ([#18751](https://github.com/NVIDIA/cudf/pull/18751)) [@mroeschke](https://github.com/mroeschke) +- Support creating a pylibcudf Column from a general ndarray ([#18744](https://github.com/NVIDIA/cudf/pull/18744)) [@Matt711](https://github.com/Matt711) +- Improve lowering of ``Distinct`` IR nodes for high-cardinality data ([#18725](https://github.com/NVIDIA/cudf/pull/18725)) [@rjzamora](https://github.com/rjzamora) +- Simplify Numba-CUDA MVC logic ([#18724](https://github.com/NVIDIA/cudf/pull/18724)) [@bdice](https://github.com/bdice) +- Test with CUDA 12.9.0 ([#18721](https://github.com/NVIDIA/cudf/pull/18721)) [@bdice](https://github.com/bdice) +- Add more `cudf.Series` microbenchmarks ([#18718](https://github.com/NVIDIA/cudf/pull/18718)) [@Matt711](https://github.com/Matt711) +- Run unit-tests-cudf-pandas on branch-25.06 for nightly tests ([#18717](https://github.com/NVIDIA/cudf/pull/18717)) [@davidwendt](https://github.com/davidwendt) +- Move `test_large_unique_categories_repr` to benchmarks ([#18715](https://github.com/NVIDIA/cudf/pull/18715)) [@galipremsagar](https://github.com/galipremsagar) +- Allow `pylibcudf.Column` to consume objects exposing `__arrow_c_stream__` ([#18712](https://github.com/NVIDIA/cudf/pull/18712)) [@mroeschke](https://github.com/mroeschke) +- Switch from printing to logging ([#18711](https://github.com/NVIDIA/cudf/pull/18711)) [@vyasr](https://github.com/vyasr) +- Add Python tests for different compression implementations ([#18710](https://github.com/NVIDIA/cudf/pull/18710)) [@vuule](https://github.com/vuule) +- Remove redundant xfails in cuml integration tests ([#18699](https://github.com/NVIDIA/cudf/pull/18699)) [@Matt711](https://github.com/Matt711) +- ci: run unit-tests-cudf-pandas on `branch-25.06` workflow ([#18692](https://github.com/NVIDIA/cudf/pull/18692)) [@gforsyth](https://github.com/gforsyth) +- Exclude librmm.so from auditwheel ([#18691](https://github.com/NVIDIA/cudf/pull/18691)) [@bdice](https://github.com/bdice) +- Add C++ tests for different compression implementations ([#18690](https://github.com/NVIDIA/cudf/pull/18690)) [@vuule](https://github.com/vuule) +- Improve runtime of cuDF Python unit tests ([#18689](https://github.com/NVIDIA/cudf/pull/18689)) [@mroeschke](https://github.com/mroeschke) +- Require at least numba-cuda `0.10.1` ([#18688](https://github.com/NVIDIA/cudf/pull/18688)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add `nvidia-cuda-{nvrtc, nvcc}` as a dependency for cuDF wheels ([#18686](https://github.com/NVIDIA/cudf/pull/18686)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support rolling aggregations in in-memory cudf-polars execution ([#18681](https://github.com/NVIDIA/cudf/pull/18681)) [@wence-](https://github.com/wence-) +- Replace ``parquet_blocksize`` with ``target_partition_size`` ([#18669](https://github.com/NVIDIA/cudf/pull/18669)) [@rjzamora](https://github.com/rjzamora) +- Skip test_large_unique_categories_repr in CI ([#18666](https://github.com/NVIDIA/cudf/pull/18666)) [@bdice](https://github.com/bdice) +- Locally import pyarrow.dataset and fsspec for `import cudf` performance ([#18663](https://github.com/NVIDIA/cudf/pull/18663)) [@mroeschke](https://github.com/mroeschke) +- Disable `arm64` python tests ([#18662](https://github.com/NVIDIA/cudf/pull/18662)) [@galipremsagar](https://github.com/galipremsagar) +- Pin numba-cuda>=0.9.0,!=0.10.0 due to CI hangs on ARM ([#18661](https://github.com/NVIDIA/cudf/pull/18661)) [@mroeschke](https://github.com/mroeschke) +- Fix compile warnings in Java JNI ([#18660](https://github.com/NVIDIA/cudf/pull/18660)) [@ttnghia](https://github.com/ttnghia) +- Drop ``Empty`` nodes from IR graph ([#18658](https://github.com/NVIDIA/cudf/pull/18658)) [@rjzamora](https://github.com/rjzamora) +- Add support for Python 3.13 ([#18648](https://github.com/NVIDIA/cudf/pull/18648)) [@gforsyth](https://github.com/gforsyth) +- Cleanup libcudf detail/aggregation.hpp/.cuh ([#18642](https://github.com/NVIDIA/cudf/pull/18642)) [@davidwendt](https://github.com/davidwendt) +- Skip all known pytest failures in pandas-tests ([#18641](https://github.com/NVIDIA/cudf/pull/18641)) [@galipremsagar](https://github.com/galipremsagar) +- Preserve partitioning after ``Filter`` and ``Projection`` in cudf-polars ([#18638](https://github.com/NVIDIA/cudf/pull/18638)) [@rjzamora](https://github.com/rjzamora) +- Support quantile in cudf-polars grouped aggregations ([#18634](https://github.com/NVIDIA/cudf/pull/18634)) [@wence-](https://github.com/wence-) +- Deprecate Series.nullmask, Series.nullable, Series.from_categorical, Series.from_masked_array, cudf.isclose ([#18631](https://github.com/NVIDIA/cudf/pull/18631)) [@mroeschke](https://github.com/mroeschke) +- Access private objects by importing from module instead of `cudf.core/util` namespace ([#18629](https://github.com/NVIDIA/cudf/pull/18629)) [@mroeschke](https://github.com/mroeschke) +- Replace unnecessary cudf::size_of() calls with sizeof() ([#18628](https://github.com/NVIDIA/cudf/pull/18628)) [@davidwendt](https://github.com/davidwendt) +- Improve cold cache dropping ([#18626](https://github.com/NVIDIA/cudf/pull/18626)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Improve default config values for cudf-polars streaming ([#18623](https://github.com/NVIDIA/cudf/pull/18623)) [@rjzamora](https://github.com/rjzamora) +- Add gtest error check for nvtext::wordpiece_tokenize ([#18621](https://github.com/NVIDIA/cudf/pull/18621)) [@davidwendt](https://github.com/davidwendt) +- Polars dataframe serialize using chunked pack ([#18614](https://github.com/NVIDIA/cudf/pull/18614)) [@madsbk](https://github.com/madsbk) +- xfail all known errors in pandas-test suite ([#18612](https://github.com/NVIDIA/cudf/pull/18612)) [@galipremsagar](https://github.com/galipremsagar) +- Add `TemporalBaseColumn` as a parent class to `DatetimeColumn` and `TimedeltaColumn` ([#18611](https://github.com/NVIDIA/cudf/pull/18611)) [@mroeschke](https://github.com/mroeschke) +- Update cudf::cast internal function to use sizeof instead of cudf::size_of ([#18607](https://github.com/NVIDIA/cudf/pull/18607)) [@davidwendt](https://github.com/davidwendt) +- Move cudf/utils/utils.py methods to appropriate locations ([#18605](https://github.com/NVIDIA/cudf/pull/18605)) [@mroeschke](https://github.com/mroeschke) +- pylibcudf.Column: add `device_buffer_size` and register a dask.sizeof function for cudf-polars Column and DataFrame ([#18602](https://github.com/NVIDIA/cudf/pull/18602)) [@madsbk](https://github.com/madsbk) +- Use `cached_property` for Datetime and Timedelta column properties ([#18601](https://github.com/NVIDIA/cudf/pull/18601)) [@mroeschke](https://github.com/mroeschke) +- Annotate and simplify `from_arrow` ([#18600](https://github.com/NVIDIA/cudf/pull/18600)) [@mroeschke](https://github.com/mroeschke) +- Enable reporting peak memory usage for gtests ([#18599](https://github.com/NVIDIA/cudf/pull/18599)) [@davidwendt](https://github.com/davidwendt) +- Prune methods from Frame that are specific to subclasses ([#18597](https://github.com/NVIDIA/cudf/pull/18597)) [@mroeschke](https://github.com/mroeschke) +- Switch `tensorflow` integration tests to use 12.x ([#18596](https://github.com/NVIDIA/cudf/pull/18596)) [@galipremsagar](https://github.com/galipremsagar) +- refactor: use `libnvcomp` from `libkvikio` wheel to unblock Python 3.13 upgrade ([#18593](https://github.com/NVIDIA/cudf/pull/18593)) [@gforsyth](https://github.com/gforsyth) +- Add temporary pdsh benchmarks to ``cudf_polars.experimental`` ([#18592](https://github.com/NVIDIA/cudf/pull/18592)) [@rjzamora](https://github.com/rjzamora) +- Update `numba-cuda` dependency to `>=0.9.0` ([#18591](https://github.com/NVIDIA/cudf/pull/18591)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- use 'certifi' certificates in fetch_pandas_versions script ([#18588](https://github.com/NVIDIA/cudf/pull/18588)) [@jameslamb](https://github.com/jameslamb) +- Add nvtext substring duplication APIs (Part 1) ([#18585](https://github.com/NVIDIA/cudf/pull/18585)) [@davidwendt](https://github.com/davidwendt) +- Bump polars version to <1.29 ([#18581](https://github.com/NVIDIA/cudf/pull/18581)) [@Matt711](https://github.com/Matt711) +- Allow datetime.timedelta objects in pylibcudf.Scalar.from_py ([#18577](https://github.com/NVIDIA/cudf/pull/18577)) [@mroeschke](https://github.com/mroeschke) +- Rework strings split_helper utility for better reuse ([#18575](https://github.com/NVIDIA/cudf/pull/18575)) [@davidwendt](https://github.com/davidwendt) +- Additional tests strings for strings split APIs ([#18574](https://github.com/NVIDIA/cudf/pull/18574)) [@davidwendt](https://github.com/davidwendt) +- Support datetime.datetime objects in pylibcudf.Scalar.from_py ([#18572](https://github.com/NVIDIA/cudf/pull/18572)) [@mroeschke](https://github.com/mroeschke) +- Store Python scalars instead of PyArrow Scalars in cudf_polars Literal expr ([#18563](https://github.com/NVIDIA/cudf/pull/18563)) [@mroeschke](https://github.com/mroeschke) +- Support `plc.Scalar.from_py(None)` and `plc.Scalar.from_py(int, float type)` ([#18559](https://github.com/NVIDIA/cudf/pull/18559)) [@mroeschke](https://github.com/mroeschke) +- Add xfail window function tests for cudf_polars ([#18557](https://github.com/NVIDIA/cudf/pull/18557)) [@btepera](https://github.com/btepera) +- Add fast paths to `Series.to_cupy` and `Series.values` ([#18555](https://github.com/NVIDIA/cudf/pull/18555)) [@Matt711](https://github.com/Matt711) +- Reduce cudf-polars pyarrow usage ([#18554](https://github.com/NVIDIA/cudf/pull/18554)) [@vyasr](https://github.com/vyasr) +- Avoid possible invalid kernel grid error in `cudf::set_null_masks` if no bitmasks to set ([#18553](https://github.com/NVIDIA/cudf/pull/18553)) [@mhaseeb123](https://github.com/mhaseeb123) +- Adjust cudf Python groupby test for cuCollections update ([#18550](https://github.com/NVIDIA/cudf/pull/18550)) [@mroeschke](https://github.com/mroeschke) +- Refactor scan test I/O logic into shared `make_partitioned_source` helper ([#18542](https://github.com/NVIDIA/cudf/pull/18542)) [@Matt711](https://github.com/Matt711) +- Download build artifacts from Github for CI jobs ([#18539](https://github.com/NVIDIA/cudf/pull/18539)) [@VenkateshJaya](https://github.com/VenkateshJaya) +- Update hypothesis version ([#18537](https://github.com/NVIDIA/cudf/pull/18537)) [@galipremsagar](https://github.com/galipremsagar) +- Make Python testing dependencies more specific to pylibcudf vs cudf ([#18535](https://github.com/NVIDIA/cudf/pull/18535)) [@mroeschke](https://github.com/mroeschke) +- Pin hypothesis<6.131.1 due to performance issues ([#18532](https://github.com/NVIDIA/cudf/pull/18532)) [@mroeschke](https://github.com/mroeschke) +- Deduplicate parquet physical type enums ([#18526](https://github.com/NVIDIA/cudf/pull/18526)) [@mhaseeb123](https://github.com/mhaseeb123) +- Reduce the number of miscellaenous pandas unit tests run with cudf.pandas ([#18524](https://github.com/NVIDIA/cudf/pull/18524)) [@mroeschke](https://github.com/mroeschke) +- Improve nvtext::tokenize_with_vocabulary performance ([#18522](https://github.com/NVIDIA/cudf/pull/18522)) [@davidwendt](https://github.com/davidwendt) +- Make pylibcudf.Column.from_rmm_buffer a Python staticmethod ([#18521](https://github.com/NVIDIA/cudf/pull/18521)) [@mroeschke](https://github.com/mroeschke) +- Add more short circuit checks for .equals ([#18520](https://github.com/NVIDIA/cudf/pull/18520)) [@mroeschke](https://github.com/mroeschke) +- Add synchronous task scheduler to cudf-polars ([#18519](https://github.com/NVIDIA/cudf/pull/18519)) [@rjzamora](https://github.com/rjzamora) +- Don't fetch dlpack headers when building cuDF Python ([#18518](https://github.com/NVIDIA/cudf/pull/18518)) [@mroeschke](https://github.com/mroeschke) +- Refactor polars configuration ([#18516](https://github.com/NVIDIA/cudf/pull/18516)) [@TomAugspurger](https://github.com/TomAugspurger) +- Refactor internal strings utility to separate header and definition file ([#18514](https://github.com/NVIDIA/cudf/pull/18514)) [@davidwendt](https://github.com/davidwendt) +- Fix `print()` keyword argument in cudf pandas test ([#18513](https://github.com/NVIDIA/cudf/pull/18513)) [@trxcllnt](https://github.com/trxcllnt) +- Improve performance of strings split-record on whitespace ([#18510](https://github.com/NVIDIA/cudf/pull/18510)) [@davidwendt](https://github.com/davidwendt) +- Use `cuda::std::iter_value_t` instead of thrust iterator traits ([#18509](https://github.com/NVIDIA/cudf/pull/18509)) [@miscco](https://github.com/miscco) +- Remove redundant task-graph logic for streaming ``GroupBy`` ([#18507](https://github.com/NVIDIA/cudf/pull/18507)) [@rjzamora](https://github.com/rjzamora) +- Replace `GPU_ARCHS` build variable by `CMAKE_CUDA_ARCHITECTURES` ([#18506](https://github.com/NVIDIA/cudf/pull/18506)) [@ttnghia](https://github.com/ttnghia) +- Optimize pandas metadata generation to reduce memory pressure ([#18505](https://github.com/NVIDIA/cudf/pull/18505)) [@galipremsagar](https://github.com/galipremsagar) +- Replace deprecated host_buffer in favor of host_span in SourceInfo ([#18503](https://github.com/NVIDIA/cudf/pull/18503)) [@Matt711](https://github.com/Matt711) +- Add pylibcudf.Column.from_rmm_buffer ([#18502](https://github.com/NVIDIA/cudf/pull/18502)) [@mroeschke](https://github.com/mroeschke) +- Replace thrust functors with libcu++ ones ([#18500](https://github.com/NVIDIA/cudf/pull/18500)) [@miscco](https://github.com/miscco) +- Rename cudf-polars executors ([#18499](https://github.com/NVIDIA/cudf/pull/18499)) [@rjzamora](https://github.com/rjzamora) +- Remove casting functions in pylibcudf utils ([#18497](https://github.com/NVIDIA/cudf/pull/18497)) [@Matt711](https://github.com/Matt711) +- Increase wheel size limit. ([#18487](https://github.com/NVIDIA/cudf/pull/18487)) [@bdice](https://github.com/bdice) +- Add CategoricalIndex.from_codes ([#18485](https://github.com/NVIDIA/cudf/pull/18485)) [@mroeschke](https://github.com/mroeschke) +- Split join header ([#18484](https://github.com/NVIDIA/cudf/pull/18484)) [@shrshi](https://github.com/shrshi) +- Fix unspecified behavior involving move semantics and order of evaluation ([#18481](https://github.com/NVIDIA/cudf/pull/18481)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Remove need for to_cudf_compatible_scalar ([#18477](https://github.com/NVIDIA/cudf/pull/18477)) [@mroeschke](https://github.com/mroeschke) +- Rerun flaky pytests in CI ([#18476](https://github.com/NVIDIA/cudf/pull/18476)) [@galipremsagar](https://github.com/galipremsagar) +- Vendor RAPIDS.cmake ([#18473](https://github.com/NVIDIA/cudf/pull/18473)) [@bdice](https://github.com/bdice) +- Add ARM conda environments. ([#18470](https://github.com/NVIDIA/cudf/pull/18470)) [@bdice](https://github.com/bdice) +- Bump polars version to <1.28 ([#18469](https://github.com/NVIDIA/cudf/pull/18469)) [@Matt711](https://github.com/Matt711) +- Add sink support in cudf_polars ([#18468](https://github.com/NVIDIA/cudf/pull/18468)) [@mroeschke](https://github.com/mroeschke) +- Enable rapidsmpf spilling in cudf-polars ([#18461](https://github.com/NVIDIA/cudf/pull/18461)) [@madsbk](https://github.com/madsbk) +- Promote Parquet type enums to enum classes ([#18441](https://github.com/NVIDIA/cudf/pull/18441)) [@mhaseeb123](https://github.com/mhaseeb123) +- Consolidate logic in DataFrame.__init__ for listlike arguments ([#18439](https://github.com/NVIDIA/cudf/pull/18439)) [@mroeschke](https://github.com/mroeschke) +- Update compression formats supported in JSON reader ([#18438](https://github.com/NVIDIA/cudf/pull/18438)) [@shrshi](https://github.com/shrshi) +- Disabled Jitify Minification ([#18436](https://github.com/NVIDIA/cudf/pull/18436)) [@lamarrr](https://github.com/lamarrr) +- Fix printing decimal128 types that are zero ([#18435](https://github.com/NVIDIA/cudf/pull/18435)) [@trxcllnt](https://github.com/trxcllnt) +- Replace direct use of nvCOMP and of its adapter with the higher-level decompression API ([#18434](https://github.com/NVIDIA/cudf/pull/18434)) [@vuule](https://github.com/vuule) +- Add more `cudf.DataFrame` constructor pytest benchmarks ([#18433](https://github.com/NVIDIA/cudf/pull/18433)) [@mroeschke](https://github.com/mroeschke) +- Test against stable tags for narwhals ([#18431](https://github.com/NVIDIA/cudf/pull/18431)) [@Matt711](https://github.com/Matt711) +- Refcount-based dropping of cached evaluations in cudf-polars executor ([#18430](https://github.com/NVIDIA/cudf/pull/18430)) [@wence-](https://github.com/wence-) +- Replace `Thrust` iterator facilities with libcu++ ones ([#18427](https://github.com/NVIDIA/cudf/pull/18427)) [@miscco](https://github.com/miscco) +- Remove numpy requirement when converting 2d cuda array interface objects to pylibcudf Columns ([#18426](https://github.com/NVIDIA/cudf/pull/18426)) [@Matt711](https://github.com/Matt711) +- Share more cudf.Column methods for `indices_of`/`isin` ([#18423](https://github.com/NVIDIA/cudf/pull/18423)) [@mroeschke](https://github.com/mroeschke) +- Switch the ptr type in gpumemoryview from Py_ssize_t to uintptr_t ([#18419](https://github.com/NVIDIA/cudf/pull/18419)) [@Matt711](https://github.com/Matt711) +- Add strings::extract_single API ([#18417](https://github.com/NVIDIA/cudf/pull/18417)) [@davidwendt](https://github.com/davidwendt) +- Add to_arrow_host_stringview interop API ([#18416](https://github.com/NVIDIA/cudf/pull/18416)) [@davidwendt](https://github.com/davidwendt) +- Start removal of vector factories with `_sync` suffix by deprecating them and adding versions without the suffix ([#18414](https://github.com/NVIDIA/cudf/pull/18414)) [@vuule](https://github.com/vuule) +- Allow polars arrow conversion to produce string_view ([#18413](https://github.com/NVIDIA/cudf/pull/18413)) [@wence-](https://github.com/wence-) +- Change `dask_cudf.to_parquet` behavior for local filesystems ([#18408](https://github.com/NVIDIA/cudf/pull/18408)) [@rjzamora](https://github.com/rjzamora) +- Add rank and label_bin methods to ColumnBase ([#18407](https://github.com/NVIDIA/cudf/pull/18407)) [@mroeschke](https://github.com/mroeschke) +- Improve performance of strings::like for long strings ([#18406](https://github.com/NVIDIA/cudf/pull/18406)) [@davidwendt](https://github.com/davidwendt) +- Automatic single-partition fallback in cudf-polars ([#18405](https://github.com/NVIDIA/cudf/pull/18405)) [@rjzamora](https://github.com/rjzamora) +- Remove `_sync` suffix from hostdevice types ([#18404](https://github.com/NVIDIA/cudf/pull/18404)) [@vuule](https://github.com/vuule) +- Use owning Arrow types in C++ to expose data to Python ([#18402](https://github.com/NVIDIA/cudf/pull/18402)) [@vyasr](https://github.com/vyasr) +- add static push and pop methods to NvtxRange ([#18401](https://github.com/NVIDIA/cudf/pull/18401)) [@zpuller](https://github.com/zpuller) +- Deprecate cudf.Scalar ([#18394](https://github.com/NVIDIA/cudf/pull/18394)) [@mroeschke](https://github.com/mroeschke) +- Bump polars version to <1.27 ([#18387](https://github.com/NVIDIA/cudf/pull/18387)) [@Matt711](https://github.com/Matt711) +- Branch 25.06 merge 25.04 ([#18380](https://github.com/NVIDIA/cudf/pull/18380)) [@Matt711](https://github.com/Matt711) +- Silence warning by setting BUILD_SHARED_LIBS ([#18371](https://github.com/NVIDIA/cudf/pull/18371)) [@vyasr](https://github.com/vyasr) +- Rewrite groupby aggregations in cudf-polars to simplify evaluation ([#18369](https://github.com/NVIDIA/cudf/pull/18369)) [@wence-](https://github.com/wence-) +- Pass stream through when taking ownership from libcudf ([#18367](https://github.com/NVIDIA/cudf/pull/18367)) [@wence-](https://github.com/wence-) +- Expose new grouped_range_rolling API in pylibcudf ([#18365](https://github.com/NVIDIA/cudf/pull/18365)) [@wence-](https://github.com/wence-) +- Avoid patching sort algorithms from CCCL ([#18364](https://github.com/NVIDIA/cudf/pull/18364)) [@miscco](https://github.com/miscco) +- Deprecate old nvtext::normalize_characters ([#18360](https://github.com/NVIDIA/cudf/pull/18360)) [@davidwendt](https://github.com/davidwendt) +- refactor(rattler): enable strict channel priority for builds ([#18358](https://github.com/NVIDIA/cudf/pull/18358)) [@gforsyth](https://github.com/gforsyth) +- Optimize `sequences` by introducing `make_offsets_child_column` ([#18357](https://github.com/NVIDIA/cudf/pull/18357)) [@ustcfy](https://github.com/ustcfy) +- Decompress all data in a single `decompress_page_data` when reading Parquet input in a single chunk ([#18352](https://github.com/NVIDIA/cudf/pull/18352)) [@vuule](https://github.com/vuule) +- Moving wheel builds to specified location and uploading build artifacts to Github ([#18346](https://github.com/NVIDIA/cudf/pull/18346)) [@VenkateshJaya](https://github.com/VenkateshJaya) +- Performance improvement for to_lower/to_upper for multi-byte UTF-8 characters ([#18345](https://github.com/NVIDIA/cudf/pull/18345)) [@davidwendt](https://github.com/davidwendt) +- Branch 25.06 merge branch 25.04 ([#18344](https://github.com/NVIDIA/cudf/pull/18344)) [@vyasr](https://github.com/vyasr) +- Use dask-cuda for cudf-polars experimental testing ([#18343](https://github.com/NVIDIA/cudf/pull/18343)) [@rjzamora](https://github.com/rjzamora) +- Deprecate nvtext subword tokenizer ([#18334](https://github.com/NVIDIA/cudf/pull/18334)) [@davidwendt](https://github.com/davidwendt) +- Remove cudf.Scalar in as_column ([#18331](https://github.com/NVIDIA/cudf/pull/18331)) [@mroeschke](https://github.com/mroeschke) +- Add tests for `cudf.polars` to be able to work on a cpu-only machine ([#18327](https://github.com/NVIDIA/cudf/pull/18327)) [@galipremsagar](https://github.com/galipremsagar) +- Allow `cudf.DataFrame.from_pylibcudf` to accept a `pylibcudf.io.TableWithMetadata` ([#18319](https://github.com/NVIDIA/cudf/pull/18319)) [@mroeschke](https://github.com/mroeschke) +- Avoid stateful construction in `DataFrame.__init__` ([#18306](https://github.com/NVIDIA/cudf/pull/18306)) [@mroeschke](https://github.com/mroeschke) +- Improve the groupby performance for extremely low cardinality ([#18290](https://github.com/NVIDIA/cudf/pull/18290)) [@PointKernel](https://github.com/PointKernel) +- Remove extranous modules from top level cudf namespace ([#18287](https://github.com/NVIDIA/cudf/pull/18287)) [@mroeschke](https://github.com/mroeschke) +- Require type annotations in cudf.polars ([#18285](https://github.com/NVIDIA/cudf/pull/18285)) [@TomAugspurger](https://github.com/TomAugspurger) +- Removing unnecessary StreamSynchronization in reading ([#18279](https://github.com/NVIDIA/cudf/pull/18279)) [@JigaoLuo](https://github.com/JigaoLuo) +- Update to CCCL 2.8.x with no CCCL patches ([#18235](https://github.com/NVIDIA/cudf/pull/18235)) [@bdice](https://github.com/bdice) +- Reduce register pressure for compute_column_kernel ([#18226](https://github.com/NVIDIA/cudf/pull/18226)) [@matal-nvidia](https://github.com/matal-nvidia) +- Use the mapped buffer for all read operations in the memory-mapped source; switch default source to the kvikIO one ([#18204](https://github.com/NVIDIA/cudf/pull/18204)) [@vuule](https://github.com/vuule) +- Improve test coverage in the catboost integration tests ([#18126](https://github.com/NVIDIA/cudf/pull/18126)) [@Matt711](https://github.com/Matt711) +- Create file sources in parallel ([#18094](https://github.com/NVIDIA/cudf/pull/18094)) [@vuule](https://github.com/vuule) +- Enable `stumpy_distributed` tests ([#17969](https://github.com/NVIDIA/cudf/pull/17969)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor distinct join to use primitive row operators when proper ([#17726](https://github.com/NVIDIA/cudf/pull/17726)) [@PointKernel](https://github.com/PointKernel) +- Update chunked parquet reader benchmarks ([#16543](https://github.com/NVIDIA/cudf/pull/16543)) [@sdrp713](https://github.com/sdrp713) # cudf 25.04.00 (9 Apr 2025) ## 🚨 Breaking Changes -- Remove unused `group_range_rolling_window` API ([#18313](https://github.com/rapidsai/cudf/pull/18313)) [@wence-](https://github.com/wence-) -- [BUG] Disabled JIT for CUDA Runtime < 11.5 ([#18296](https://github.com/rapidsai/cudf/pull/18296)) [@lamarrr](https://github.com/lamarrr) -- Remove cudf.Scalar from binops ([#18240](https://github.com/rapidsai/cudf/pull/18240)) [@mroeschke](https://github.com/mroeschke) -- Enforce deprecation of dtype parameter in sum/product ([#18070](https://github.com/rapidsai/cudf/pull/18070)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated single component datetime extract APIs ([#18010](https://github.com/rapidsai/cudf/pull/18010)) [@Matt711](https://github.com/Matt711) -- Remove deprecated rolling window functionality ([#17993](https://github.com/rapidsai/cudf/pull/17993)) [@wence-](https://github.com/wence-) -- Remove deprecated nvtext::minhash_permuted APIs ([#17939](https://github.com/rapidsai/cudf/pull/17939)) [@davidwendt](https://github.com/davidwendt) -- Remove dataframe protocol ([#17909](https://github.com/rapidsai/cudf/pull/17909)) [@vyasr](https://github.com/vyasr) -- Use new rapids-logger library ([#17899](https://github.com/rapidsai/cudf/pull/17899)) [@vyasr](https://github.com/vyasr) -- Added Multi-input & Scalar Support for Transform UDFs ([#17881](https://github.com/rapidsai/cudf/pull/17881)) [@lamarrr](https://github.com/lamarrr) -- Fixed incorrect PTX parsing of `ret` instruction after branch label ([#17859](https://github.com/rapidsai/cudf/pull/17859)) [@lamarrr](https://github.com/lamarrr) -- Use KvikIO to enable file's fast host read and host write ([#17764](https://github.com/rapidsai/cudf/pull/17764)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Remove unused `group_range_rolling_window` API ([#18313](https://github.com/NVIDIA/cudf/pull/18313)) [@wence-](https://github.com/wence-) +- [BUG] Disabled JIT for CUDA Runtime < 11.5 ([#18296](https://github.com/NVIDIA/cudf/pull/18296)) [@lamarrr](https://github.com/lamarrr) +- Remove cudf.Scalar from binops ([#18240](https://github.com/NVIDIA/cudf/pull/18240)) [@mroeschke](https://github.com/mroeschke) +- Enforce deprecation of dtype parameter in sum/product ([#18070](https://github.com/NVIDIA/cudf/pull/18070)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated single component datetime extract APIs ([#18010](https://github.com/NVIDIA/cudf/pull/18010)) [@Matt711](https://github.com/Matt711) +- Remove deprecated rolling window functionality ([#17993](https://github.com/NVIDIA/cudf/pull/17993)) [@wence-](https://github.com/wence-) +- Remove deprecated nvtext::minhash_permuted APIs ([#17939](https://github.com/NVIDIA/cudf/pull/17939)) [@davidwendt](https://github.com/davidwendt) +- Remove dataframe protocol ([#17909](https://github.com/NVIDIA/cudf/pull/17909)) [@vyasr](https://github.com/vyasr) +- Use new rapids-logger library ([#17899](https://github.com/NVIDIA/cudf/pull/17899)) [@vyasr](https://github.com/vyasr) +- Added Multi-input & Scalar Support for Transform UDFs ([#17881](https://github.com/NVIDIA/cudf/pull/17881)) [@lamarrr](https://github.com/lamarrr) +- Fixed incorrect PTX parsing of `ret` instruction after branch label ([#17859](https://github.com/NVIDIA/cudf/pull/17859)) [@lamarrr](https://github.com/lamarrr) +- Use KvikIO to enable file's fast host read and host write ([#17764](https://github.com/NVIDIA/cudf/pull/17764)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) ## 🐛 Bug Fixes -- Fix alpha versions of cudf package. ([#18429](https://github.com/rapidsai/cudf/pull/18429)) [@bdice](https://github.com/bdice) -- Backport: Deterministic hashing for DataFrameScan nodes in cudf-polars multi-partition executor ([#18351) (#18420](https://github.com/rapidsai/cudf/pull/18351) (#18420)) [@bdice](https://github.com/bdice) -- Skip failing Narwhals rolling groupy tests ([#18398](https://github.com/rapidsai/cudf/pull/18398)) [@Matt711](https://github.com/Matt711) -- Pin cmake in test_java to be less than 4.0.0 ([#18392](https://github.com/rapidsai/cudf/pull/18392)) [@abellina](https://github.com/abellina) -- Skip polars tests that fail with pydantic deprecation warnings ([#18388](https://github.com/rapidsai/cudf/pull/18388)) [@Matt711](https://github.com/Matt711) -- Backport: Fix index of right table in unary operators in AST, in Joins ([#18342](https://github.com/rapidsai/cudf/pull/18342)) [@bdice](https://github.com/bdice) -- xfail narwhals sqlframe tests ([#18297](https://github.com/rapidsai/cudf/pull/18297)) [@Matt711](https://github.com/Matt711) -- [BUG] Disabled JIT for CUDA Runtime < 11.5 ([#18296](https://github.com/rapidsai/cudf/pull/18296)) [@lamarrr](https://github.com/lamarrr) -- Make a pylibcudf Column from a device array object with `strides=None` ([#18295](https://github.com/rapidsai/cudf/pull/18295)) [@Matt711](https://github.com/Matt711) -- Fix `cudf.pandas` objects to not be `Callable` ([#18288](https://github.com/rapidsai/cudf/pull/18288)) [@galipremsagar](https://github.com/galipremsagar) -- Skip failing polars test test_general_prefiltering ([#18264](https://github.com/rapidsai/cudf/pull/18264)) [@Matt711](https://github.com/Matt711) -- Filter all cudf.pandas profiler tests from running in parallel ([#18262](https://github.com/rapidsai/cudf/pull/18262)) [@Matt711](https://github.com/Matt711) -- Allow cudf.Series([pd.NA], dtype=, nan_as_null=False) ([#18259](https://github.com/rapidsai/cudf/pull/18259)) [@mroeschke](https://github.com/mroeschke) -- Fix `cross` join with extra columns ([#18256](https://github.com/rapidsai/cudf/pull/18256)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `Dataframe.loc` to not modify the actual dataframe ([#18254](https://github.com/rapidsai/cudf/pull/18254)) [@galipremsagar](https://github.com/galipremsagar) -- Remove RMM macro usage from to_arrow_device.cu ([#18252](https://github.com/rapidsai/cudf/pull/18252)) [@davidwendt](https://github.com/davidwendt) -- Skip Narwhals cross join tests for cudf.pandas CI run ([#18249](https://github.com/rapidsai/cudf/pull/18249)) [@Matt711](https://github.com/Matt711) -- Fix cudf-polars tests for polars < 1.24 ([#18246](https://github.com/rapidsai/cudf/pull/18246)) [@wence-](https://github.com/wence-) -- Fix experimental cudf-polars tests ([#18244](https://github.com/rapidsai/cudf/pull/18244)) [@rjzamora](https://github.com/rjzamora) -- Fix `datetime64` vs `datetime` binops max resolution ([#18241](https://github.com/rapidsai/cudf/pull/18241)) [@galipremsagar](https://github.com/galipremsagar) -- Use CCCL::libcudacxx include directories in Jitify preprocessing. ([#18233](https://github.com/rapidsai/cudf/pull/18233)) [@bdice](https://github.com/bdice) -- Disable conda prefix patching to avoid mangling binaries ([#18225](https://github.com/rapidsai/cudf/pull/18225)) [@vyasr](https://github.com/vyasr) -- Workaround for ARM compiler issue with single space literal string ([#18220](https://github.com/rapidsai/cudf/pull/18220)) [@davidwendt](https://github.com/davidwendt) -- Bump nightly check limit ([#18213](https://github.com/rapidsai/cudf/pull/18213)) [@Matt711](https://github.com/Matt711) -- Support comparitive binops between catgorical and non categorical ([#18200](https://github.com/rapidsai/cudf/pull/18200)) [@mroeschke](https://github.com/mroeschke) -- Make the version file inside cudf.pandas not a symlink ([#18198](https://github.com/rapidsai/cudf/pull/18198)) [@vyasr](https://github.com/vyasr) -- Ensure RAPIDS_ARTIFACTS_DIR is set for build metrics reports. ([#18192](https://github.com/rapidsai/cudf/pull/18192)) [@bdice](https://github.com/bdice) -- Ignore run exports of libcufile. ([#18190](https://github.com/rapidsai/cudf/pull/18190)) [@bdice](https://github.com/bdice) -- Skip flaky multi GPU test ([#18187](https://github.com/rapidsai/cudf/pull/18187)) [@Matt711](https://github.com/Matt711) -- Fix BPE merges table static-map capacity size ([#18184](https://github.com/rapidsai/cudf/pull/18184)) [@davidwendt](https://github.com/davidwendt) -- Drop `CUB_QUOTIENT_CEILING` ([#18179](https://github.com/rapidsai/cudf/pull/18179)) [@miscco](https://github.com/miscco) -- Disable ARM CI in C++ and Python test CI jobs ([#18175](https://github.com/rapidsai/cudf/pull/18175)) [@Matt711](https://github.com/Matt711) -- Add fmt to the test/benchmarks env ([#18173](https://github.com/rapidsai/cudf/pull/18173)) [@vyasr](https://github.com/vyasr) -- Fix merge(how=left, left_on=, right_index=True, sort=True) ([#18166](https://github.com/rapidsai/cudf/pull/18166)) [@mroeschke](https://github.com/mroeschke) -- Allow nonnative cupy dtype in cudf.Series ([#18164](https://github.com/rapidsai/cudf/pull/18164)) [@mroeschke](https://github.com/mroeschke) -- Fix Series construction from numpy array with non-native byte order ([#18151](https://github.com/rapidsai/cudf/pull/18151)) [@mroeschke](https://github.com/mroeschke) -- Use protocol for dlpack instead of deprecated function in cupy notebook ([#18147](https://github.com/rapidsai/cudf/pull/18147)) [@Matt711](https://github.com/Matt711) -- Skip failing test ([#18146](https://github.com/rapidsai/cudf/pull/18146)) [@vyasr](https://github.com/vyasr) -- Update calls to KvikIO's config setter ([#18144](https://github.com/rapidsai/cudf/pull/18144)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Reduce memory use when writing tables with very short columns to ORC ([#18136](https://github.com/rapidsai/cudf/pull/18136)) [@vuule](https://github.com/vuule) -- Handle empty dictionary in to_arrow_device interop ([#18121](https://github.com/rapidsai/cudf/pull/18121)) [@davidwendt](https://github.com/davidwendt) -- Allow pivot_table to accept single label index and column arguments ([#18115](https://github.com/rapidsai/cudf/pull/18115)) [@mroeschke](https://github.com/mroeschke) -- Preserve DataFrame.column subclass and type during binop ([#18113](https://github.com/rapidsai/cudf/pull/18113)) [@mroeschke](https://github.com/mroeschke) -- Fix rmm macro call ([#18108](https://github.com/rapidsai/cudf/pull/18108)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Add include for `<functional>` ([#18102](https://github.com/rapidsai/cudf/pull/18102)) [@miscco](https://github.com/miscco) -- Remove static column vectors from window function tests. ([#18099](https://github.com/rapidsai/cudf/pull/18099)) [@mythrocks](https://github.com/mythrocks) -- Fix scatter_by_map with spilling enabled ([#18095](https://github.com/rapidsai/cudf/pull/18095)) [@mroeschke](https://github.com/mroeschke) -- Use the right version macro `CCCL_MAJOR_VERSION` ([#18073](https://github.com/rapidsai/cudf/pull/18073)) [@miscco](https://github.com/miscco) -- Fix `test_scan_csv_multi` cudf-polars test ([#18064](https://github.com/rapidsai/cudf/pull/18064)) [@rjzamora](https://github.com/rjzamora) -- Fix memcopy direction for concatenate ([#18058](https://github.com/rapidsai/cudf/pull/18058)) [@tgujar](https://github.com/tgujar) -- Fix upstream dask `loc` test ([#18045](https://github.com/rapidsai/cudf/pull/18045)) [@rjzamora](https://github.com/rjzamora) -- Fix hang on invalid UTF-8 data in string_view iterator ([#18039](https://github.com/rapidsai/cudf/pull/18039)) [@davidwendt](https://github.com/davidwendt) -- Fix `dask_cudf.to_orc` deprecation ([#18038](https://github.com/rapidsai/cudf/pull/18038)) [@rjzamora](https://github.com/rjzamora) -- Compatibility with dask.dataframe's `is_scalar` ([#18030](https://github.com/rapidsai/cudf/pull/18030)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix the build error due to KvikIO update ([#18025](https://github.com/rapidsai/cudf/pull/18025)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Fix failing ibis test ([#18022](https://github.com/rapidsai/cudf/pull/18022)) [@Matt711](https://github.com/Matt711) -- Skip failing polars tests ([#18015](https://github.com/rapidsai/cudf/pull/18015)) [@Matt711](https://github.com/Matt711) -- Fix `to_arrow` to return consistent pandas-metadata ([#18009](https://github.com/rapidsai/cudf/pull/18009)) [@galipremsagar](https://github.com/galipremsagar) -- Prevent setting custom attributes to `ColumnMethods` ([#18005](https://github.com/rapidsai/cudf/pull/18005)) [@galipremsagar](https://github.com/galipremsagar) -- Compatibility with Dask `main` ([#17992](https://github.com/rapidsai/cudf/pull/17992)) [@TomAugspurger](https://github.com/TomAugspurger) -- [Bug] Fix Parquet-metadata sampling in cudf-polars ([#17991](https://github.com/rapidsai/cudf/pull/17991)) [@rjzamora](https://github.com/rjzamora) -- Add missing include for calling std::iota() ([#17983](https://github.com/rapidsai/cudf/pull/17983)) [@davidwendt](https://github.com/davidwendt) -- Fix pickle and unpickling for all objects ([#17980](https://github.com/rapidsai/cudf/pull/17980)) [@galipremsagar](https://github.com/galipremsagar) -- Install duckdb the default backend for ibis in the cudf.pandas integration tests ([#17972](https://github.com/rapidsai/cudf/pull/17972)) [@Matt711](https://github.com/Matt711) -- Check null count too in sum aggregation ([#17964](https://github.com/rapidsai/cudf/pull/17964)) [@Matt711](https://github.com/Matt711) -- Raise NotImplementedError for groupby.agg if duplicate columns would be created ([#17956](https://github.com/rapidsai/cudf/pull/17956)) [@mroeschke](https://github.com/mroeschke) -- Ensure disabling the module accelerator is thread-safe ([#17955](https://github.com/rapidsai/cudf/pull/17955)) [@vyasr](https://github.com/vyasr) -- Fix DataFrame/Series.rank for int and null data in mode.pandas_compatible ([#17954](https://github.com/rapidsai/cudf/pull/17954)) [@mroeschke](https://github.com/mroeschke) -- Limit buffer size in reallocation policy in JSON reader ([#17940](https://github.com/rapidsai/cudf/pull/17940)) [@shrshi](https://github.com/shrshi) -- Make `cudf.pandas` proxy array picklable ([#17929](https://github.com/rapidsai/cudf/pull/17929)) [@Matt711](https://github.com/Matt711) -- Add missing standard includes ([#17928](https://github.com/rapidsai/cudf/pull/17928)) [@miscco](https://github.com/miscco) -- Fix torch integration test ([#17923](https://github.com/rapidsai/cudf/pull/17923)) [@Matt711](https://github.com/Matt711) -- Fix `to_pandas` writable bug for `datetime` and `timedelta` types ([#17913](https://github.com/rapidsai/cudf/pull/17913)) [@galipremsagar](https://github.com/galipremsagar) -- Raise NotImplementedError if `.merge(suffixes=)` introduces duplicate labels ([#17905](https://github.com/rapidsai/cudf/pull/17905)) [@mroeschke](https://github.com/mroeschke) -- Fix groupby scans with int and NA data in mode.pandas_compatible ([#17895](https://github.com/rapidsai/cudf/pull/17895)) [@mroeschke](https://github.com/mroeschke) -- Patch `__init__` of `cudf` constructors to parse through `cudf.pandas` proxy objects ([#17878](https://github.com/rapidsai/cudf/pull/17878)) [@galipremsagar](https://github.com/galipremsagar) -- Fixed incorrect PTX parsing of `ret` instruction after branch label ([#17859](https://github.com/rapidsai/cudf/pull/17859)) [@lamarrr](https://github.com/lamarrr) -- Relax inconsistent schema handling in `dask_cudf.read_parquet` ([#17554](https://github.com/rapidsai/cudf/pull/17554)) [@rjzamora](https://github.com/rjzamora) +- Fix alpha versions of cudf package. ([#18429](https://github.com/NVIDIA/cudf/pull/18429)) [@bdice](https://github.com/bdice) +- Backport: Deterministic hashing for DataFrameScan nodes in cudf-polars multi-partition executor ([#18351) (#18420](https://github.com/NVIDIA/cudf/pull/18351) (#18420)) [@bdice](https://github.com/bdice) +- Skip failing Narwhals rolling groupy tests ([#18398](https://github.com/NVIDIA/cudf/pull/18398)) [@Matt711](https://github.com/Matt711) +- Pin cmake in test_java to be less than 4.0.0 ([#18392](https://github.com/NVIDIA/cudf/pull/18392)) [@abellina](https://github.com/abellina) +- Skip polars tests that fail with pydantic deprecation warnings ([#18388](https://github.com/NVIDIA/cudf/pull/18388)) [@Matt711](https://github.com/Matt711) +- Backport: Fix index of right table in unary operators in AST, in Joins ([#18342](https://github.com/NVIDIA/cudf/pull/18342)) [@bdice](https://github.com/bdice) +- xfail narwhals sqlframe tests ([#18297](https://github.com/NVIDIA/cudf/pull/18297)) [@Matt711](https://github.com/Matt711) +- [BUG] Disabled JIT for CUDA Runtime < 11.5 ([#18296](https://github.com/NVIDIA/cudf/pull/18296)) [@lamarrr](https://github.com/lamarrr) +- Make a pylibcudf Column from a device array object with `strides=None` ([#18295](https://github.com/NVIDIA/cudf/pull/18295)) [@Matt711](https://github.com/Matt711) +- Fix `cudf.pandas` objects to not be `Callable` ([#18288](https://github.com/NVIDIA/cudf/pull/18288)) [@galipremsagar](https://github.com/galipremsagar) +- Skip failing polars test test_general_prefiltering ([#18264](https://github.com/NVIDIA/cudf/pull/18264)) [@Matt711](https://github.com/Matt711) +- Filter all cudf.pandas profiler tests from running in parallel ([#18262](https://github.com/NVIDIA/cudf/pull/18262)) [@Matt711](https://github.com/Matt711) +- Allow cudf.Series([pd.NA], dtype=, nan_as_null=False) ([#18259](https://github.com/NVIDIA/cudf/pull/18259)) [@mroeschke](https://github.com/mroeschke) +- Fix `cross` join with extra columns ([#18256](https://github.com/NVIDIA/cudf/pull/18256)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `Dataframe.loc` to not modify the actual dataframe ([#18254](https://github.com/NVIDIA/cudf/pull/18254)) [@galipremsagar](https://github.com/galipremsagar) +- Remove RMM macro usage from to_arrow_device.cu ([#18252](https://github.com/NVIDIA/cudf/pull/18252)) [@davidwendt](https://github.com/davidwendt) +- Skip Narwhals cross join tests for cudf.pandas CI run ([#18249](https://github.com/NVIDIA/cudf/pull/18249)) [@Matt711](https://github.com/Matt711) +- Fix cudf-polars tests for polars < 1.24 ([#18246](https://github.com/NVIDIA/cudf/pull/18246)) [@wence-](https://github.com/wence-) +- Fix experimental cudf-polars tests ([#18244](https://github.com/NVIDIA/cudf/pull/18244)) [@rjzamora](https://github.com/rjzamora) +- Fix `datetime64` vs `datetime` binops max resolution ([#18241](https://github.com/NVIDIA/cudf/pull/18241)) [@galipremsagar](https://github.com/galipremsagar) +- Use CCCL::libcudacxx include directories in Jitify preprocessing. ([#18233](https://github.com/NVIDIA/cudf/pull/18233)) [@bdice](https://github.com/bdice) +- Disable conda prefix patching to avoid mangling binaries ([#18225](https://github.com/NVIDIA/cudf/pull/18225)) [@vyasr](https://github.com/vyasr) +- Workaround for ARM compiler issue with single space literal string ([#18220](https://github.com/NVIDIA/cudf/pull/18220)) [@davidwendt](https://github.com/davidwendt) +- Bump nightly check limit ([#18213](https://github.com/NVIDIA/cudf/pull/18213)) [@Matt711](https://github.com/Matt711) +- Support comparitive binops between catgorical and non categorical ([#18200](https://github.com/NVIDIA/cudf/pull/18200)) [@mroeschke](https://github.com/mroeschke) +- Make the version file inside cudf.pandas not a symlink ([#18198](https://github.com/NVIDIA/cudf/pull/18198)) [@vyasr](https://github.com/vyasr) +- Ensure RAPIDS_ARTIFACTS_DIR is set for build metrics reports. ([#18192](https://github.com/NVIDIA/cudf/pull/18192)) [@bdice](https://github.com/bdice) +- Ignore run exports of libcufile. ([#18190](https://github.com/NVIDIA/cudf/pull/18190)) [@bdice](https://github.com/bdice) +- Skip flaky multi GPU test ([#18187](https://github.com/NVIDIA/cudf/pull/18187)) [@Matt711](https://github.com/Matt711) +- Fix BPE merges table static-map capacity size ([#18184](https://github.com/NVIDIA/cudf/pull/18184)) [@davidwendt](https://github.com/davidwendt) +- Drop `CUB_QUOTIENT_CEILING` ([#18179](https://github.com/NVIDIA/cudf/pull/18179)) [@miscco](https://github.com/miscco) +- Disable ARM CI in C++ and Python test CI jobs ([#18175](https://github.com/NVIDIA/cudf/pull/18175)) [@Matt711](https://github.com/Matt711) +- Add fmt to the test/benchmarks env ([#18173](https://github.com/NVIDIA/cudf/pull/18173)) [@vyasr](https://github.com/vyasr) +- Fix merge(how=left, left_on=, right_index=True, sort=True) ([#18166](https://github.com/NVIDIA/cudf/pull/18166)) [@mroeschke](https://github.com/mroeschke) +- Allow nonnative cupy dtype in cudf.Series ([#18164](https://github.com/NVIDIA/cudf/pull/18164)) [@mroeschke](https://github.com/mroeschke) +- Fix Series construction from numpy array with non-native byte order ([#18151](https://github.com/NVIDIA/cudf/pull/18151)) [@mroeschke](https://github.com/mroeschke) +- Use protocol for dlpack instead of deprecated function in cupy notebook ([#18147](https://github.com/NVIDIA/cudf/pull/18147)) [@Matt711](https://github.com/Matt711) +- Skip failing test ([#18146](https://github.com/NVIDIA/cudf/pull/18146)) [@vyasr](https://github.com/vyasr) +- Update calls to KvikIO's config setter ([#18144](https://github.com/NVIDIA/cudf/pull/18144)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Reduce memory use when writing tables with very short columns to ORC ([#18136](https://github.com/NVIDIA/cudf/pull/18136)) [@vuule](https://github.com/vuule) +- Handle empty dictionary in to_arrow_device interop ([#18121](https://github.com/NVIDIA/cudf/pull/18121)) [@davidwendt](https://github.com/davidwendt) +- Allow pivot_table to accept single label index and column arguments ([#18115](https://github.com/NVIDIA/cudf/pull/18115)) [@mroeschke](https://github.com/mroeschke) +- Preserve DataFrame.column subclass and type during binop ([#18113](https://github.com/NVIDIA/cudf/pull/18113)) [@mroeschke](https://github.com/mroeschke) +- Fix rmm macro call ([#18108](https://github.com/NVIDIA/cudf/pull/18108)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Add include for `<functional>` ([#18102](https://github.com/NVIDIA/cudf/pull/18102)) [@miscco](https://github.com/miscco) +- Remove static column vectors from window function tests. ([#18099](https://github.com/NVIDIA/cudf/pull/18099)) [@mythrocks](https://github.com/mythrocks) +- Fix scatter_by_map with spilling enabled ([#18095](https://github.com/NVIDIA/cudf/pull/18095)) [@mroeschke](https://github.com/mroeschke) +- Use the right version macro `CCCL_MAJOR_VERSION` ([#18073](https://github.com/NVIDIA/cudf/pull/18073)) [@miscco](https://github.com/miscco) +- Fix `test_scan_csv_multi` cudf-polars test ([#18064](https://github.com/NVIDIA/cudf/pull/18064)) [@rjzamora](https://github.com/rjzamora) +- Fix memcopy direction for concatenate ([#18058](https://github.com/NVIDIA/cudf/pull/18058)) [@tgujar](https://github.com/tgujar) +- Fix upstream dask `loc` test ([#18045](https://github.com/NVIDIA/cudf/pull/18045)) [@rjzamora](https://github.com/rjzamora) +- Fix hang on invalid UTF-8 data in string_view iterator ([#18039](https://github.com/NVIDIA/cudf/pull/18039)) [@davidwendt](https://github.com/davidwendt) +- Fix `dask_cudf.to_orc` deprecation ([#18038](https://github.com/NVIDIA/cudf/pull/18038)) [@rjzamora](https://github.com/rjzamora) +- Compatibility with dask.dataframe's `is_scalar` ([#18030](https://github.com/NVIDIA/cudf/pull/18030)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix the build error due to KvikIO update ([#18025](https://github.com/NVIDIA/cudf/pull/18025)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Fix failing ibis test ([#18022](https://github.com/NVIDIA/cudf/pull/18022)) [@Matt711](https://github.com/Matt711) +- Skip failing polars tests ([#18015](https://github.com/NVIDIA/cudf/pull/18015)) [@Matt711](https://github.com/Matt711) +- Fix `to_arrow` to return consistent pandas-metadata ([#18009](https://github.com/NVIDIA/cudf/pull/18009)) [@galipremsagar](https://github.com/galipremsagar) +- Prevent setting custom attributes to `ColumnMethods` ([#18005](https://github.com/NVIDIA/cudf/pull/18005)) [@galipremsagar](https://github.com/galipremsagar) +- Compatibility with Dask `main` ([#17992](https://github.com/NVIDIA/cudf/pull/17992)) [@TomAugspurger](https://github.com/TomAugspurger) +- [Bug] Fix Parquet-metadata sampling in cudf-polars ([#17991](https://github.com/NVIDIA/cudf/pull/17991)) [@rjzamora](https://github.com/rjzamora) +- Add missing include for calling std::iota() ([#17983](https://github.com/NVIDIA/cudf/pull/17983)) [@davidwendt](https://github.com/davidwendt) +- Fix pickle and unpickling for all objects ([#17980](https://github.com/NVIDIA/cudf/pull/17980)) [@galipremsagar](https://github.com/galipremsagar) +- Install duckdb the default backend for ibis in the cudf.pandas integration tests ([#17972](https://github.com/NVIDIA/cudf/pull/17972)) [@Matt711](https://github.com/Matt711) +- Check null count too in sum aggregation ([#17964](https://github.com/NVIDIA/cudf/pull/17964)) [@Matt711](https://github.com/Matt711) +- Raise NotImplementedError for groupby.agg if duplicate columns would be created ([#17956](https://github.com/NVIDIA/cudf/pull/17956)) [@mroeschke](https://github.com/mroeschke) +- Ensure disabling the module accelerator is thread-safe ([#17955](https://github.com/NVIDIA/cudf/pull/17955)) [@vyasr](https://github.com/vyasr) +- Fix DataFrame/Series.rank for int and null data in mode.pandas_compatible ([#17954](https://github.com/NVIDIA/cudf/pull/17954)) [@mroeschke](https://github.com/mroeschke) +- Limit buffer size in reallocation policy in JSON reader ([#17940](https://github.com/NVIDIA/cudf/pull/17940)) [@shrshi](https://github.com/shrshi) +- Make `cudf.pandas` proxy array picklable ([#17929](https://github.com/NVIDIA/cudf/pull/17929)) [@Matt711](https://github.com/Matt711) +- Add missing standard includes ([#17928](https://github.com/NVIDIA/cudf/pull/17928)) [@miscco](https://github.com/miscco) +- Fix torch integration test ([#17923](https://github.com/NVIDIA/cudf/pull/17923)) [@Matt711](https://github.com/Matt711) +- Fix `to_pandas` writable bug for `datetime` and `timedelta` types ([#17913](https://github.com/NVIDIA/cudf/pull/17913)) [@galipremsagar](https://github.com/galipremsagar) +- Raise NotImplementedError if `.merge(suffixes=)` introduces duplicate labels ([#17905](https://github.com/NVIDIA/cudf/pull/17905)) [@mroeschke](https://github.com/mroeschke) +- Fix groupby scans with int and NA data in mode.pandas_compatible ([#17895](https://github.com/NVIDIA/cudf/pull/17895)) [@mroeschke](https://github.com/mroeschke) +- Patch `__init__` of `cudf` constructors to parse through `cudf.pandas` proxy objects ([#17878](https://github.com/NVIDIA/cudf/pull/17878)) [@galipremsagar](https://github.com/galipremsagar) +- Fixed incorrect PTX parsing of `ret` instruction after branch label ([#17859](https://github.com/NVIDIA/cudf/pull/17859)) [@lamarrr](https://github.com/lamarrr) +- Relax inconsistent schema handling in `dask_cudf.read_parquet` ([#17554](https://github.com/NVIDIA/cudf/pull/17554)) [@rjzamora](https://github.com/rjzamora) ## 📖 Documentation -- Clarify that cudf.pandas should be enabled before importing pandas. ([#18339](https://github.com/rapidsai/cudf/pull/18339)) [@bdice](https://github.com/bdice) -- [DOC] Add wordpiece tokenizer to cudf documentation ([#18247](https://github.com/rapidsai/cudf/pull/18247)) [@davidwendt](https://github.com/davidwendt) -- Added pylibcudf.contiguous_split to API docs ([#18194](https://github.com/rapidsai/cudf/pull/18194)) [@TomAugspurger](https://github.com/TomAugspurger) -- Fix build.sh docs for default behavior ([#18180](https://github.com/rapidsai/cudf/pull/18180)) [@bdice](https://github.com/bdice) -- Update Dask-cuDF documentation to fix all warnings and errors ([#18157](https://github.com/rapidsai/cudf/pull/18157)) [@TomAugspurger](https://github.com/TomAugspurger) -- [DOC] Document character normalizer ([#18125](https://github.com/rapidsai/cudf/pull/18125)) [@Matt711](https://github.com/Matt711) +- Clarify that cudf.pandas should be enabled before importing pandas. ([#18339](https://github.com/NVIDIA/cudf/pull/18339)) [@bdice](https://github.com/bdice) +- [DOC] Add wordpiece tokenizer to cudf documentation ([#18247](https://github.com/NVIDIA/cudf/pull/18247)) [@davidwendt](https://github.com/davidwendt) +- Added pylibcudf.contiguous_split to API docs ([#18194](https://github.com/NVIDIA/cudf/pull/18194)) [@TomAugspurger](https://github.com/TomAugspurger) +- Fix build.sh docs for default behavior ([#18180](https://github.com/NVIDIA/cudf/pull/18180)) [@bdice](https://github.com/bdice) +- Update Dask-cuDF documentation to fix all warnings and errors ([#18157](https://github.com/NVIDIA/cudf/pull/18157)) [@TomAugspurger](https://github.com/TomAugspurger) +- [DOC] Document character normalizer ([#18125](https://github.com/NVIDIA/cudf/pull/18125)) [@Matt711](https://github.com/Matt711) ## 🚀 New Features -- Add and revise experimental cudf-polars config options ([#18284](https://github.com/rapidsai/cudf/pull/18284)) [@rjzamora](https://github.com/rjzamora) -- Support `top-k` and `bottom_k` expressions ([#18222](https://github.com/rapidsai/cudf/pull/18222)) [@Matt711](https://github.com/Matt711) -- Support `cudf-polars` `is_leap_year` ([#18212](https://github.com/rapidsai/cudf/pull/18212)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `cudf-polars` `month_start`/`month_end` ([#18211](https://github.com/rapidsai/cudf/pull/18211)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `cudf-polars` `ordinal_day` ([#18152](https://github.com/rapidsai/cudf/pull/18152)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add `pylibcudf.gpumemoryview` support for `len()`/`nbytes` ([#18133](https://github.com/rapidsai/cudf/pull/18133)) [@pentschev](https://github.com/pentschev) -- Link to libzstd for ZSTD compression and decompression APIs ([#18129](https://github.com/rapidsai/cudf/pull/18129)) [@shrshi](https://github.com/shrshi) -- Added NDSH Q09 Benchmark for Transforms ([#18127](https://github.com/rapidsai/cudf/pull/18127)) [@lamarrr](https://github.com/lamarrr) -- Make pylibcudf traits raise exceptions gracefully rather than terminating in C++ ([#18117](https://github.com/rapidsai/cudf/pull/18117)) [@Matt711](https://github.com/Matt711) -- Host decompression ([#18114](https://github.com/rapidsai/cudf/pull/18114)) [@vuule](https://github.com/vuule) -- Add owning types to hold Arrow data ([#18084](https://github.com/rapidsai/cudf/pull/18084)) [@vyasr](https://github.com/vyasr) -- Bump polars version to <1.24 ([#18076](https://github.com/rapidsai/cudf/pull/18076)) [@Matt711](https://github.com/Matt711) -- Support sorted merges in cudf.polars ([#18075](https://github.com/rapidsai/cudf/pull/18075)) [@Matt711](https://github.com/Matt711) -- Add a slice expression to polars IR ([#18050](https://github.com/rapidsai/cudf/pull/18050)) [@Matt711](https://github.com/Matt711) -- Expose `num_rows_per_source` (IO metadata) to pylibcudf ([#18049](https://github.com/rapidsai/cudf/pull/18049)) [@Matt711](https://github.com/Matt711) -- Added Imbalanced Tree Benchmarks for Transforms ([#18032](https://github.com/rapidsai/cudf/pull/18032)) [@lamarrr](https://github.com/lamarrr) -- Run the narwhals test suite with cudf.pandas ([#18031](https://github.com/rapidsai/cudf/pull/18031)) [@Matt711](https://github.com/Matt711) -- Add `host_read_async` interfaces to `datasource` ([#18018](https://github.com/rapidsai/cudf/pull/18018)) [@vuule](https://github.com/vuule) -- Make most cudf-polars `Node` objects pickleable ([#17998](https://github.com/rapidsai/cudf/pull/17998)) [@rjzamora](https://github.com/rjzamora) -- Add `Column.serialize` to cudf-polars ([#17990](https://github.com/rapidsai/cudf/pull/17990)) [@rjzamora](https://github.com/rjzamora) -- Bump polars version to <1.23 ([#17986](https://github.com/rapidsai/cudf/pull/17986)) [@Matt711](https://github.com/Matt711) -- Implemented Decimal Transforms ([#17968](https://github.com/rapidsai/cudf/pull/17968)) [@lamarrr](https://github.com/lamarrr) -- Introduce ZSTD host-side compression and decompression APIs ([#17935](https://github.com/rapidsai/cudf/pull/17935)) [@shrshi](https://github.com/shrshi) -- Add catboost integration tests ([#17931](https://github.com/rapidsai/cudf/pull/17931)) [@Matt711](https://github.com/Matt711) -- [FEA] Expose `stripe_size_rows` setting for `ORCWriterOptions` ([#17927](https://github.com/rapidsai/cudf/pull/17927)) [@ustcfy](https://github.com/ustcfy) -- Test narwhals in CI ([#17884](https://github.com/rapidsai/cudf/pull/17884)) [@bdice](https://github.com/bdice) -- Added Multi-input & Scalar Support for Transform UDFs ([#17881](https://github.com/rapidsai/cudf/pull/17881)) [@lamarrr](https://github.com/lamarrr) -- Host Snappy compression ([#17824](https://github.com/rapidsai/cudf/pull/17824)) [@vuule](https://github.com/vuule) -- Run spark-rapids-jni CI ([#17781](https://github.com/rapidsai/cudf/pull/17781)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Add multi-partition `Shuffle` operation to cuDF Polars ([#17744](https://github.com/rapidsai/cudf/pull/17744)) [@rjzamora](https://github.com/rjzamora) -- Added polynomials benchmark ([#17695](https://github.com/rapidsai/cudf/pull/17695)) [@lamarrr](https://github.com/lamarrr) -- Add stream parameters in pylibcudf IO APIs ([#17620](https://github.com/rapidsai/cudf/pull/17620)) [@Matt711](https://github.com/Matt711) -- New nvtext::wordpiece_tokenizer APIs ([#17600](https://github.com/rapidsai/cudf/pull/17600)) [@davidwendt](https://github.com/davidwendt) -- Add support for unary negation operator ([#17560](https://github.com/rapidsai/cudf/pull/17560)) [@Matt711](https://github.com/Matt711) -- Add multi-partition `Join` support to cuDF-Polars ([#17518](https://github.com/rapidsai/cudf/pull/17518)) [@rjzamora](https://github.com/rjzamora) -- Add basic multi-partition `GroupBy` support to cuDF-Polars ([#17503](https://github.com/rapidsai/cudf/pull/17503)) [@rjzamora](https://github.com/rjzamora) -- Support Distributed in cudf-polars tests and IR evaluation ([#17364](https://github.com/rapidsai/cudf/pull/17364)) [@pentschev](https://github.com/pentschev) +- Add and revise experimental cudf-polars config options ([#18284](https://github.com/NVIDIA/cudf/pull/18284)) [@rjzamora](https://github.com/rjzamora) +- Support `top-k` and `bottom_k` expressions ([#18222](https://github.com/NVIDIA/cudf/pull/18222)) [@Matt711](https://github.com/Matt711) +- Support `cudf-polars` `is_leap_year` ([#18212](https://github.com/NVIDIA/cudf/pull/18212)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `cudf-polars` `month_start`/`month_end` ([#18211](https://github.com/NVIDIA/cudf/pull/18211)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `cudf-polars` `ordinal_day` ([#18152](https://github.com/NVIDIA/cudf/pull/18152)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add `pylibcudf.gpumemoryview` support for `len()`/`nbytes` ([#18133](https://github.com/NVIDIA/cudf/pull/18133)) [@pentschev](https://github.com/pentschev) +- Link to libzstd for ZSTD compression and decompression APIs ([#18129](https://github.com/NVIDIA/cudf/pull/18129)) [@shrshi](https://github.com/shrshi) +- Added NDSH Q09 Benchmark for Transforms ([#18127](https://github.com/NVIDIA/cudf/pull/18127)) [@lamarrr](https://github.com/lamarrr) +- Make pylibcudf traits raise exceptions gracefully rather than terminating in C++ ([#18117](https://github.com/NVIDIA/cudf/pull/18117)) [@Matt711](https://github.com/Matt711) +- Host decompression ([#18114](https://github.com/NVIDIA/cudf/pull/18114)) [@vuule](https://github.com/vuule) +- Add owning types to hold Arrow data ([#18084](https://github.com/NVIDIA/cudf/pull/18084)) [@vyasr](https://github.com/vyasr) +- Bump polars version to <1.24 ([#18076](https://github.com/NVIDIA/cudf/pull/18076)) [@Matt711](https://github.com/Matt711) +- Support sorted merges in cudf.polars ([#18075](https://github.com/NVIDIA/cudf/pull/18075)) [@Matt711](https://github.com/Matt711) +- Add a slice expression to polars IR ([#18050](https://github.com/NVIDIA/cudf/pull/18050)) [@Matt711](https://github.com/Matt711) +- Expose `num_rows_per_source` (IO metadata) to pylibcudf ([#18049](https://github.com/NVIDIA/cudf/pull/18049)) [@Matt711](https://github.com/Matt711) +- Added Imbalanced Tree Benchmarks for Transforms ([#18032](https://github.com/NVIDIA/cudf/pull/18032)) [@lamarrr](https://github.com/lamarrr) +- Run the narwhals test suite with cudf.pandas ([#18031](https://github.com/NVIDIA/cudf/pull/18031)) [@Matt711](https://github.com/Matt711) +- Add `host_read_async` interfaces to `datasource` ([#18018](https://github.com/NVIDIA/cudf/pull/18018)) [@vuule](https://github.com/vuule) +- Make most cudf-polars `Node` objects pickleable ([#17998](https://github.com/NVIDIA/cudf/pull/17998)) [@rjzamora](https://github.com/rjzamora) +- Add `Column.serialize` to cudf-polars ([#17990](https://github.com/NVIDIA/cudf/pull/17990)) [@rjzamora](https://github.com/rjzamora) +- Bump polars version to <1.23 ([#17986](https://github.com/NVIDIA/cudf/pull/17986)) [@Matt711](https://github.com/Matt711) +- Implemented Decimal Transforms ([#17968](https://github.com/NVIDIA/cudf/pull/17968)) [@lamarrr](https://github.com/lamarrr) +- Introduce ZSTD host-side compression and decompression APIs ([#17935](https://github.com/NVIDIA/cudf/pull/17935)) [@shrshi](https://github.com/shrshi) +- Add catboost integration tests ([#17931](https://github.com/NVIDIA/cudf/pull/17931)) [@Matt711](https://github.com/Matt711) +- [FEA] Expose `stripe_size_rows` setting for `ORCWriterOptions` ([#17927](https://github.com/NVIDIA/cudf/pull/17927)) [@ustcfy](https://github.com/ustcfy) +- Test narwhals in CI ([#17884](https://github.com/NVIDIA/cudf/pull/17884)) [@bdice](https://github.com/bdice) +- Added Multi-input & Scalar Support for Transform UDFs ([#17881](https://github.com/NVIDIA/cudf/pull/17881)) [@lamarrr](https://github.com/lamarrr) +- Host Snappy compression ([#17824](https://github.com/NVIDIA/cudf/pull/17824)) [@vuule](https://github.com/vuule) +- Run spark-rapids-jni CI ([#17781](https://github.com/NVIDIA/cudf/pull/17781)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Add multi-partition `Shuffle` operation to cuDF Polars ([#17744](https://github.com/NVIDIA/cudf/pull/17744)) [@rjzamora](https://github.com/rjzamora) +- Added polynomials benchmark ([#17695](https://github.com/NVIDIA/cudf/pull/17695)) [@lamarrr](https://github.com/lamarrr) +- Add stream parameters in pylibcudf IO APIs ([#17620](https://github.com/NVIDIA/cudf/pull/17620)) [@Matt711](https://github.com/Matt711) +- New nvtext::wordpiece_tokenizer APIs ([#17600](https://github.com/NVIDIA/cudf/pull/17600)) [@davidwendt](https://github.com/davidwendt) +- Add support for unary negation operator ([#17560](https://github.com/NVIDIA/cudf/pull/17560)) [@Matt711](https://github.com/Matt711) +- Add multi-partition `Join` support to cuDF-Polars ([#17518](https://github.com/NVIDIA/cudf/pull/17518)) [@rjzamora](https://github.com/rjzamora) +- Add basic multi-partition `GroupBy` support to cuDF-Polars ([#17503](https://github.com/NVIDIA/cudf/pull/17503)) [@rjzamora](https://github.com/rjzamora) +- Support Distributed in cudf-polars tests and IR evaluation ([#17364](https://github.com/NVIDIA/cudf/pull/17364)) [@pentschev](https://github.com/pentschev) ## 🛠️ Improvements -- Use pyarrow 15 in oldest dependency CI jobs ([#18409](https://github.com/rapidsai/cudf/pull/18409)) [@bdice](https://github.com/bdice) -- Bump librdkafka to 2.8.0 ([#18370](https://github.com/rapidsai/cudf/pull/18370)) [@raydouglass](https://github.com/raydouglass) -- fix(rattler): ignore `libzlib` run dependency to avoid `pandoc` collision ([#18368](https://github.com/rapidsai/cudf/pull/18368)) [@gforsyth](https://github.com/gforsyth) -- Fix zstd build interface include definition ([#18366](https://github.com/rapidsai/cudf/pull/18366)) [@trxcllnt](https://github.com/trxcllnt) -- test: Install pytest-env and hypothesis in test_narwhals.sh ([#18337](https://github.com/rapidsai/cudf/pull/18337)) [@MarcoGorelli](https://github.com/MarcoGorelli) -- Remove unused `group_range_rolling_window` API ([#18313](https://github.com/rapidsai/cudf/pull/18313)) [@wence-](https://github.com/wence-) -- Cache column view creation from arrow types ([#18302](https://github.com/rapidsai/cudf/pull/18302)) [@vyasr](https://github.com/vyasr) -- Split Narwhals cudf.pandas tests failures into to fix and to skip ([#18267](https://github.com/rapidsai/cudf/pull/18267)) [@mroeschke](https://github.com/mroeschke) -- Support BinOp, min, and max Aggregations in cudf-polars parallel groupby ([#18266](https://github.com/rapidsai/cudf/pull/18266)) [@TomAugspurger](https://github.com/TomAugspurger) -- Minor clean up and optimizations in the Parquet writer ([#18258](https://github.com/rapidsai/cudf/pull/18258)) [@vuule](https://github.com/vuule) -- Fix `cudf_kafka` run export for `cudatoolkit` ([#18245](https://github.com/rapidsai/cudf/pull/18245)) [@gforsyth](https://github.com/gforsyth) -- dask-polars: use splat everywhere. ([#18243](https://github.com/rapidsai/cudf/pull/18243)) [@madsbk](https://github.com/madsbk) -- Remove cudf.Scalar from binops ([#18240](https://github.com/rapidsai/cudf/pull/18240)) [@mroeschke](https://github.com/mroeschke) -- Remove warning in the stream pool when asking for more streams than available ([#18236](https://github.com/rapidsai/cudf/pull/18236)) [@vuule](https://github.com/vuule) -- Explain why we disable parallelism for profiler tests to avoid pytest-cov issue ([#18234](https://github.com/rapidsai/cudf/pull/18234)) [@Matt711](https://github.com/Matt711) -- Ignore `cudatoolkit` run exports by name, not package ([#18230](https://github.com/rapidsai/cudf/pull/18230)) [@gforsyth](https://github.com/gforsyth) -- Revert "Bump nightly check limit" ([#18227](https://github.com/rapidsai/cudf/pull/18227)) [@Matt711](https://github.com/Matt711) -- Fix `cudf.pandas` to be able to work on a cpu-only machine ([#18224](https://github.com/rapidsai/cudf/pull/18224)) [@galipremsagar](https://github.com/galipremsagar) -- Add missing `cudatoolkit` run_export ignore to `pylibcudf` ([#18223](https://github.com/rapidsai/cudf/pull/18223)) [@gforsyth](https://github.com/gforsyth) -- Remove cudf.Scalar from Column.__setitem__ ([#18221](https://github.com/rapidsai/cudf/pull/18221)) [@mroeschke](https://github.com/mroeschke) -- Remove unused round_up_pow2 utility ([#18218](https://github.com/rapidsai/cudf/pull/18218)) [@PointKernel](https://github.com/PointKernel) -- Add flake8-print/debugger Ruff rules ([#18217](https://github.com/rapidsai/cudf/pull/18217)) [@mroeschke](https://github.com/mroeschke) -- Bump polars version to <1.25 ([#18209](https://github.com/rapidsai/cudf/pull/18209)) [@Matt711](https://github.com/Matt711) -- Export RAPIDS_ARTIFACTS_DIR. ([#18208](https://github.com/rapidsai/cudf/pull/18208)) [@bdice](https://github.com/bdice) -- Drop more thrust functions with libcu++ ones ([#18207](https://github.com/rapidsai/cudf/pull/18207)) [@miscco](https://github.com/miscco) -- Update Numpy <2.1 unpinning xfail condition ([#18203](https://github.com/rapidsai/cudf/pull/18203)) [@mroeschke](https://github.com/mroeschke) -- Run conda import tests on Python packages ([#18197](https://github.com/rapidsai/cudf/pull/18197)) [@bdice](https://github.com/bdice) -- fix(rattler): add `cudatoolkit` ignore run export to `cudf` ([#18195](https://github.com/rapidsai/cudf/pull/18195)) [@gforsyth](https://github.com/gforsyth) -- Revert "Disable ARM CI in C++ and Python test CI jobs" ([#18188](https://github.com/rapidsai/cudf/pull/18188)) [@Matt711](https://github.com/Matt711) -- Define Column.where to be used across DataFrame/Series ([#18186](https://github.com/rapidsai/cudf/pull/18186)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf.Scalar in where ([#18178](https://github.com/rapidsai/cudf/pull/18178)) [@mroeschke](https://github.com/mroeschke) -- Drop unnecessary fmt dep ([#18177](https://github.com/rapidsai/cudf/pull/18177)) [@vyasr](https://github.com/vyasr) -- Refactor join internals: separate hash_join declaration and cleanup ([#18170](https://github.com/rapidsai/cudf/pull/18170)) [@PointKernel](https://github.com/PointKernel) -- Add Ruff rule to enforce cudf dtype utils over numpy/pandas dtype utils ([#18169](https://github.com/rapidsai/cudf/pull/18169)) [@mroeschke](https://github.com/mroeschke) -- Combine multiple str.minhash() APIs into one call ([#18168](https://github.com/rapidsai/cudf/pull/18168)) [@davidwendt](https://github.com/davidwendt) -- Move nanoarrow_utils.hpp from cpp/tests/interop to cpp/include/cudf_test ([#18163](https://github.com/rapidsai/cudf/pull/18163)) [@davidwendt](https://github.com/davidwendt) -- Test cudf against the latest stable branch of Narwhals ([#18162](https://github.com/rapidsai/cudf/pull/18162)) [@Matt711](https://github.com/Matt711) -- fix libcudf pins cu11 ([#18161](https://github.com/rapidsai/cudf/pull/18161)) [@gforsyth](https://github.com/gforsyth) -- Combine separate ConfigureNVBench calls to fix cpp conda builds ([#18155](https://github.com/rapidsai/cudf/pull/18155)) [@gforsyth](https://github.com/gforsyth) -- Add telemetry to build workflows ([#18154](https://github.com/rapidsai/cudf/pull/18154)) [@gforsyth](https://github.com/gforsyth) -- Prune more seldom used dtype utils ([#18150](https://github.com/rapidsai/cudf/pull/18150)) [@mroeschke](https://github.com/mroeschke) -- Remove some unnecessary module imports ([#18143](https://github.com/rapidsai/cudf/pull/18143)) [@mroeschke](https://github.com/mroeschke) -- Branch 25.04 merge branch 25.02 ([#18142](https://github.com/rapidsai/cudf/pull/18142)) [@vyasr](https://github.com/vyasr) -- Prune some seldom used dtype utils ([#18141](https://github.com/rapidsai/cudf/pull/18141)) [@mroeschke](https://github.com/mroeschke) -- Use more, cheaper dtype checking utilities in cudf Python ([#18139](https://github.com/rapidsai/cudf/pull/18139)) [@mroeschke](https://github.com/mroeschke) -- Support deserializing cudf-polars objects composed of RMM frames ([#18138](https://github.com/rapidsai/cudf/pull/18138)) [@pentschev](https://github.com/pentschev) -- Add `ConfigOptions` convenience class to cudf-polars ([#18137](https://github.com/rapidsai/cudf/pull/18137)) [@rjzamora](https://github.com/rjzamora) -- Support new callback API for lazyframe.profile ([#18132](https://github.com/rapidsai/cudf/pull/18132)) [@wence-](https://github.com/wence-) -- Optimized compilation of CUDFTESTUTIL's interface sources ([#18131](https://github.com/rapidsai/cudf/pull/18131)) [@lamarrr](https://github.com/lamarrr) -- Unpin numpy<2.1 ([#18128](https://github.com/rapidsai/cudf/pull/18128)) [@mroeschke](https://github.com/mroeschke) -- Use cpu16 for build CI jobs ([#18124](https://github.com/rapidsai/cudf/pull/18124)) [@bdice](https://github.com/bdice) -- Remove now non-existent job ([#18123](https://github.com/rapidsai/cudf/pull/18123)) [@vyasr](https://github.com/vyasr) -- Minor typo fix in filling.pxd ([#18120](https://github.com/rapidsai/cudf/pull/18120)) [@davidwendt](https://github.com/davidwendt) -- Replace more deprecated `CUB` functors ([#18119](https://github.com/rapidsai/cudf/pull/18119)) [@miscco](https://github.com/miscco) -- Simplify DecimalDtype and DecimalColumn operations ([#18111](https://github.com/rapidsai/cudf/pull/18111)) [@mroeschke](https://github.com/mroeschke) -- Add interop support from arrow StringView to libcudf strings column ([#18107](https://github.com/rapidsai/cudf/pull/18107)) [@davidwendt](https://github.com/davidwendt) -- Expose the Number of Filtered Parquet Rowgroups (IO Metadata) to pylibcudf ([#18106](https://github.com/rapidsai/cudf/pull/18106)) [@JigaoLuo](https://github.com/JigaoLuo) -- Add a list of expected failures to narwhals tests ([#18097](https://github.com/rapidsai/cudf/pull/18097)) [@Matt711](https://github.com/Matt711) -- Remove unused var ([#18096](https://github.com/rapidsai/cudf/pull/18096)) [@vyasr](https://github.com/vyasr) -- Run narwhals tests nightly. ([#18093](https://github.com/rapidsai/cudf/pull/18093)) [@bdice](https://github.com/bdice) -- Use conda-build instead of conda-mambabuild ([#18092](https://github.com/rapidsai/cudf/pull/18092)) [@bdice](https://github.com/bdice) -- Remove static configure step ([#18091](https://github.com/rapidsai/cudf/pull/18091)) [@vyasr](https://github.com/vyasr) -- Remove `FindCUDAToolkit.cmake` from `.pre-commit-config.yaml` ([#18087](https://github.com/rapidsai/cudf/pull/18087)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Align StringColumn constructor with ColumnBase base class ([#18086](https://github.com/rapidsai/cudf/pull/18086)) [@mroeschke](https://github.com/mroeschke) -- Remove `FindCUDAToolkit` backport ([#18081](https://github.com/rapidsai/cudf/pull/18081)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Support melt(ignore_index=False) ([#18080](https://github.com/rapidsai/cudf/pull/18080)) [@mroeschke](https://github.com/mroeschke) -- Update numba dep and upper-bound numpy ([#18078](https://github.com/rapidsai/cudf/pull/18078)) [@vyasr](https://github.com/vyasr) -- Add `as_proxy_object` API to `cudf.pandas` ([#18072](https://github.com/rapidsai/cudf/pull/18072)) [@galipremsagar](https://github.com/galipremsagar) -- Enforce deprecation of dtype parameter in sum/product ([#18070](https://github.com/rapidsai/cudf/pull/18070)) [@mroeschke](https://github.com/mroeschke) -- send sccache logs to telemetry ([#18069](https://github.com/rapidsai/cudf/pull/18069)) [@msarahan](https://github.com/msarahan) -- Short circuit Index.equal if compared Index isn't same type ([#18067](https://github.com/rapidsai/cudf/pull/18067)) [@mroeschke](https://github.com/mroeschke) -- Make Column.view/can_cast_safely accept a dtype object ([#18066](https://github.com/rapidsai/cudf/pull/18066)) [@mroeschke](https://github.com/mroeschke) -- Optimization improvement for substr in cudf::string_view ([#18062](https://github.com/rapidsai/cudf/pull/18062)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-25.02 to branch-25.04 ([#18061](https://github.com/rapidsai/cudf/pull/18061)) [@bdice](https://github.com/bdice) -- Port all conda recipes to `rattler-build` ([#18054](https://github.com/rapidsai/cudf/pull/18054)) [@gforsyth](https://github.com/gforsyth) -- Minor improvements in arrow interop ([#18053](https://github.com/rapidsai/cudf/pull/18053)) [@wence-](https://github.com/wence-) -- Pass more dtype objects to `astype` calls ([#18044](https://github.com/rapidsai/cudf/pull/18044)) [@mroeschke](https://github.com/mroeschke) -- Forward merge branch-25.02 to branch-25.04 ([#18041](https://github.com/rapidsai/cudf/pull/18041)) [@Matt711](https://github.com/Matt711) -- Replace deprecated CCCL features ([#18036](https://github.com/rapidsai/cudf/pull/18036)) [@miscco](https://github.com/miscco) -- Separate stats filtering helpers to reuse in page pruning ([#18034](https://github.com/rapidsai/cudf/pull/18034)) [@mhaseeb123](https://github.com/mhaseeb123) -- Update spark-rapids-jni CI image version to cuda12.8.0 ([#18024](https://github.com/rapidsai/cudf/pull/18024)) [@pxLi](https://github.com/pxLi) -- Add pylibcudf.Scalar.from_numpy for bool/int/float/str types ([#18020](https://github.com/rapidsai/cudf/pull/18020)) [@mroeschke](https://github.com/mroeschke) -- Support IntervalDtype(subtype=None) ([#18017](https://github.com/rapidsai/cudf/pull/18017)) [@mroeschke](https://github.com/mroeschke) -- Enable pytest-xdist runs for py-polars tests ([#18016](https://github.com/rapidsai/cudf/pull/18016)) [@galipremsagar](https://github.com/galipremsagar) -- consolidate more conda solves in CI ([#18014](https://github.com/rapidsai/cudf/pull/18014)) [@jameslamb](https://github.com/jameslamb) -- Replace `cub::Int2Type` with `cuda::std::integral_constant` ([#18013](https://github.com/rapidsai/cudf/pull/18013)) [@miscco](https://github.com/miscco) -- Remove deprecated single component datetime extract APIs ([#18010](https://github.com/rapidsai/cudf/pull/18010)) [@Matt711](https://github.com/Matt711) -- Pass dtype objects to Column.astype ([#18008](https://github.com/rapidsai/cudf/pull/18008)) [@mroeschke](https://github.com/mroeschke) -- Require CMake 3.30.4 ([#18007](https://github.com/rapidsai/cudf/pull/18007)) [@robertmaynard](https://github.com/robertmaynard) -- Refactor math_ops.cu dispatcher logic ([#18006](https://github.com/rapidsai/cudf/pull/18006)) [@davidwendt](https://github.com/davidwendt) -- Move cudf::lists::detail::make_empty_lists_column to public API ([#17996](https://github.com/rapidsai/cudf/pull/17996)) [@davidwendt](https://github.com/davidwendt) -- Create Conda CI test env in one step ([#17995](https://github.com/rapidsai/cudf/pull/17995)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Add seed parameter to cudf hash_character_ngrams ([#17994](https://github.com/rapidsai/cudf/pull/17994)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated rolling window functionality ([#17993](https://github.com/rapidsai/cudf/pull/17993)) [@wence-](https://github.com/wence-) -- Continue on failures in cudf.pandas integration tests CI job ([#17987](https://github.com/rapidsai/cudf/pull/17987)) [@Matt711](https://github.com/Matt711) -- Avoid cudf.dtype calls in build_column/column_empty/.where ([#17979](https://github.com/rapidsai/cudf/pull/17979)) [@mroeschke](https://github.com/mroeschke) -- Ensure dtype objects are passed within Column.astype ([#17978](https://github.com/rapidsai/cudf/pull/17978)) [@mroeschke](https://github.com/mroeschke) -- Use Conda XGBoost ([#17959](https://github.com/rapidsai/cudf/pull/17959)) [@jakirkham](https://github.com/jakirkham) -- Read the footers in parallel when reading multiple Parquet files ([#17957](https://github.com/rapidsai/cudf/pull/17957)) [@vuule](https://github.com/vuule) -- Refactor predicate pushdown to reuse row group pruning in experimental PQ reader ([#17946](https://github.com/rapidsai/cudf/pull/17946)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add new nvtext tokenized minhash API ([#17944](https://github.com/rapidsai/cudf/pull/17944)) [@davidwendt](https://github.com/davidwendt) -- Use shared-workflows branch-25.04 ([#17943](https://github.com/rapidsai/cudf/pull/17943)) [@bdice](https://github.com/bdice) -- Get rid of the deprecated `thrust::identity` ([#17942](https://github.com/rapidsai/cudf/pull/17942)) [@PointKernel](https://github.com/PointKernel) -- Remove deprecated nvtext::minhash_permuted APIs ([#17939](https://github.com/rapidsai/cudf/pull/17939)) [@davidwendt](https://github.com/davidwendt) -- Enable third party library integration tests in CI with `cudf.pandas` ([#17936](https://github.com/rapidsai/cudf/pull/17936)) [@galipremsagar](https://github.com/galipremsagar) -- Add build_type input field for `test.yaml` ([#17925](https://github.com/rapidsai/cudf/pull/17925)) [@gforsyth](https://github.com/gforsyth) -- Remove cudf.Scalar from shift/fillna ([#17922](https://github.com/rapidsai/cudf/pull/17922)) [@mroeschke](https://github.com/mroeschke) -- Enabling `cross` join in `cudf` python ([#17921](https://github.com/rapidsai/cudf/pull/17921)) [@galipremsagar](https://github.com/galipremsagar) -- Use `rapids-pip-retry` in CI jobs that might need retries ([#17920](https://github.com/rapidsai/cudf/pull/17920)) [@gforsyth](https://github.com/gforsyth) -- More avoid cudf.dtype internally in favor of pre-defined, supported types ([#17918](https://github.com/rapidsai/cudf/pull/17918)) [@mroeschke](https://github.com/mroeschke) -- Initialize inout parameter ([#17911](https://github.com/rapidsai/cudf/pull/17911)) [@miscco](https://github.com/miscco) -- Remove dataframe protocol ([#17909](https://github.com/rapidsai/cudf/pull/17909)) [@vyasr](https://github.com/vyasr) -- Rename PascalCase functions and types to to snake_case to improve consistency ([#17908](https://github.com/rapidsai/cudf/pull/17908)) [@vuule](https://github.com/vuule) -- Use new rapids-logger library ([#17899](https://github.com/rapidsai/cudf/pull/17899)) [@vyasr](https://github.com/vyasr) -- Add `pylibcudf.Scalar.from_py` for construction from Python strings, bool, int, float ([#17898](https://github.com/rapidsai/cudf/pull/17898)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf.Scalar from factorize ([#17897](https://github.com/rapidsai/cudf/pull/17897)) [@mroeschke](https://github.com/mroeschke) -- disallow fallback to Make in Python builds ([#17894](https://github.com/rapidsai/cudf/pull/17894)) [@jameslamb](https://github.com/jameslamb) -- Remove `orc::gpu` namespace ([#17891](https://github.com/rapidsai/cudf/pull/17891)) [@vuule](https://github.com/vuule) -- Only run Auto Assign PR workflow if PR is not merged ([#17888](https://github.com/rapidsai/cudf/pull/17888)) [@mroeschke](https://github.com/mroeschke) -- Update pre-commit-hooks to version 0.6.0 ([#17887](https://github.com/rapidsai/cudf/pull/17887)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Forward-merge branch-25.02 to branch-25.04 ([#17885](https://github.com/rapidsai/cudf/pull/17885)) [@bdice](https://github.com/bdice) -- Add script to run pylibcudf tests ([#17882](https://github.com/rapidsai/cudf/pull/17882)) [@bdice](https://github.com/bdice) -- Migrate to NVKS for amd64 CI runners ([#17877](https://github.com/rapidsai/cudf/pull/17877)) [@bdice](https://github.com/bdice) -- Fix merge conflict for branch-25.02 into branch-25.04 ([#17874](https://github.com/rapidsai/cudf/pull/17874)) [@davidwendt](https://github.com/davidwendt) -- Remove decimal32/64 to decimal128 conversion in Parquet writer ([#17869](https://github.com/rapidsai/cudf/pull/17869)) [@mhaseeb123](https://github.com/mhaseeb123) -- Expose JSON reader options to builder in pylibcudf ([#17866](https://github.com/rapidsai/cudf/pull/17866)) [@shrshi](https://github.com/shrshi) -- Remove cudf.Scalar from .dt timedelta properties ([#17863](https://github.com/rapidsai/cudf/pull/17863)) [@mroeschke](https://github.com/mroeschke) -- Added support for custom types in PTX parser ([#17861](https://github.com/rapidsai/cudf/pull/17861)) [@lamarrr](https://github.com/lamarrr) -- Remove cudf.Scalar from date_range/to_datetime ([#17860](https://github.com/rapidsai/cudf/pull/17860)) [@mroeschke](https://github.com/mroeschke) -- Avoid `cudf.dtype` internally in favor of pre-defined, supported types ([#17839](https://github.com/rapidsai/cudf/pull/17839)) [@mroeschke](https://github.com/mroeschke) -- Allow cudf::type_to_id<T const>() ([#17831](https://github.com/rapidsai/cudf/pull/17831)) [@esoha-nvidia](https://github.com/esoha-nvidia) -- Fixing auto-merge branch-25.02 into branch-25.04 ([#17828](https://github.com/rapidsai/cudf/pull/17828)) [@davidwendt](https://github.com/davidwendt) -- Add new nvtext::normalize_characters API ([#17818](https://github.com/rapidsai/cudf/pull/17818)) [@davidwendt](https://github.com/davidwendt) -- Include more information in error messages in the nvcomp adapter ([#17814](https://github.com/rapidsai/cudf/pull/17814)) [@vuule](https://github.com/vuule) -- Extend and simplify API for calculation of range-based rolling window offsets ([#17807](https://github.com/rapidsai/cudf/pull/17807)) [@wence-](https://github.com/wence-) -- More minor fixes for CCCL ([#17793](https://github.com/rapidsai/cudf/pull/17793)) [@miscco](https://github.com/miscco) -- Use KvikIO to enable file's fast host read and host write ([#17764](https://github.com/rapidsai/cudf/pull/17764)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Remove cudf._lib.column in favor of pylibcudf. ([#17760](https://github.com/rapidsai/cudf/pull/17760)) [@mroeschke](https://github.com/mroeschke) -- Replaced std::string with std::string_view and removed excessive copies in cudf::io ([#17734](https://github.com/rapidsai/cudf/pull/17734)) [@lamarrr](https://github.com/lamarrr) -- Use xdist worksteal on the `cudf.pandas` test suite ([#16930](https://github.com/rapidsai/cudf/pull/16930)) [@Matt711](https://github.com/Matt711) +- Use pyarrow 15 in oldest dependency CI jobs ([#18409](https://github.com/NVIDIA/cudf/pull/18409)) [@bdice](https://github.com/bdice) +- Bump librdkafka to 2.8.0 ([#18370](https://github.com/NVIDIA/cudf/pull/18370)) [@raydouglass](https://github.com/raydouglass) +- fix(rattler): ignore `libzlib` run dependency to avoid `pandoc` collision ([#18368](https://github.com/NVIDIA/cudf/pull/18368)) [@gforsyth](https://github.com/gforsyth) +- Fix zstd build interface include definition ([#18366](https://github.com/NVIDIA/cudf/pull/18366)) [@trxcllnt](https://github.com/trxcllnt) +- test: Install pytest-env and hypothesis in test_narwhals.sh ([#18337](https://github.com/NVIDIA/cudf/pull/18337)) [@MarcoGorelli](https://github.com/MarcoGorelli) +- Remove unused `group_range_rolling_window` API ([#18313](https://github.com/NVIDIA/cudf/pull/18313)) [@wence-](https://github.com/wence-) +- Cache column view creation from arrow types ([#18302](https://github.com/NVIDIA/cudf/pull/18302)) [@vyasr](https://github.com/vyasr) +- Split Narwhals cudf.pandas tests failures into to fix and to skip ([#18267](https://github.com/NVIDIA/cudf/pull/18267)) [@mroeschke](https://github.com/mroeschke) +- Support BinOp, min, and max Aggregations in cudf-polars parallel groupby ([#18266](https://github.com/NVIDIA/cudf/pull/18266)) [@TomAugspurger](https://github.com/TomAugspurger) +- Minor clean up and optimizations in the Parquet writer ([#18258](https://github.com/NVIDIA/cudf/pull/18258)) [@vuule](https://github.com/vuule) +- Fix `cudf_kafka` run export for `cudatoolkit` ([#18245](https://github.com/NVIDIA/cudf/pull/18245)) [@gforsyth](https://github.com/gforsyth) +- dask-polars: use splat everywhere. ([#18243](https://github.com/NVIDIA/cudf/pull/18243)) [@madsbk](https://github.com/madsbk) +- Remove cudf.Scalar from binops ([#18240](https://github.com/NVIDIA/cudf/pull/18240)) [@mroeschke](https://github.com/mroeschke) +- Remove warning in the stream pool when asking for more streams than available ([#18236](https://github.com/NVIDIA/cudf/pull/18236)) [@vuule](https://github.com/vuule) +- Explain why we disable parallelism for profiler tests to avoid pytest-cov issue ([#18234](https://github.com/NVIDIA/cudf/pull/18234)) [@Matt711](https://github.com/Matt711) +- Ignore `cudatoolkit` run exports by name, not package ([#18230](https://github.com/NVIDIA/cudf/pull/18230)) [@gforsyth](https://github.com/gforsyth) +- Revert "Bump nightly check limit" ([#18227](https://github.com/NVIDIA/cudf/pull/18227)) [@Matt711](https://github.com/Matt711) +- Fix `cudf.pandas` to be able to work on a cpu-only machine ([#18224](https://github.com/NVIDIA/cudf/pull/18224)) [@galipremsagar](https://github.com/galipremsagar) +- Add missing `cudatoolkit` run_export ignore to `pylibcudf` ([#18223](https://github.com/NVIDIA/cudf/pull/18223)) [@gforsyth](https://github.com/gforsyth) +- Remove cudf.Scalar from Column.__setitem__ ([#18221](https://github.com/NVIDIA/cudf/pull/18221)) [@mroeschke](https://github.com/mroeschke) +- Remove unused round_up_pow2 utility ([#18218](https://github.com/NVIDIA/cudf/pull/18218)) [@PointKernel](https://github.com/PointKernel) +- Add flake8-print/debugger Ruff rules ([#18217](https://github.com/NVIDIA/cudf/pull/18217)) [@mroeschke](https://github.com/mroeschke) +- Bump polars version to <1.25 ([#18209](https://github.com/NVIDIA/cudf/pull/18209)) [@Matt711](https://github.com/Matt711) +- Export RAPIDS_ARTIFACTS_DIR. ([#18208](https://github.com/NVIDIA/cudf/pull/18208)) [@bdice](https://github.com/bdice) +- Drop more thrust functions with libcu++ ones ([#18207](https://github.com/NVIDIA/cudf/pull/18207)) [@miscco](https://github.com/miscco) +- Update Numpy <2.1 unpinning xfail condition ([#18203](https://github.com/NVIDIA/cudf/pull/18203)) [@mroeschke](https://github.com/mroeschke) +- Run conda import tests on Python packages ([#18197](https://github.com/NVIDIA/cudf/pull/18197)) [@bdice](https://github.com/bdice) +- fix(rattler): add `cudatoolkit` ignore run export to `cudf` ([#18195](https://github.com/NVIDIA/cudf/pull/18195)) [@gforsyth](https://github.com/gforsyth) +- Revert "Disable ARM CI in C++ and Python test CI jobs" ([#18188](https://github.com/NVIDIA/cudf/pull/18188)) [@Matt711](https://github.com/Matt711) +- Define Column.where to be used across DataFrame/Series ([#18186](https://github.com/NVIDIA/cudf/pull/18186)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf.Scalar in where ([#18178](https://github.com/NVIDIA/cudf/pull/18178)) [@mroeschke](https://github.com/mroeschke) +- Drop unnecessary fmt dep ([#18177](https://github.com/NVIDIA/cudf/pull/18177)) [@vyasr](https://github.com/vyasr) +- Refactor join internals: separate hash_join declaration and cleanup ([#18170](https://github.com/NVIDIA/cudf/pull/18170)) [@PointKernel](https://github.com/PointKernel) +- Add Ruff rule to enforce cudf dtype utils over numpy/pandas dtype utils ([#18169](https://github.com/NVIDIA/cudf/pull/18169)) [@mroeschke](https://github.com/mroeschke) +- Combine multiple str.minhash() APIs into one call ([#18168](https://github.com/NVIDIA/cudf/pull/18168)) [@davidwendt](https://github.com/davidwendt) +- Move nanoarrow_utils.hpp from cpp/tests/interop to cpp/include/cudf_test ([#18163](https://github.com/NVIDIA/cudf/pull/18163)) [@davidwendt](https://github.com/davidwendt) +- Test cudf against the latest stable branch of Narwhals ([#18162](https://github.com/NVIDIA/cudf/pull/18162)) [@Matt711](https://github.com/Matt711) +- fix libcudf pins cu11 ([#18161](https://github.com/NVIDIA/cudf/pull/18161)) [@gforsyth](https://github.com/gforsyth) +- Combine separate ConfigureNVBench calls to fix cpp conda builds ([#18155](https://github.com/NVIDIA/cudf/pull/18155)) [@gforsyth](https://github.com/gforsyth) +- Add telemetry to build workflows ([#18154](https://github.com/NVIDIA/cudf/pull/18154)) [@gforsyth](https://github.com/gforsyth) +- Prune more seldom used dtype utils ([#18150](https://github.com/NVIDIA/cudf/pull/18150)) [@mroeschke](https://github.com/mroeschke) +- Remove some unnecessary module imports ([#18143](https://github.com/NVIDIA/cudf/pull/18143)) [@mroeschke](https://github.com/mroeschke) +- Branch 25.04 merge branch 25.02 ([#18142](https://github.com/NVIDIA/cudf/pull/18142)) [@vyasr](https://github.com/vyasr) +- Prune some seldom used dtype utils ([#18141](https://github.com/NVIDIA/cudf/pull/18141)) [@mroeschke](https://github.com/mroeschke) +- Use more, cheaper dtype checking utilities in cudf Python ([#18139](https://github.com/NVIDIA/cudf/pull/18139)) [@mroeschke](https://github.com/mroeschke) +- Support deserializing cudf-polars objects composed of RMM frames ([#18138](https://github.com/NVIDIA/cudf/pull/18138)) [@pentschev](https://github.com/pentschev) +- Add `ConfigOptions` convenience class to cudf-polars ([#18137](https://github.com/NVIDIA/cudf/pull/18137)) [@rjzamora](https://github.com/rjzamora) +- Support new callback API for lazyframe.profile ([#18132](https://github.com/NVIDIA/cudf/pull/18132)) [@wence-](https://github.com/wence-) +- Optimized compilation of CUDFTESTUTIL's interface sources ([#18131](https://github.com/NVIDIA/cudf/pull/18131)) [@lamarrr](https://github.com/lamarrr) +- Unpin numpy<2.1 ([#18128](https://github.com/NVIDIA/cudf/pull/18128)) [@mroeschke](https://github.com/mroeschke) +- Use cpu16 for build CI jobs ([#18124](https://github.com/NVIDIA/cudf/pull/18124)) [@bdice](https://github.com/bdice) +- Remove now non-existent job ([#18123](https://github.com/NVIDIA/cudf/pull/18123)) [@vyasr](https://github.com/vyasr) +- Minor typo fix in filling.pxd ([#18120](https://github.com/NVIDIA/cudf/pull/18120)) [@davidwendt](https://github.com/davidwendt) +- Replace more deprecated `CUB` functors ([#18119](https://github.com/NVIDIA/cudf/pull/18119)) [@miscco](https://github.com/miscco) +- Simplify DecimalDtype and DecimalColumn operations ([#18111](https://github.com/NVIDIA/cudf/pull/18111)) [@mroeschke](https://github.com/mroeschke) +- Add interop support from arrow StringView to libcudf strings column ([#18107](https://github.com/NVIDIA/cudf/pull/18107)) [@davidwendt](https://github.com/davidwendt) +- Expose the Number of Filtered Parquet Rowgroups (IO Metadata) to pylibcudf ([#18106](https://github.com/NVIDIA/cudf/pull/18106)) [@JigaoLuo](https://github.com/JigaoLuo) +- Add a list of expected failures to narwhals tests ([#18097](https://github.com/NVIDIA/cudf/pull/18097)) [@Matt711](https://github.com/Matt711) +- Remove unused var ([#18096](https://github.com/NVIDIA/cudf/pull/18096)) [@vyasr](https://github.com/vyasr) +- Run narwhals tests nightly. ([#18093](https://github.com/NVIDIA/cudf/pull/18093)) [@bdice](https://github.com/bdice) +- Use conda-build instead of conda-mambabuild ([#18092](https://github.com/NVIDIA/cudf/pull/18092)) [@bdice](https://github.com/bdice) +- Remove static configure step ([#18091](https://github.com/NVIDIA/cudf/pull/18091)) [@vyasr](https://github.com/vyasr) +- Remove `FindCUDAToolkit.cmake` from `.pre-commit-config.yaml` ([#18087](https://github.com/NVIDIA/cudf/pull/18087)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Align StringColumn constructor with ColumnBase base class ([#18086](https://github.com/NVIDIA/cudf/pull/18086)) [@mroeschke](https://github.com/mroeschke) +- Remove `FindCUDAToolkit` backport ([#18081](https://github.com/NVIDIA/cudf/pull/18081)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Support melt(ignore_index=False) ([#18080](https://github.com/NVIDIA/cudf/pull/18080)) [@mroeschke](https://github.com/mroeschke) +- Update numba dep and upper-bound numpy ([#18078](https://github.com/NVIDIA/cudf/pull/18078)) [@vyasr](https://github.com/vyasr) +- Add `as_proxy_object` API to `cudf.pandas` ([#18072](https://github.com/NVIDIA/cudf/pull/18072)) [@galipremsagar](https://github.com/galipremsagar) +- Enforce deprecation of dtype parameter in sum/product ([#18070](https://github.com/NVIDIA/cudf/pull/18070)) [@mroeschke](https://github.com/mroeschke) +- send sccache logs to telemetry ([#18069](https://github.com/NVIDIA/cudf/pull/18069)) [@msarahan](https://github.com/msarahan) +- Short circuit Index.equal if compared Index isn't same type ([#18067](https://github.com/NVIDIA/cudf/pull/18067)) [@mroeschke](https://github.com/mroeschke) +- Make Column.view/can_cast_safely accept a dtype object ([#18066](https://github.com/NVIDIA/cudf/pull/18066)) [@mroeschke](https://github.com/mroeschke) +- Optimization improvement for substr in cudf::string_view ([#18062](https://github.com/NVIDIA/cudf/pull/18062)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-25.02 to branch-25.04 ([#18061](https://github.com/NVIDIA/cudf/pull/18061)) [@bdice](https://github.com/bdice) +- Port all conda recipes to `rattler-build` ([#18054](https://github.com/NVIDIA/cudf/pull/18054)) [@gforsyth](https://github.com/gforsyth) +- Minor improvements in arrow interop ([#18053](https://github.com/NVIDIA/cudf/pull/18053)) [@wence-](https://github.com/wence-) +- Pass more dtype objects to `astype` calls ([#18044](https://github.com/NVIDIA/cudf/pull/18044)) [@mroeschke](https://github.com/mroeschke) +- Forward merge branch-25.02 to branch-25.04 ([#18041](https://github.com/NVIDIA/cudf/pull/18041)) [@Matt711](https://github.com/Matt711) +- Replace deprecated CCCL features ([#18036](https://github.com/NVIDIA/cudf/pull/18036)) [@miscco](https://github.com/miscco) +- Separate stats filtering helpers to reuse in page pruning ([#18034](https://github.com/NVIDIA/cudf/pull/18034)) [@mhaseeb123](https://github.com/mhaseeb123) +- Update spark-rapids-jni CI image version to cuda12.8.0 ([#18024](https://github.com/NVIDIA/cudf/pull/18024)) [@pxLi](https://github.com/pxLi) +- Add pylibcudf.Scalar.from_numpy for bool/int/float/str types ([#18020](https://github.com/NVIDIA/cudf/pull/18020)) [@mroeschke](https://github.com/mroeschke) +- Support IntervalDtype(subtype=None) ([#18017](https://github.com/NVIDIA/cudf/pull/18017)) [@mroeschke](https://github.com/mroeschke) +- Enable pytest-xdist runs for py-polars tests ([#18016](https://github.com/NVIDIA/cudf/pull/18016)) [@galipremsagar](https://github.com/galipremsagar) +- consolidate more conda solves in CI ([#18014](https://github.com/NVIDIA/cudf/pull/18014)) [@jameslamb](https://github.com/jameslamb) +- Replace `cub::Int2Type` with `cuda::std::integral_constant` ([#18013](https://github.com/NVIDIA/cudf/pull/18013)) [@miscco](https://github.com/miscco) +- Remove deprecated single component datetime extract APIs ([#18010](https://github.com/NVIDIA/cudf/pull/18010)) [@Matt711](https://github.com/Matt711) +- Pass dtype objects to Column.astype ([#18008](https://github.com/NVIDIA/cudf/pull/18008)) [@mroeschke](https://github.com/mroeschke) +- Require CMake 3.30.4 ([#18007](https://github.com/NVIDIA/cudf/pull/18007)) [@robertmaynard](https://github.com/robertmaynard) +- Refactor math_ops.cu dispatcher logic ([#18006](https://github.com/NVIDIA/cudf/pull/18006)) [@davidwendt](https://github.com/davidwendt) +- Move cudf::lists::detail::make_empty_lists_column to public API ([#17996](https://github.com/NVIDIA/cudf/pull/17996)) [@davidwendt](https://github.com/davidwendt) +- Create Conda CI test env in one step ([#17995](https://github.com/NVIDIA/cudf/pull/17995)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Add seed parameter to cudf hash_character_ngrams ([#17994](https://github.com/NVIDIA/cudf/pull/17994)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated rolling window functionality ([#17993](https://github.com/NVIDIA/cudf/pull/17993)) [@wence-](https://github.com/wence-) +- Continue on failures in cudf.pandas integration tests CI job ([#17987](https://github.com/NVIDIA/cudf/pull/17987)) [@Matt711](https://github.com/Matt711) +- Avoid cudf.dtype calls in build_column/column_empty/.where ([#17979](https://github.com/NVIDIA/cudf/pull/17979)) [@mroeschke](https://github.com/mroeschke) +- Ensure dtype objects are passed within Column.astype ([#17978](https://github.com/NVIDIA/cudf/pull/17978)) [@mroeschke](https://github.com/mroeschke) +- Use Conda XGBoost ([#17959](https://github.com/NVIDIA/cudf/pull/17959)) [@jakirkham](https://github.com/jakirkham) +- Read the footers in parallel when reading multiple Parquet files ([#17957](https://github.com/NVIDIA/cudf/pull/17957)) [@vuule](https://github.com/vuule) +- Refactor predicate pushdown to reuse row group pruning in experimental PQ reader ([#17946](https://github.com/NVIDIA/cudf/pull/17946)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add new nvtext tokenized minhash API ([#17944](https://github.com/NVIDIA/cudf/pull/17944)) [@davidwendt](https://github.com/davidwendt) +- Use shared-workflows branch-25.04 ([#17943](https://github.com/NVIDIA/cudf/pull/17943)) [@bdice](https://github.com/bdice) +- Get rid of the deprecated `thrust::identity` ([#17942](https://github.com/NVIDIA/cudf/pull/17942)) [@PointKernel](https://github.com/PointKernel) +- Remove deprecated nvtext::minhash_permuted APIs ([#17939](https://github.com/NVIDIA/cudf/pull/17939)) [@davidwendt](https://github.com/davidwendt) +- Enable third party library integration tests in CI with `cudf.pandas` ([#17936](https://github.com/NVIDIA/cudf/pull/17936)) [@galipremsagar](https://github.com/galipremsagar) +- Add build_type input field for `test.yaml` ([#17925](https://github.com/NVIDIA/cudf/pull/17925)) [@gforsyth](https://github.com/gforsyth) +- Remove cudf.Scalar from shift/fillna ([#17922](https://github.com/NVIDIA/cudf/pull/17922)) [@mroeschke](https://github.com/mroeschke) +- Enabling `cross` join in `cudf` python ([#17921](https://github.com/NVIDIA/cudf/pull/17921)) [@galipremsagar](https://github.com/galipremsagar) +- Use `rapids-pip-retry` in CI jobs that might need retries ([#17920](https://github.com/NVIDIA/cudf/pull/17920)) [@gforsyth](https://github.com/gforsyth) +- More avoid cudf.dtype internally in favor of pre-defined, supported types ([#17918](https://github.com/NVIDIA/cudf/pull/17918)) [@mroeschke](https://github.com/mroeschke) +- Initialize inout parameter ([#17911](https://github.com/NVIDIA/cudf/pull/17911)) [@miscco](https://github.com/miscco) +- Remove dataframe protocol ([#17909](https://github.com/NVIDIA/cudf/pull/17909)) [@vyasr](https://github.com/vyasr) +- Rename PascalCase functions and types to to snake_case to improve consistency ([#17908](https://github.com/NVIDIA/cudf/pull/17908)) [@vuule](https://github.com/vuule) +- Use new rapids-logger library ([#17899](https://github.com/NVIDIA/cudf/pull/17899)) [@vyasr](https://github.com/vyasr) +- Add `pylibcudf.Scalar.from_py` for construction from Python strings, bool, int, float ([#17898](https://github.com/NVIDIA/cudf/pull/17898)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf.Scalar from factorize ([#17897](https://github.com/NVIDIA/cudf/pull/17897)) [@mroeschke](https://github.com/mroeschke) +- disallow fallback to Make in Python builds ([#17894](https://github.com/NVIDIA/cudf/pull/17894)) [@jameslamb](https://github.com/jameslamb) +- Remove `orc::gpu` namespace ([#17891](https://github.com/NVIDIA/cudf/pull/17891)) [@vuule](https://github.com/vuule) +- Only run Auto Assign PR workflow if PR is not merged ([#17888](https://github.com/NVIDIA/cudf/pull/17888)) [@mroeschke](https://github.com/mroeschke) +- Update pre-commit-hooks to version 0.6.0 ([#17887](https://github.com/NVIDIA/cudf/pull/17887)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Forward-merge branch-25.02 to branch-25.04 ([#17885](https://github.com/NVIDIA/cudf/pull/17885)) [@bdice](https://github.com/bdice) +- Add script to run pylibcudf tests ([#17882](https://github.com/NVIDIA/cudf/pull/17882)) [@bdice](https://github.com/bdice) +- Migrate to NVKS for amd64 CI runners ([#17877](https://github.com/NVIDIA/cudf/pull/17877)) [@bdice](https://github.com/bdice) +- Fix merge conflict for branch-25.02 into branch-25.04 ([#17874](https://github.com/NVIDIA/cudf/pull/17874)) [@davidwendt](https://github.com/davidwendt) +- Remove decimal32/64 to decimal128 conversion in Parquet writer ([#17869](https://github.com/NVIDIA/cudf/pull/17869)) [@mhaseeb123](https://github.com/mhaseeb123) +- Expose JSON reader options to builder in pylibcudf ([#17866](https://github.com/NVIDIA/cudf/pull/17866)) [@shrshi](https://github.com/shrshi) +- Remove cudf.Scalar from .dt timedelta properties ([#17863](https://github.com/NVIDIA/cudf/pull/17863)) [@mroeschke](https://github.com/mroeschke) +- Added support for custom types in PTX parser ([#17861](https://github.com/NVIDIA/cudf/pull/17861)) [@lamarrr](https://github.com/lamarrr) +- Remove cudf.Scalar from date_range/to_datetime ([#17860](https://github.com/NVIDIA/cudf/pull/17860)) [@mroeschke](https://github.com/mroeschke) +- Avoid `cudf.dtype` internally in favor of pre-defined, supported types ([#17839](https://github.com/NVIDIA/cudf/pull/17839)) [@mroeschke](https://github.com/mroeschke) +- Allow cudf::type_to_id<T const>() ([#17831](https://github.com/NVIDIA/cudf/pull/17831)) [@esoha-nvidia](https://github.com/esoha-nvidia) +- Fixing auto-merge branch-25.02 into branch-25.04 ([#17828](https://github.com/NVIDIA/cudf/pull/17828)) [@davidwendt](https://github.com/davidwendt) +- Add new nvtext::normalize_characters API ([#17818](https://github.com/NVIDIA/cudf/pull/17818)) [@davidwendt](https://github.com/davidwendt) +- Include more information in error messages in the nvcomp adapter ([#17814](https://github.com/NVIDIA/cudf/pull/17814)) [@vuule](https://github.com/vuule) +- Extend and simplify API for calculation of range-based rolling window offsets ([#17807](https://github.com/NVIDIA/cudf/pull/17807)) [@wence-](https://github.com/wence-) +- More minor fixes for CCCL ([#17793](https://github.com/NVIDIA/cudf/pull/17793)) [@miscco](https://github.com/miscco) +- Use KvikIO to enable file's fast host read and host write ([#17764](https://github.com/NVIDIA/cudf/pull/17764)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Remove cudf._lib.column in favor of pylibcudf. ([#17760](https://github.com/NVIDIA/cudf/pull/17760)) [@mroeschke](https://github.com/mroeschke) +- Replaced std::string with std::string_view and removed excessive copies in cudf::io ([#17734](https://github.com/NVIDIA/cudf/pull/17734)) [@lamarrr](https://github.com/lamarrr) +- Use xdist worksteal on the `cudf.pandas` test suite ([#16930](https://github.com/NVIDIA/cudf/pull/16930)) [@Matt711](https://github.com/Matt711) # cudf 25.02.00 (13 Feb 2025) ## 🚨 Breaking Changes -- Expose stream-ordering in scalar and avro APIs ([#17766](https://github.com/rapidsai/cudf/pull/17766)) [@shrshi](https://github.com/shrshi) -- Add seed parameter to hash_character_ngrams ([#17643](https://github.com/rapidsai/cudf/pull/17643)) [@davidwendt](https://github.com/davidwendt) -- Performance improvements and simplifications for fixed size row-based rolling windows ([#17623](https://github.com/rapidsai/cudf/pull/17623)) [@wence-](https://github.com/wence-) -- Refactor distinct hash join to handle multiple probes with the same build table ([#17609](https://github.com/rapidsai/cudf/pull/17609)) [@PointKernel](https://github.com/PointKernel) -- Deprecate cudf::grouped_time_range_rolling_window ([#17589](https://github.com/rapidsai/cudf/pull/17589)) [@wence-](https://github.com/wence-) -- Remove "legacy" Dask DataFrame support from Dask cuDF ([#17558](https://github.com/rapidsai/cudf/pull/17558)) [@rjzamora](https://github.com/rjzamora) -- Return empty result for segmented_reduce if input and offsets are both empty ([#17437](https://github.com/rapidsai/cudf/pull/17437)) [@davidwendt](https://github.com/davidwendt) -- Rework minhash APIs for deprecation cycle ([#17421](https://github.com/rapidsai/cudf/pull/17421)) [@davidwendt](https://github.com/davidwendt) -- Change indices for dictionary column to signed integer type ([#17390](https://github.com/rapidsai/cudf/pull/17390)) [@davidwendt](https://github.com/davidwendt) +- Expose stream-ordering in scalar and avro APIs ([#17766](https://github.com/NVIDIA/cudf/pull/17766)) [@shrshi](https://github.com/shrshi) +- Add seed parameter to hash_character_ngrams ([#17643](https://github.com/NVIDIA/cudf/pull/17643)) [@davidwendt](https://github.com/davidwendt) +- Performance improvements and simplifications for fixed size row-based rolling windows ([#17623](https://github.com/NVIDIA/cudf/pull/17623)) [@wence-](https://github.com/wence-) +- Refactor distinct hash join to handle multiple probes with the same build table ([#17609](https://github.com/NVIDIA/cudf/pull/17609)) [@PointKernel](https://github.com/PointKernel) +- Deprecate cudf::grouped_time_range_rolling_window ([#17589](https://github.com/NVIDIA/cudf/pull/17589)) [@wence-](https://github.com/wence-) +- Remove "legacy" Dask DataFrame support from Dask cuDF ([#17558](https://github.com/NVIDIA/cudf/pull/17558)) [@rjzamora](https://github.com/rjzamora) +- Return empty result for segmented_reduce if input and offsets are both empty ([#17437](https://github.com/NVIDIA/cudf/pull/17437)) [@davidwendt](https://github.com/davidwendt) +- Rework minhash APIs for deprecation cycle ([#17421](https://github.com/NVIDIA/cudf/pull/17421)) [@davidwendt](https://github.com/davidwendt) +- Change indices for dictionary column to signed integer type ([#17390](https://github.com/NVIDIA/cudf/pull/17390)) [@davidwendt](https://github.com/davidwendt) ## 🐛 Bug Fixes -- Fix race check failures in shared memory groupby ([#17985](https://github.com/rapidsai/cudf/pull/17985)) [@PointKernel](https://github.com/PointKernel) -- Pin `ibis` version in the cudf.pandas integration tests <10.0.0 ([#17975](https://github.com/rapidsai/cudf/pull/17975)) [@Matt711](https://github.com/Matt711) -- Fix the index type in the indexing operator of the span types ([#17971](https://github.com/rapidsai/cudf/pull/17971)) [@vuule](https://github.com/vuule) -- Add missing pin ([#17915](https://github.com/rapidsai/cudf/pull/17915)) [@vyasr](https://github.com/vyasr) -- Fix third-party `cudf.pandas` tests ([#17900](https://github.com/rapidsai/cudf/pull/17900)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `numpy` data access by making attribute private ([#17890](https://github.com/rapidsai/cudf/pull/17890)) [@galipremsagar](https://github.com/galipremsagar) -- Remove extra local var declaration from cudf.pandas 3rd-party integration shell script ([#17886](https://github.com/rapidsai/cudf/pull/17886)) [@Matt711](https://github.com/Matt711) -- Move `isinstance_cudf_pandas` to `fast_slow_proxy` ([#17875](https://github.com/rapidsai/cudf/pull/17875)) [@galipremsagar](https://github.com/galipremsagar) -- Make `_Series_dtype` method a property ([#17854](https://github.com/rapidsai/cudf/pull/17854)) [@Matt711](https://github.com/Matt711) -- Fix the bug in determining the heuristics for shared memory groupby ([#17851](https://github.com/rapidsai/cudf/pull/17851)) [@PointKernel](https://github.com/PointKernel) -- Fix possible OOB mem access in Parquet decoder ([#17841](https://github.com/rapidsai/cudf/pull/17841)) [@mhaseeb123](https://github.com/mhaseeb123) -- Require batches to be non-empty in multi-batch JSON reader ([#17837](https://github.com/rapidsai/cudf/pull/17837)) [@shrshi](https://github.com/shrshi) -- Fix rolling(min_periods=) with int and null data with mode.pandas_compat ([#17822](https://github.com/rapidsai/cudf/pull/17822)) [@mroeschke](https://github.com/mroeschke) -- Resolve race-condition in `disable_module_accelerator` ([#17811](https://github.com/rapidsai/cudf/pull/17811)) [@galipremsagar](https://github.com/galipremsagar) -- Make Series(dtype=object) raise in mode.pandas_compat with non string data ([#17804](https://github.com/rapidsai/cudf/pull/17804)) [@mroeschke](https://github.com/mroeschke) -- Disable intended disabled ORC tests ([#17790](https://github.com/rapidsai/cudf/pull/17790)) [@davidwendt](https://github.com/davidwendt) -- Fix empty DataFrame construction not returning RangeIndex columns ([#17784](https://github.com/rapidsai/cudf/pull/17784)) [@mroeschke](https://github.com/mroeschke) -- Fix various `.str` methods for pandas compatability ([#17782](https://github.com/rapidsai/cudf/pull/17782)) [@mroeschke](https://github.com/mroeschke) -- Fix `count` API issue about ignoring nan values ([#17779](https://github.com/rapidsai/cudf/pull/17779)) [@galipremsagar](https://github.com/galipremsagar) -- Add `numba` pinning to `cudf` repo ([#17777](https://github.com/rapidsai/cudf/pull/17777)) [@galipremsagar](https://github.com/galipremsagar) -- Allow .sort_values(na_position=) to include NaNs in mode.pandas_compatible ([#17776](https://github.com/rapidsai/cudf/pull/17776)) [@mroeschke](https://github.com/mroeschke) -- allow deselecting nvcomp wheels ([#17774](https://github.com/rapidsai/cudf/pull/17774)) [@jameslamb](https://github.com/jameslamb) -- Use the `aligned_resource_adaptor` to allocate bloom filter device buffers ([#17758](https://github.com/rapidsai/cudf/pull/17758)) [@mhaseeb123](https://github.com/mhaseeb123) -- Avoid instantiating bloom filter query function for nested and bool types ([#17753](https://github.com/rapidsai/cudf/pull/17753)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix DataFrame.merge(Series, how="left"/"right") on column and index not resulting in a RangeIndex ([#17739](https://github.com/rapidsai/cudf/pull/17739)) [@mroeschke](https://github.com/mroeschke) -- [BUG] xfail Polars excel test ([#17731](https://github.com/rapidsai/cudf/pull/17731)) [@Matt711](https://github.com/Matt711) -- Require to implement `AutoCloseable` for the classes derived from `HostUDFWrapper` ([#17727](https://github.com/rapidsai/cudf/pull/17727)) [@ttnghia](https://github.com/ttnghia) -- Remove jlowe as a java committer since he retired ([#17725](https://github.com/rapidsai/cudf/pull/17725)) [@tgravescs](https://github.com/tgravescs) -- Prevent use of invalid grid sizes in ORC reader and writer ([#17709](https://github.com/rapidsai/cudf/pull/17709)) [@vuule](https://github.com/vuule) -- Enforce schema for partial tables in multi-source multi-batch JSON reader ([#17708](https://github.com/rapidsai/cudf/pull/17708)) [@shrshi](https://github.com/shrshi) -- Compute and use the initial string offset when building `nested` large string cols with chunked parquet reader ([#17702](https://github.com/rapidsai/cudf/pull/17702)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix writing of compressed ORC files with large stripe footers ([#17700](https://github.com/rapidsai/cudf/pull/17700)) [@vuule](https://github.com/vuule) -- Fix cudf.polars sum of empty not equalling zero ([#17685](https://github.com/rapidsai/cudf/pull/17685)) [@mroeschke](https://github.com/mroeschke) -- Fix formatting in logging ([#17680](https://github.com/rapidsai/cudf/pull/17680)) [@vuule](https://github.com/vuule) -- convert all nulls to nans in a specific scenario ([#17677](https://github.com/rapidsai/cudf/pull/17677)) [@galipremsagar](https://github.com/galipremsagar) -- Define cudf repr methods on the Column ([#17675](https://github.com/rapidsai/cudf/pull/17675)) [@mroeschke](https://github.com/mroeschke) -- Fix groupby.len with null values in cudf.polars ([#17671](https://github.com/rapidsai/cudf/pull/17671)) [@mroeschke](https://github.com/mroeschke) -- Fix: DataFrameGroupBy.get_group was raising with length>1 tuples ([#17653](https://github.com/rapidsai/cudf/pull/17653)) [@MarcoGorelli](https://github.com/MarcoGorelli) -- Fix possible int overflow in compute_mixed_join_output_size ([#17633](https://github.com/rapidsai/cudf/pull/17633)) [@davidwendt](https://github.com/davidwendt) -- Fix a minor potential i32 overflow in `thrust::transform_exclusive_scan` in PQ reader preprocessing ([#17617](https://github.com/rapidsai/cudf/pull/17617)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix failing xgboost test in the cudf.pandas third-party integration tests ([#17616](https://github.com/rapidsai/cudf/pull/17616)) [@Matt711](https://github.com/Matt711) -- Fix ``dask_cudf.read_csv`` ([#17612](https://github.com/rapidsai/cudf/pull/17612)) [@rjzamora](https://github.com/rjzamora) -- Fix memcheck error in ReplaceTest.NormalizeNansAndZerosMutable gtest ([#17610](https://github.com/rapidsai/cudf/pull/17610)) [@davidwendt](https://github.com/davidwendt) -- Correctly accept a `pandas.CategoricalDtype(pandas.IntervalDtype(...), ...)` type ([#17604](https://github.com/rapidsai/cudf/pull/17604)) [@mroeschke](https://github.com/mroeschke) -- Add ability to modify and propagate `names` of `columns` object ([#17597](https://github.com/rapidsai/cudf/pull/17597)) [@galipremsagar](https://github.com/galipremsagar) -- Ignore NaN correctly in .quantile ([#17593](https://github.com/rapidsai/cudf/pull/17593)) [@mroeschke](https://github.com/mroeschke) -- Fix groupby argmin/max gather of sorted-order indices ([#17591](https://github.com/rapidsai/cudf/pull/17591)) [@davidwendt](https://github.com/davidwendt) -- Fix ctest fail running libcudf tests in a Debug build ([#17576](https://github.com/rapidsai/cudf/pull/17576)) [@davidwendt](https://github.com/davidwendt) -- Specify a version for rapids_logger dependency ([#17573](https://github.com/rapidsai/cudf/pull/17573)) [@jlowe](https://github.com/jlowe) -- Fix the ORC decoding bug for the timestamp data ([#17570](https://github.com/rapidsai/cudf/pull/17570)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- [JNI] remove rmm argument to set rw access for fabric handles ([#17553](https://github.com/rapidsai/cudf/pull/17553)) [@abellina](https://github.com/abellina) -- Document undefined behavior in div_rounding_up_safe ([#17542](https://github.com/rapidsai/cudf/pull/17542)) [@davidwendt](https://github.com/davidwendt) -- Fix nvcc-imposed UB in `constexpr` functions ([#17534](https://github.com/rapidsai/cudf/pull/17534)) [@vuule](https://github.com/vuule) -- Add anonymous namespace to libcudf test source ([#17529](https://github.com/rapidsai/cudf/pull/17529)) [@davidwendt](https://github.com/davidwendt) -- Propagate failures in pandas integration tests and Skip failing tests ([#17521](https://github.com/rapidsai/cudf/pull/17521)) [@Matt711](https://github.com/Matt711) -- Fix libcudf compile error when logging is disabled ([#17512](https://github.com/rapidsai/cudf/pull/17512)) [@davidwendt](https://github.com/davidwendt) -- Fix Dask-cuDF `clip` APIs ([#17509](https://github.com/rapidsai/cudf/pull/17509)) [@rjzamora](https://github.com/rjzamora) -- Fix pylibcudf to_arrow with multiple nested data types ([#17504](https://github.com/rapidsai/cudf/pull/17504)) [@mroeschke](https://github.com/mroeschke) -- Fix groupby(as_index=False).size not reseting index ([#17499](https://github.com/rapidsai/cudf/pull/17499)) [@mroeschke](https://github.com/mroeschke) -- Revert "Temporarily skip tests due to dask/distributed#8953" ([#17492](https://github.com/rapidsai/cudf/pull/17492)) [@Matt711](https://github.com/Matt711) -- Workaround for a misaligned access in `read_csv` on some CUDA versions ([#17477](https://github.com/rapidsai/cudf/pull/17477)) [@vuule](https://github.com/vuule) -- Fix some possible thread-id overflow calculations ([#17473](https://github.com/rapidsai/cudf/pull/17473)) [@davidwendt](https://github.com/davidwendt) -- Temporarily skip tests due to dask/distributed#8953 ([#17472](https://github.com/rapidsai/cudf/pull/17472)) [@wence-](https://github.com/wence-) -- Detect mismatches in begin and end tokens returned by JSON tokenizer FST ([#17471](https://github.com/rapidsai/cudf/pull/17471)) [@shrshi](https://github.com/shrshi) -- Support dask>=2024.11.2 in Dask cuDF ([#17439](https://github.com/rapidsai/cudf/pull/17439)) [@rjzamora](https://github.com/rjzamora) -- Fix write_json failure for zero columns in table/struct ([#17414](https://github.com/rapidsai/cudf/pull/17414)) [@karthikeyann](https://github.com/karthikeyann) -- Fix Debug-mode failing Arrow test ([#17405](https://github.com/rapidsai/cudf/pull/17405)) [@zeroshade](https://github.com/zeroshade) -- Fix all null list column with missing child column in JSON reader ([#17348](https://github.com/rapidsai/cudf/pull/17348)) [@karthikeyann](https://github.com/karthikeyann) +- Fix race check failures in shared memory groupby ([#17985](https://github.com/NVIDIA/cudf/pull/17985)) [@PointKernel](https://github.com/PointKernel) +- Pin `ibis` version in the cudf.pandas integration tests <10.0.0 ([#17975](https://github.com/NVIDIA/cudf/pull/17975)) [@Matt711](https://github.com/Matt711) +- Fix the index type in the indexing operator of the span types ([#17971](https://github.com/NVIDIA/cudf/pull/17971)) [@vuule](https://github.com/vuule) +- Add missing pin ([#17915](https://github.com/NVIDIA/cudf/pull/17915)) [@vyasr](https://github.com/vyasr) +- Fix third-party `cudf.pandas` tests ([#17900](https://github.com/NVIDIA/cudf/pull/17900)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `numpy` data access by making attribute private ([#17890](https://github.com/NVIDIA/cudf/pull/17890)) [@galipremsagar](https://github.com/galipremsagar) +- Remove extra local var declaration from cudf.pandas 3rd-party integration shell script ([#17886](https://github.com/NVIDIA/cudf/pull/17886)) [@Matt711](https://github.com/Matt711) +- Move `isinstance_cudf_pandas` to `fast_slow_proxy` ([#17875](https://github.com/NVIDIA/cudf/pull/17875)) [@galipremsagar](https://github.com/galipremsagar) +- Make `_Series_dtype` method a property ([#17854](https://github.com/NVIDIA/cudf/pull/17854)) [@Matt711](https://github.com/Matt711) +- Fix the bug in determining the heuristics for shared memory groupby ([#17851](https://github.com/NVIDIA/cudf/pull/17851)) [@PointKernel](https://github.com/PointKernel) +- Fix possible OOB mem access in Parquet decoder ([#17841](https://github.com/NVIDIA/cudf/pull/17841)) [@mhaseeb123](https://github.com/mhaseeb123) +- Require batches to be non-empty in multi-batch JSON reader ([#17837](https://github.com/NVIDIA/cudf/pull/17837)) [@shrshi](https://github.com/shrshi) +- Fix rolling(min_periods=) with int and null data with mode.pandas_compat ([#17822](https://github.com/NVIDIA/cudf/pull/17822)) [@mroeschke](https://github.com/mroeschke) +- Resolve race-condition in `disable_module_accelerator` ([#17811](https://github.com/NVIDIA/cudf/pull/17811)) [@galipremsagar](https://github.com/galipremsagar) +- Make Series(dtype=object) raise in mode.pandas_compat with non string data ([#17804](https://github.com/NVIDIA/cudf/pull/17804)) [@mroeschke](https://github.com/mroeschke) +- Disable intended disabled ORC tests ([#17790](https://github.com/NVIDIA/cudf/pull/17790)) [@davidwendt](https://github.com/davidwendt) +- Fix empty DataFrame construction not returning RangeIndex columns ([#17784](https://github.com/NVIDIA/cudf/pull/17784)) [@mroeschke](https://github.com/mroeschke) +- Fix various `.str` methods for pandas compatability ([#17782](https://github.com/NVIDIA/cudf/pull/17782)) [@mroeschke](https://github.com/mroeschke) +- Fix `count` API issue about ignoring nan values ([#17779](https://github.com/NVIDIA/cudf/pull/17779)) [@galipremsagar](https://github.com/galipremsagar) +- Add `numba` pinning to `cudf` repo ([#17777](https://github.com/NVIDIA/cudf/pull/17777)) [@galipremsagar](https://github.com/galipremsagar) +- Allow .sort_values(na_position=) to include NaNs in mode.pandas_compatible ([#17776](https://github.com/NVIDIA/cudf/pull/17776)) [@mroeschke](https://github.com/mroeschke) +- allow deselecting nvcomp wheels ([#17774](https://github.com/NVIDIA/cudf/pull/17774)) [@jameslamb](https://github.com/jameslamb) +- Use the `aligned_resource_adaptor` to allocate bloom filter device buffers ([#17758](https://github.com/NVIDIA/cudf/pull/17758)) [@mhaseeb123](https://github.com/mhaseeb123) +- Avoid instantiating bloom filter query function for nested and bool types ([#17753](https://github.com/NVIDIA/cudf/pull/17753)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix DataFrame.merge(Series, how="left"/"right") on column and index not resulting in a RangeIndex ([#17739](https://github.com/NVIDIA/cudf/pull/17739)) [@mroeschke](https://github.com/mroeschke) +- [BUG] xfail Polars excel test ([#17731](https://github.com/NVIDIA/cudf/pull/17731)) [@Matt711](https://github.com/Matt711) +- Require to implement `AutoCloseable` for the classes derived from `HostUDFWrapper` ([#17727](https://github.com/NVIDIA/cudf/pull/17727)) [@ttnghia](https://github.com/ttnghia) +- Remove jlowe as a java committer since he retired ([#17725](https://github.com/NVIDIA/cudf/pull/17725)) [@tgravescs](https://github.com/tgravescs) +- Prevent use of invalid grid sizes in ORC reader and writer ([#17709](https://github.com/NVIDIA/cudf/pull/17709)) [@vuule](https://github.com/vuule) +- Enforce schema for partial tables in multi-source multi-batch JSON reader ([#17708](https://github.com/NVIDIA/cudf/pull/17708)) [@shrshi](https://github.com/shrshi) +- Compute and use the initial string offset when building `nested` large string cols with chunked parquet reader ([#17702](https://github.com/NVIDIA/cudf/pull/17702)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix writing of compressed ORC files with large stripe footers ([#17700](https://github.com/NVIDIA/cudf/pull/17700)) [@vuule](https://github.com/vuule) +- Fix cudf.polars sum of empty not equalling zero ([#17685](https://github.com/NVIDIA/cudf/pull/17685)) [@mroeschke](https://github.com/mroeschke) +- Fix formatting in logging ([#17680](https://github.com/NVIDIA/cudf/pull/17680)) [@vuule](https://github.com/vuule) +- convert all nulls to nans in a specific scenario ([#17677](https://github.com/NVIDIA/cudf/pull/17677)) [@galipremsagar](https://github.com/galipremsagar) +- Define cudf repr methods on the Column ([#17675](https://github.com/NVIDIA/cudf/pull/17675)) [@mroeschke](https://github.com/mroeschke) +- Fix groupby.len with null values in cudf.polars ([#17671](https://github.com/NVIDIA/cudf/pull/17671)) [@mroeschke](https://github.com/mroeschke) +- Fix: DataFrameGroupBy.get_group was raising with length>1 tuples ([#17653](https://github.com/NVIDIA/cudf/pull/17653)) [@MarcoGorelli](https://github.com/MarcoGorelli) +- Fix possible int overflow in compute_mixed_join_output_size ([#17633](https://github.com/NVIDIA/cudf/pull/17633)) [@davidwendt](https://github.com/davidwendt) +- Fix a minor potential i32 overflow in `thrust::transform_exclusive_scan` in PQ reader preprocessing ([#17617](https://github.com/NVIDIA/cudf/pull/17617)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix failing xgboost test in the cudf.pandas third-party integration tests ([#17616](https://github.com/NVIDIA/cudf/pull/17616)) [@Matt711](https://github.com/Matt711) +- Fix ``dask_cudf.read_csv`` ([#17612](https://github.com/NVIDIA/cudf/pull/17612)) [@rjzamora](https://github.com/rjzamora) +- Fix memcheck error in ReplaceTest.NormalizeNansAndZerosMutable gtest ([#17610](https://github.com/NVIDIA/cudf/pull/17610)) [@davidwendt](https://github.com/davidwendt) +- Correctly accept a `pandas.CategoricalDtype(pandas.IntervalDtype(...), ...)` type ([#17604](https://github.com/NVIDIA/cudf/pull/17604)) [@mroeschke](https://github.com/mroeschke) +- Add ability to modify and propagate `names` of `columns` object ([#17597](https://github.com/NVIDIA/cudf/pull/17597)) [@galipremsagar](https://github.com/galipremsagar) +- Ignore NaN correctly in .quantile ([#17593](https://github.com/NVIDIA/cudf/pull/17593)) [@mroeschke](https://github.com/mroeschke) +- Fix groupby argmin/max gather of sorted-order indices ([#17591](https://github.com/NVIDIA/cudf/pull/17591)) [@davidwendt](https://github.com/davidwendt) +- Fix ctest fail running libcudf tests in a Debug build ([#17576](https://github.com/NVIDIA/cudf/pull/17576)) [@davidwendt](https://github.com/davidwendt) +- Specify a version for rapids_logger dependency ([#17573](https://github.com/NVIDIA/cudf/pull/17573)) [@jlowe](https://github.com/jlowe) +- Fix the ORC decoding bug for the timestamp data ([#17570](https://github.com/NVIDIA/cudf/pull/17570)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- [JNI] remove rmm argument to set rw access for fabric handles ([#17553](https://github.com/NVIDIA/cudf/pull/17553)) [@abellina](https://github.com/abellina) +- Document undefined behavior in div_rounding_up_safe ([#17542](https://github.com/NVIDIA/cudf/pull/17542)) [@davidwendt](https://github.com/davidwendt) +- Fix nvcc-imposed UB in `constexpr` functions ([#17534](https://github.com/NVIDIA/cudf/pull/17534)) [@vuule](https://github.com/vuule) +- Add anonymous namespace to libcudf test source ([#17529](https://github.com/NVIDIA/cudf/pull/17529)) [@davidwendt](https://github.com/davidwendt) +- Propagate failures in pandas integration tests and Skip failing tests ([#17521](https://github.com/NVIDIA/cudf/pull/17521)) [@Matt711](https://github.com/Matt711) +- Fix libcudf compile error when logging is disabled ([#17512](https://github.com/NVIDIA/cudf/pull/17512)) [@davidwendt](https://github.com/davidwendt) +- Fix Dask-cuDF `clip` APIs ([#17509](https://github.com/NVIDIA/cudf/pull/17509)) [@rjzamora](https://github.com/rjzamora) +- Fix pylibcudf to_arrow with multiple nested data types ([#17504](https://github.com/NVIDIA/cudf/pull/17504)) [@mroeschke](https://github.com/mroeschke) +- Fix groupby(as_index=False).size not reseting index ([#17499](https://github.com/NVIDIA/cudf/pull/17499)) [@mroeschke](https://github.com/mroeschke) +- Revert "Temporarily skip tests due to dask/distributed#8953" ([#17492](https://github.com/NVIDIA/cudf/pull/17492)) [@Matt711](https://github.com/Matt711) +- Workaround for a misaligned access in `read_csv` on some CUDA versions ([#17477](https://github.com/NVIDIA/cudf/pull/17477)) [@vuule](https://github.com/vuule) +- Fix some possible thread-id overflow calculations ([#17473](https://github.com/NVIDIA/cudf/pull/17473)) [@davidwendt](https://github.com/davidwendt) +- Temporarily skip tests due to dask/distributed#8953 ([#17472](https://github.com/NVIDIA/cudf/pull/17472)) [@wence-](https://github.com/wence-) +- Detect mismatches in begin and end tokens returned by JSON tokenizer FST ([#17471](https://github.com/NVIDIA/cudf/pull/17471)) [@shrshi](https://github.com/shrshi) +- Support dask>=2024.11.2 in Dask cuDF ([#17439](https://github.com/NVIDIA/cudf/pull/17439)) [@rjzamora](https://github.com/rjzamora) +- Fix write_json failure for zero columns in table/struct ([#17414](https://github.com/NVIDIA/cudf/pull/17414)) [@karthikeyann](https://github.com/karthikeyann) +- Fix Debug-mode failing Arrow test ([#17405](https://github.com/NVIDIA/cudf/pull/17405)) [@zeroshade](https://github.com/zeroshade) +- Fix all null list column with missing child column in JSON reader ([#17348](https://github.com/NVIDIA/cudf/pull/17348)) [@karthikeyann](https://github.com/karthikeyann) ## 📖 Documentation -- Fix incorrect example in pylibcudf docs ([#17912](https://github.com/rapidsai/cudf/pull/17912)) [@Matt711](https://github.com/Matt711) -- Explicitly call out that the GPU open beta runs on a single GPU ([#17872](https://github.com/rapidsai/cudf/pull/17872)) [@taureandyernv](https://github.com/taureandyernv) -- Update cudf.pandas colab link in docs ([#17846](https://github.com/rapidsai/cudf/pull/17846)) [@taureandyernv](https://github.com/taureandyernv) -- [DOC] Make pylibcudf docs more visible ([#17803](https://github.com/rapidsai/cudf/pull/17803)) [@Matt711](https://github.com/Matt711) -- Cross-link cudf.pandas profiler documentation. ([#17668](https://github.com/rapidsai/cudf/pull/17668)) [@bdice](https://github.com/bdice) -- Document interpreter install command for cudf.pandas ([#17358](https://github.com/rapidsai/cudf/pull/17358)) [@bdice](https://github.com/bdice) -- add comment to Series.tolist method ([#17350](https://github.com/rapidsai/cudf/pull/17350)) [@tequilayu](https://github.com/tequilayu) +- Fix incorrect example in pylibcudf docs ([#17912](https://github.com/NVIDIA/cudf/pull/17912)) [@Matt711](https://github.com/Matt711) +- Explicitly call out that the GPU open beta runs on a single GPU ([#17872](https://github.com/NVIDIA/cudf/pull/17872)) [@taureandyernv](https://github.com/taureandyernv) +- Update cudf.pandas colab link in docs ([#17846](https://github.com/NVIDIA/cudf/pull/17846)) [@taureandyernv](https://github.com/taureandyernv) +- [DOC] Make pylibcudf docs more visible ([#17803](https://github.com/NVIDIA/cudf/pull/17803)) [@Matt711](https://github.com/Matt711) +- Cross-link cudf.pandas profiler documentation. ([#17668](https://github.com/NVIDIA/cudf/pull/17668)) [@bdice](https://github.com/bdice) +- Document interpreter install command for cudf.pandas ([#17358](https://github.com/NVIDIA/cudf/pull/17358)) [@bdice](https://github.com/bdice) +- add comment to Series.tolist method ([#17350](https://github.com/NVIDIA/cudf/pull/17350)) [@tequilayu](https://github.com/tequilayu) ## 🚀 New Features -- Bump polars version to <1.22 ([#17771](https://github.com/rapidsai/cudf/pull/17771)) [@Matt711](https://github.com/Matt711) -- Make more constexpr available on device for cuIO ([#17746](https://github.com/rapidsai/cudf/pull/17746)) [@PointKernel](https://github.com/PointKernel) -- Add public interop functions between pylibcudf and cudf classic ([#17730](https://github.com/rapidsai/cudf/pull/17730)) [@Matt711](https://github.com/Matt711) -- Support `dask_expr` migration into `dask.dataframe` ([#17704](https://github.com/rapidsai/cudf/pull/17704)) [@rjzamora](https://github.com/rjzamora) -- Make tests build without relaxed constexpr ([#17691](https://github.com/rapidsai/cudf/pull/17691)) [@PointKernel](https://github.com/PointKernel) -- Set default logger level to warn ([#17684](https://github.com/rapidsai/cudf/pull/17684)) [@vyasr](https://github.com/vyasr) -- Support multithreaded reading of compressed buffers in JSON reader ([#17670](https://github.com/rapidsai/cudf/pull/17670)) [@shrshi](https://github.com/shrshi) -- Control pinned memory use with environment variables ([#17657](https://github.com/rapidsai/cudf/pull/17657)) [@vuule](https://github.com/vuule) -- Host compression ([#17656](https://github.com/rapidsai/cudf/pull/17656)) [@vuule](https://github.com/vuule) -- Enable text build without relying on relaxed constexpr ([#17647](https://github.com/rapidsai/cudf/pull/17647)) [@PointKernel](https://github.com/PointKernel) -- Implement `HOST_UDF` aggregation for reduction and segmented reduction ([#17645](https://github.com/rapidsai/cudf/pull/17645)) [@ttnghia](https://github.com/ttnghia) -- Add JSON reader options structs to pylibcudf ([#17614](https://github.com/rapidsai/cudf/pull/17614)) [@Matt711](https://github.com/Matt711) -- Refactor distinct hash join to handle multiple probes with the same build table ([#17609](https://github.com/rapidsai/cudf/pull/17609)) [@PointKernel](https://github.com/PointKernel) -- Add JSON Writer options classes to pylibcudf ([#17606](https://github.com/rapidsai/cudf/pull/17606)) [@Matt711](https://github.com/Matt711) -- Add ORC reader options structs to pylibcudf ([#17601](https://github.com/rapidsai/cudf/pull/17601)) [@Matt711](https://github.com/Matt711) -- Add Avro Reader options classes to pylibcudf ([#17599](https://github.com/rapidsai/cudf/pull/17599)) [@Matt711](https://github.com/Matt711) -- Enable binaryop build without relying on relaxed constexpr ([#17598](https://github.com/rapidsai/cudf/pull/17598)) [@PointKernel](https://github.com/PointKernel) -- Measure the number of Parquet row groups filtered by predicate pushdown ([#17594](https://github.com/rapidsai/cudf/pull/17594)) [@mhaseeb123](https://github.com/mhaseeb123) -- Implement `HOST_UDF` aggregation for groupby ([#17592](https://github.com/rapidsai/cudf/pull/17592)) [@ttnghia](https://github.com/ttnghia) -- Plumb pylibcudf.io.parquet options classes through cudf python ([#17506](https://github.com/rapidsai/cudf/pull/17506)) [@Matt711](https://github.com/Matt711) -- Add partition-wise `Select` support to cuDF-Polars ([#17495](https://github.com/rapidsai/cudf/pull/17495)) [@rjzamora](https://github.com/rjzamora) -- Add multi-partition `Scan` support to cuDF-Polars ([#17494](https://github.com/rapidsai/cudf/pull/17494)) [@rjzamora](https://github.com/rjzamora) -- Migrate `cudf::io::merge_row_group_metadata` to pylibcudf ([#17491](https://github.com/rapidsai/cudf/pull/17491)) [@Matt711](https://github.com/Matt711) -- Add Parquet Reader options classes to pylibcudf ([#17464](https://github.com/rapidsai/cudf/pull/17464)) [@Matt711](https://github.com/Matt711) -- Add multi-partition `DataFrameScan` support to cuDF-Polars ([#17441](https://github.com/rapidsai/cudf/pull/17441)) [@rjzamora](https://github.com/rjzamora) -- Return empty result for segmented_reduce if input and offsets are both empty ([#17437](https://github.com/rapidsai/cudf/pull/17437)) [@davidwendt](https://github.com/davidwendt) -- Abstract polars function expression nodes to ensure they are serializable ([#17418](https://github.com/rapidsai/cudf/pull/17418)) [@pentschev](https://github.com/pentschev) -- Add CSV Reader options classes to pylibcudf ([#17412](https://github.com/rapidsai/cudf/pull/17412)) [@Matt711](https://github.com/Matt711) -- Add support for `pylibcudf.DataType` serialization ([#17352](https://github.com/rapidsai/cudf/pull/17352)) [@pentschev](https://github.com/pentschev) -- Enable rounding for Decimal32 and Decimal64 in cuDF ([#17332](https://github.com/rapidsai/cudf/pull/17332)) [@a-hirota](https://github.com/a-hirota) -- Remove upper bounds on cuda-python to allow 12.6.2 and 11.8.5 ([#17326](https://github.com/rapidsai/cudf/pull/17326)) [@bdice](https://github.com/bdice) -- Expose stream-ordering to groupby APIs ([#17324](https://github.com/rapidsai/cudf/pull/17324)) [@shrshi](https://github.com/shrshi) -- Migrate ORC Writer to pylibcudf ([#17310](https://github.com/rapidsai/cudf/pull/17310)) [@Matt711](https://github.com/Matt711) -- Support reading bloom filters from Parquet files and filter row groups using them ([#17289](https://github.com/rapidsai/cudf/pull/17289)) [@mhaseeb123](https://github.com/mhaseeb123) +- Bump polars version to <1.22 ([#17771](https://github.com/NVIDIA/cudf/pull/17771)) [@Matt711](https://github.com/Matt711) +- Make more constexpr available on device for cuIO ([#17746](https://github.com/NVIDIA/cudf/pull/17746)) [@PointKernel](https://github.com/PointKernel) +- Add public interop functions between pylibcudf and cudf classic ([#17730](https://github.com/NVIDIA/cudf/pull/17730)) [@Matt711](https://github.com/Matt711) +- Support `dask_expr` migration into `dask.dataframe` ([#17704](https://github.com/NVIDIA/cudf/pull/17704)) [@rjzamora](https://github.com/rjzamora) +- Make tests build without relaxed constexpr ([#17691](https://github.com/NVIDIA/cudf/pull/17691)) [@PointKernel](https://github.com/PointKernel) +- Set default logger level to warn ([#17684](https://github.com/NVIDIA/cudf/pull/17684)) [@vyasr](https://github.com/vyasr) +- Support multithreaded reading of compressed buffers in JSON reader ([#17670](https://github.com/NVIDIA/cudf/pull/17670)) [@shrshi](https://github.com/shrshi) +- Control pinned memory use with environment variables ([#17657](https://github.com/NVIDIA/cudf/pull/17657)) [@vuule](https://github.com/vuule) +- Host compression ([#17656](https://github.com/NVIDIA/cudf/pull/17656)) [@vuule](https://github.com/vuule) +- Enable text build without relying on relaxed constexpr ([#17647](https://github.com/NVIDIA/cudf/pull/17647)) [@PointKernel](https://github.com/PointKernel) +- Implement `HOST_UDF` aggregation for reduction and segmented reduction ([#17645](https://github.com/NVIDIA/cudf/pull/17645)) [@ttnghia](https://github.com/ttnghia) +- Add JSON reader options structs to pylibcudf ([#17614](https://github.com/NVIDIA/cudf/pull/17614)) [@Matt711](https://github.com/Matt711) +- Refactor distinct hash join to handle multiple probes with the same build table ([#17609](https://github.com/NVIDIA/cudf/pull/17609)) [@PointKernel](https://github.com/PointKernel) +- Add JSON Writer options classes to pylibcudf ([#17606](https://github.com/NVIDIA/cudf/pull/17606)) [@Matt711](https://github.com/Matt711) +- Add ORC reader options structs to pylibcudf ([#17601](https://github.com/NVIDIA/cudf/pull/17601)) [@Matt711](https://github.com/Matt711) +- Add Avro Reader options classes to pylibcudf ([#17599](https://github.com/NVIDIA/cudf/pull/17599)) [@Matt711](https://github.com/Matt711) +- Enable binaryop build without relying on relaxed constexpr ([#17598](https://github.com/NVIDIA/cudf/pull/17598)) [@PointKernel](https://github.com/PointKernel) +- Measure the number of Parquet row groups filtered by predicate pushdown ([#17594](https://github.com/NVIDIA/cudf/pull/17594)) [@mhaseeb123](https://github.com/mhaseeb123) +- Implement `HOST_UDF` aggregation for groupby ([#17592](https://github.com/NVIDIA/cudf/pull/17592)) [@ttnghia](https://github.com/ttnghia) +- Plumb pylibcudf.io.parquet options classes through cudf python ([#17506](https://github.com/NVIDIA/cudf/pull/17506)) [@Matt711](https://github.com/Matt711) +- Add partition-wise `Select` support to cuDF-Polars ([#17495](https://github.com/NVIDIA/cudf/pull/17495)) [@rjzamora](https://github.com/rjzamora) +- Add multi-partition `Scan` support to cuDF-Polars ([#17494](https://github.com/NVIDIA/cudf/pull/17494)) [@rjzamora](https://github.com/rjzamora) +- Migrate `cudf::io::merge_row_group_metadata` to pylibcudf ([#17491](https://github.com/NVIDIA/cudf/pull/17491)) [@Matt711](https://github.com/Matt711) +- Add Parquet Reader options classes to pylibcudf ([#17464](https://github.com/NVIDIA/cudf/pull/17464)) [@Matt711](https://github.com/Matt711) +- Add multi-partition `DataFrameScan` support to cuDF-Polars ([#17441](https://github.com/NVIDIA/cudf/pull/17441)) [@rjzamora](https://github.com/rjzamora) +- Return empty result for segmented_reduce if input and offsets are both empty ([#17437](https://github.com/NVIDIA/cudf/pull/17437)) [@davidwendt](https://github.com/davidwendt) +- Abstract polars function expression nodes to ensure they are serializable ([#17418](https://github.com/NVIDIA/cudf/pull/17418)) [@pentschev](https://github.com/pentschev) +- Add CSV Reader options classes to pylibcudf ([#17412](https://github.com/NVIDIA/cudf/pull/17412)) [@Matt711](https://github.com/Matt711) +- Add support for `pylibcudf.DataType` serialization ([#17352](https://github.com/NVIDIA/cudf/pull/17352)) [@pentschev](https://github.com/pentschev) +- Enable rounding for Decimal32 and Decimal64 in cuDF ([#17332](https://github.com/NVIDIA/cudf/pull/17332)) [@a-hirota](https://github.com/a-hirota) +- Remove upper bounds on cuda-python to allow 12.6.2 and 11.8.5 ([#17326](https://github.com/NVIDIA/cudf/pull/17326)) [@bdice](https://github.com/bdice) +- Expose stream-ordering to groupby APIs ([#17324](https://github.com/NVIDIA/cudf/pull/17324)) [@shrshi](https://github.com/shrshi) +- Migrate ORC Writer to pylibcudf ([#17310](https://github.com/NVIDIA/cudf/pull/17310)) [@Matt711](https://github.com/Matt711) +- Support reading bloom filters from Parquet files and filter row groups using them ([#17289](https://github.com/NVIDIA/cudf/pull/17289)) [@mhaseeb123](https://github.com/mhaseeb123) ## 🛠️ Improvements -- Remove pandas backend from `cudf.pandas` - ibis integration tests ([#17945](https://github.com/rapidsai/cudf/pull/17945)) [@Matt711](https://github.com/Matt711) -- Revert CUDA 12.8 shared workflow branch changes ([#17879](https://github.com/rapidsai/cudf/pull/17879)) [@vyasr](https://github.com/vyasr) -- Remove predicate param from `DataFrameScan` IR ([#17852](https://github.com/rapidsai/cudf/pull/17852)) [@Matt711](https://github.com/Matt711) -- Remove cudf.Scalar from scatter APIs ([#17847](https://github.com/rapidsai/cudf/pull/17847)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf.Scalar from interval_range ([#17844](https://github.com/rapidsai/cudf/pull/17844)) [@mroeschke](https://github.com/mroeschke) -- Add `verify-codeowners` hook ([#17840](https://github.com/rapidsai/cudf/pull/17840)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Build and test with CUDA 12.8.0 ([#17834](https://github.com/rapidsai/cudf/pull/17834)) [@bdice](https://github.com/bdice) -- Increase timeout for recently added test ([#17829](https://github.com/rapidsai/cudf/pull/17829)) [@galipremsagar](https://github.com/galipremsagar) -- Apply ruff everywhere (notebooks and scripts) ([#17820](https://github.com/rapidsai/cudf/pull/17820)) [@bdice](https://github.com/bdice) -- Fix pre-commit.ci failures ([#17819](https://github.com/rapidsai/cudf/pull/17819)) [@bdice](https://github.com/bdice) -- Remove incorrect calls to set architectures ([#17813](https://github.com/rapidsai/cudf/pull/17813)) [@vyasr](https://github.com/vyasr) -- Fix typo in exception raised when attempting to convert a string column to cupy ([#17800](https://github.com/rapidsai/cudf/pull/17800)) [@dagardner-nv](https://github.com/dagardner-nv) -- Add support for `pyarrow-19` ([#17794](https://github.com/rapidsai/cudf/pull/17794)) [@galipremsagar](https://github.com/galipremsagar) -- increase parallelism in nightly builds ([#17792](https://github.com/rapidsai/cudf/pull/17792)) [@jameslamb](https://github.com/jameslamb) -- Reduce libcudf memcheck tests output ([#17791](https://github.com/rapidsai/cudf/pull/17791)) [@davidwendt](https://github.com/davidwendt) -- Make cudf build with latest CCCL ([#17788](https://github.com/rapidsai/cudf/pull/17788)) [@miscco](https://github.com/miscco) -- Introduce some more rolling window benchmarks ([#17787](https://github.com/rapidsai/cudf/pull/17787)) [@wence-](https://github.com/wence-) -- Add shellcheck to pre-commit and fix warnings ([#17778](https://github.com/rapidsai/cudf/pull/17778)) [@gforsyth](https://github.com/gforsyth) -- Improve parquet reader very-long string performance ([#17773](https://github.com/rapidsai/cudf/pull/17773)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Update how to manage host UDF instance ([#17770](https://github.com/rapidsai/cudf/pull/17770)) [@res-life](https://github.com/res-life) -- Add getInts api for HostMemoryBuffer and UnsafeMemoryAccessor ([#17767](https://github.com/rapidsai/cudf/pull/17767)) [@liurenjie1024](https://github.com/liurenjie1024) -- Expose stream-ordering in scalar and avro APIs ([#17766](https://github.com/rapidsai/cudf/pull/17766)) [@shrshi](https://github.com/shrshi) -- Standarize methods used from `cudf.core._internals` ([#17765](https://github.com/rapidsai/cudf/pull/17765)) [@mroeschke](https://github.com/mroeschke) -- Implement string join in cudf-polars ([#17755](https://github.com/rapidsai/cudf/pull/17755)) [@wence-](https://github.com/wence-) -- Deprecate dataframe protocol ([#17736](https://github.com/rapidsai/cudf/pull/17736)) [@vyasr](https://github.com/vyasr) -- Add parquet reader long row test ([#17735](https://github.com/rapidsai/cudf/pull/17735)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Update kvikio call due to upstream changes ([#17733](https://github.com/rapidsai/cudf/pull/17733)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Delay setting MultiIndex.level/codes until needed ([#17728](https://github.com/rapidsai/cudf/pull/17728)) [@mroeschke](https://github.com/mroeschke) -- Bounding pool size in multi-batch JSON reader ([#17724](https://github.com/rapidsai/cudf/pull/17724)) [@shrshi](https://github.com/shrshi) -- Use GCC 13 in CUDA 12 conda builds. ([#17721](https://github.com/rapidsai/cudf/pull/17721)) [@bdice](https://github.com/bdice) -- Update minimal sphinx theme version so that we can use parallel doc builds ([#17719](https://github.com/rapidsai/cudf/pull/17719)) [@vyasr](https://github.com/vyasr) -- Add more aggregation methods in pylibcudf ([#17717](https://github.com/rapidsai/cudf/pull/17717)) [@mroeschke](https://github.com/mroeschke) -- Make cudf._lib.string_udf work with pylibcudf Columns instead of cudf._lib Columns ([#17715](https://github.com/rapidsai/cudf/pull/17715)) [@mroeschke](https://github.com/mroeschke) -- Add special orc test data: timestamp interspersed with null values ([#17713](https://github.com/rapidsai/cudf/pull/17713)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Add pylibcudf.null_mask.null_count ([#17711](https://github.com/rapidsai/cudf/pull/17711)) [@mroeschke](https://github.com/mroeschke) -- Ensure pyarrow.Scalar to pylibcudf.Scalar is cached ([#17707](https://github.com/rapidsai/cudf/pull/17707)) [@mroeschke](https://github.com/mroeschke) -- Adapt cudf numba config for numba 0.61 removal ([#17705](https://github.com/rapidsai/cudf/pull/17705)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.scalar in favor of pylibcudf ([#17701](https://github.com/rapidsai/cudf/pull/17701)) [@mroeschke](https://github.com/mroeschke) -- Fix parquet reader list bug ([#17699](https://github.com/rapidsai/cudf/pull/17699)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Migrated Dynamic AST Expression Trees in Benchmarks and Tests to use AST Tree ([#17697](https://github.com/rapidsai/cudf/pull/17697)) [@lamarrr](https://github.com/lamarrr) -- Skip polars test that can generate timezones that chrono_tz doesn't know ([#17694](https://github.com/rapidsai/cudf/pull/17694)) [@wence-](https://github.com/wence-) -- Use 64-bit offsets only if the current strings column output chunk size exceeds threshold ([#17693](https://github.com/rapidsai/cudf/pull/17693)) [@mhaseeb123](https://github.com/mhaseeb123) -- Use latest ci-conda images ([#17690](https://github.com/rapidsai/cudf/pull/17690)) [@bdice](https://github.com/bdice) -- Add multi-source reading to JSON reader benchmarks ([#17688](https://github.com/rapidsai/cudf/pull/17688)) [@shrshi](https://github.com/shrshi) -- Convert cudf.Scalar usage to pylibcudf and pyarrow usage ([#17686](https://github.com/rapidsai/cudf/pull/17686)) [@mroeschke](https://github.com/mroeschke) -- remove find_package(Python) in libcudf build ([#17683](https://github.com/rapidsai/cudf/pull/17683)) [@jameslamb](https://github.com/jameslamb) -- Fix build metrics report format with long placehold filenames ([#17679](https://github.com/rapidsai/cudf/pull/17679)) [@davidwendt](https://github.com/davidwendt) -- Use rapids-cmake for the logger ([#17674](https://github.com/rapidsai/cudf/pull/17674)) [@vyasr](https://github.com/vyasr) -- Java Parquet reads via multiple host buffers ([#17673](https://github.com/rapidsai/cudf/pull/17673)) [@jlowe](https://github.com/jlowe) -- Remove cudf._libs.types.pyx ([#17665](https://github.com/rapidsai/cudf/pull/17665)) [@mroeschke](https://github.com/mroeschke) -- Add support for `Groupby.cumprod` ([#17661](https://github.com/rapidsai/cudf/pull/17661)) [@galipremsagar](https://github.com/galipremsagar) -- Implement `.dt.total_seconds` ([#17659](https://github.com/rapidsai/cudf/pull/17659)) [@galipremsagar](https://github.com/galipremsagar) -- Avoid shallow copies in groupby methods ([#17646](https://github.com/rapidsai/cudf/pull/17646)) [@mroeschke](https://github.com/mroeschke) -- Avoid double MultiIndex factorization in groupby index result ([#17644](https://github.com/rapidsai/cudf/pull/17644)) [@mroeschke](https://github.com/mroeschke) -- Add seed parameter to hash_character_ngrams ([#17643](https://github.com/rapidsai/cudf/pull/17643)) [@davidwendt](https://github.com/davidwendt) -- Fix possible overflow in WriteCoalescingCallbackWrapper::TearDown ([#17642](https://github.com/rapidsai/cudf/pull/17642)) [@davidwendt](https://github.com/davidwendt) -- Remove pragma GCC diagnostic from source files ([#17637](https://github.com/rapidsai/cudf/pull/17637)) [@davidwendt](https://github.com/davidwendt) -- Move unnecessary utilities from cudf._lib.scalar ([#17636](https://github.com/rapidsai/cudf/pull/17636)) [@mroeschke](https://github.com/mroeschke) -- Support compression= in DataFrame.to_json ([#17634](https://github.com/rapidsai/cudf/pull/17634)) [@mroeschke](https://github.com/mroeschke) -- Bump Polars version to <1.18 ([#17632](https://github.com/rapidsai/cudf/pull/17632)) [@Matt711](https://github.com/Matt711) -- Add public APIs to Access Underlying `cudf` and `pandas` Objects from `cudf.pandas` Proxy Objects ([#17629](https://github.com/rapidsai/cudf/pull/17629)) [@galipremsagar](https://github.com/galipremsagar) -- Use Numba Config to turn on Pynvjitlink Features ([#17628](https://github.com/rapidsai/cudf/pull/17628)) [@isVoid](https://github.com/isVoid) -- Use PyNVML 12 ([#17627](https://github.com/rapidsai/cudf/pull/17627)) [@jakirkham](https://github.com/jakirkham) -- Remove cudf._lib.utils in favor of python APIs ([#17625](https://github.com/rapidsai/cudf/pull/17625)) [@mroeschke](https://github.com/mroeschke) -- Performance improvements and simplifications for fixed size row-based rolling windows ([#17623](https://github.com/rapidsai/cudf/pull/17623)) [@wence-](https://github.com/wence-) -- Fix return types for MurmurHash3_x86_32 template specializations ([#17622](https://github.com/rapidsai/cudf/pull/17622)) [@davidwendt](https://github.com/davidwendt) -- Clean up namespaces and improve compression-related headers ([#17621](https://github.com/rapidsai/cudf/pull/17621)) [@vuule](https://github.com/vuule) -- Use more pylibcudf.types instead of cudf._lib.types ([#17619](https://github.com/rapidsai/cudf/pull/17619)) [@mroeschke](https://github.com/mroeschke) -- Remove patch that is only needed for clang-tidy to run on test files ([#17618](https://github.com/rapidsai/cudf/pull/17618)) [@vyasr](https://github.com/vyasr) -- update telemetry actions to fluent-bit friendly style ([#17615](https://github.com/rapidsai/cudf/pull/17615)) [@msarahan](https://github.com/msarahan) -- Introduce some simple benchmarks for rolling window aggregations ([#17613](https://github.com/rapidsai/cudf/pull/17613)) [@wence-](https://github.com/wence-) -- Bump the oldest `pyarrow` version to `14.0.2` in test matrix ([#17611](https://github.com/rapidsai/cudf/pull/17611)) [@galipremsagar](https://github.com/galipremsagar) -- Use `[[nodiscard]]` attribute before `__device__` ([#17608](https://github.com/rapidsai/cudf/pull/17608)) [@vuule](https://github.com/vuule) -- Use `host_vector` in `flatten_single_pass_aggs` ([#17605](https://github.com/rapidsai/cudf/pull/17605)) [@vuule](https://github.com/vuule) -- Stop memory_resource.hpp from including itself ([#17603](https://github.com/rapidsai/cudf/pull/17603)) [@vyasr](https://github.com/vyasr) -- Replace the outdated cuco window concept with buckets ([#17602](https://github.com/rapidsai/cudf/pull/17602)) [@PointKernel](https://github.com/PointKernel) -- Check if nightlies have succeeded recently enough ([#17596](https://github.com/rapidsai/cudf/pull/17596)) [@vyasr](https://github.com/vyasr) -- Deprecate cudf::grouped_time_range_rolling_window ([#17589](https://github.com/rapidsai/cudf/pull/17589)) [@wence-](https://github.com/wence-) -- A couple of fixes in rapids-logger usage ([#17588](https://github.com/rapidsai/cudf/pull/17588)) [@vyasr](https://github.com/vyasr) -- Simplify expression transformer in Parquet predicate pushdown with `ast::tree` ([#17587](https://github.com/rapidsai/cudf/pull/17587)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove unused functionality in cudf._lib.utils.pyx ([#17586](https://github.com/rapidsai/cudf/pull/17586)) [@mroeschke](https://github.com/mroeschke) -- Use cuda-python `cuda.bindings` import names. ([#17585](https://github.com/rapidsai/cudf/pull/17585)) [@bdice](https://github.com/bdice) -- Use no-sync copy for fixed-width types in cudf::concatenate ([#17584](https://github.com/rapidsai/cudf/pull/17584)) [@davidwendt](https://github.com/davidwendt) -- Remove cudf._lib.groupby in favor of inlining pylibcudf ([#17582](https://github.com/rapidsai/cudf/pull/17582)) [@mroeschke](https://github.com/mroeschke) -- Remove unused code of json schema in JSON reader ([#17581](https://github.com/rapidsai/cudf/pull/17581)) [@karthikeyann](https://github.com/karthikeyann) -- Expose Scalar's constructor and `Scalar#getScalarHandle()` to public ([#17580](https://github.com/rapidsai/cudf/pull/17580)) [@ttnghia](https://github.com/ttnghia) -- Allow large strings in nvtext benchmarks ([#17579](https://github.com/rapidsai/cudf/pull/17579)) [@davidwendt](https://github.com/davidwendt) -- Remove cudf._lib.reduce in favor of inlining pylibcudf ([#17574](https://github.com/rapidsai/cudf/pull/17574)) [@mroeschke](https://github.com/mroeschke) -- Use batched memcpy when writing ORC statistics ([#17572](https://github.com/rapidsai/cudf/pull/17572)) [@vuule](https://github.com/vuule) -- Allow large strings in nvbench strings benchmarks ([#17571](https://github.com/rapidsai/cudf/pull/17571)) [@davidwendt](https://github.com/davidwendt) -- Update version references in workflow ([#17568](https://github.com/rapidsai/cudf/pull/17568)) [@AyodeAwe](https://github.com/AyodeAwe) -- Enable all json reader options in pylibcudf read_json ([#17563](https://github.com/rapidsai/cudf/pull/17563)) [@karthikeyann](https://github.com/karthikeyann) -- Remove cudf._lib.parquet in favor of inlining pylibcudf ([#17562](https://github.com/rapidsai/cudf/pull/17562)) [@mroeschke](https://github.com/mroeschke) -- Fix CMake format in cudf/_lib/CMakeLists.txt ([#17559](https://github.com/rapidsai/cudf/pull/17559)) [@mroeschke](https://github.com/mroeschke) -- Remove "legacy" Dask DataFrame support from Dask cuDF ([#17558](https://github.com/rapidsai/cudf/pull/17558)) [@rjzamora](https://github.com/rjzamora) -- Replace direct `cudaMemcpyAsync` calls with utility functions (within `/include`) ([#17557](https://github.com/rapidsai/cudf/pull/17557)) [@vuule](https://github.com/vuule) -- Remove cudf._lib.interop in favor of inlining pylibcudf ([#17555](https://github.com/rapidsai/cudf/pull/17555)) [@mroeschke](https://github.com/mroeschke) -- gate telemetry dispatch calls on TELEMETRY_ENABLED env var ([#17551](https://github.com/rapidsai/cudf/pull/17551)) [@msarahan](https://github.com/msarahan) -- Replace direct `cudaMemcpyAsync` calls with utility functions (within `/src`) ([#17550](https://github.com/rapidsai/cudf/pull/17550)) [@vuule](https://github.com/vuule) -- Remove unused `BufferArrayFromVector` ([#17549](https://github.com/rapidsai/cudf/pull/17549)) [@Matt711](https://github.com/Matt711) -- Move cudf._lib.copying to cudf.core._internals ([#17548](https://github.com/rapidsai/cudf/pull/17548)) [@mroeschke](https://github.com/mroeschke) -- Update cuda-python lower bounds to 12.6.2 / 11.8.5 ([#17547](https://github.com/rapidsai/cudf/pull/17547)) [@bdice](https://github.com/bdice) -- Fix typos, rename types, and add null_probability benchmark axis for distinct ([#17546](https://github.com/rapidsai/cudf/pull/17546)) [@PointKernel](https://github.com/PointKernel) -- Mark more constexpr functions as device-available ([#17545](https://github.com/rapidsai/cudf/pull/17545)) [@vyasr](https://github.com/vyasr) -- Use cooperative-groups instead of cub warp-reduce for strings contains ([#17540](https://github.com/rapidsai/cudf/pull/17540)) [@davidwendt](https://github.com/davidwendt) -- Remove cudf._lib.nvtext in favor of inlining pylibcudf ([#17535](https://github.com/rapidsai/cudf/pull/17535)) [@mroeschke](https://github.com/mroeschke) -- Add XXHash_32 hasher ([#17533](https://github.com/rapidsai/cudf/pull/17533)) [@PointKernel](https://github.com/PointKernel) -- Remove unused masked keyword in column_empty ([#17530](https://github.com/rapidsai/cudf/pull/17530)) [@mroeschke](https://github.com/mroeschke) -- Remove Thrust patch in favor of CMake definition for Thrust 32-bit offset types. ([#17527](https://github.com/rapidsai/cudf/pull/17527)) [@bdice](https://github.com/bdice) -- [JNI] Enables fabric handles for CUDA async memory pools ([#17526](https://github.com/rapidsai/cudf/pull/17526)) [@abellina](https://github.com/abellina) -- Force Thrust to use 32-bit offset type. ([#17523](https://github.com/rapidsai/cudf/pull/17523)) [@bdice](https://github.com/bdice) -- Replace cudf::detail::copy_if logic with thrust::copy_if and gather ([#17520](https://github.com/rapidsai/cudf/pull/17520)) [@davidwendt](https://github.com/davidwendt) -- Replaces uses of `cudf._lib.Column.from_unique_ptr` with `pylibcudf.Column.from_libcudf` ([#17517](https://github.com/rapidsai/cudf/pull/17517)) [@Matt711](https://github.com/Matt711) -- Move cudf._lib.aggregation to cudf.core._internals ([#17516](https://github.com/rapidsai/cudf/pull/17516)) [@mroeschke](https://github.com/mroeschke) -- Migrate copy_column and Column.from_scalar to pylibcudf ([#17513](https://github.com/rapidsai/cudf/pull/17513)) [@Matt711](https://github.com/Matt711) -- Remove cudf._lib.transform in favor of inlining pylibcudf ([#17505](https://github.com/rapidsai/cudf/pull/17505)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.string.convert/split in favor of inlining pylibcudf ([#17496](https://github.com/rapidsai/cudf/pull/17496)) [@mroeschke](https://github.com/mroeschke) -- Move cudf._lib.sort to cudf.core._internals ([#17488](https://github.com/rapidsai/cudf/pull/17488)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.csv in favor in inlining pylibcudf ([#17485](https://github.com/rapidsai/cudf/pull/17485)) [@mroeschke](https://github.com/mroeschke) -- Update PyTorch to >=2.4.0 to get fix for CUDA array interface bug, and drop CUDA 11 PyTorch tests. ([#17475](https://github.com/rapidsai/cudf/pull/17475)) [@bdice](https://github.com/bdice) -- Remove cudf._lib.binops in favor of inlining pylibcudf ([#17468](https://github.com/rapidsai/cudf/pull/17468)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.orc in favor of inlining pylibcudf ([#17466](https://github.com/rapidsai/cudf/pull/17466)) [@mroeschke](https://github.com/mroeschke) -- skip most CI on devcontainer-only changes ([#17465](https://github.com/rapidsai/cudf/pull/17465)) [@jameslamb](https://github.com/jameslamb) -- Set build type for all examples ([#17463](https://github.com/rapidsai/cudf/pull/17463)) [@vyasr](https://github.com/vyasr) -- Update the hook versions in pre-commit ([#17462](https://github.com/rapidsai/cudf/pull/17462)) [@wence-](https://github.com/wence-) -- Remove cudf._lib.string_casting in favor of inlining pylibcudf ([#17460](https://github.com/rapidsai/cudf/pull/17460)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.filling in favor of inlining pylibcudf ([#17459](https://github.com/rapidsai/cudf/pull/17459)) [@mroeschke](https://github.com/mroeschke) -- Update MurmurHash3_x64_128 to use the cuco equivalent implementation ([#17457](https://github.com/rapidsai/cudf/pull/17457)) [@PointKernel](https://github.com/PointKernel) -- Move cudf._lib.stream_compaction to cudf.core._internals ([#17456](https://github.com/rapidsai/cudf/pull/17456)) [@mroeschke](https://github.com/mroeschke) -- Clean up xxhash_64 implementations ([#17455](https://github.com/rapidsai/cudf/pull/17455)) [@PointKernel](https://github.com/PointKernel) -- Update Hadoop dependency in Java pom ([#17454](https://github.com/rapidsai/cudf/pull/17454)) [@jlowe](https://github.com/jlowe) -- Adapt to rmm logger changes ([#17451](https://github.com/rapidsai/cudf/pull/17451)) [@vyasr](https://github.com/vyasr) -- Require approval to run CI on draft PRs ([#17450](https://github.com/rapidsai/cudf/pull/17450)) [@bdice](https://github.com/bdice) -- Expose stream-ordering in nvtext API ([#17446](https://github.com/rapidsai/cudf/pull/17446)) [@shrshi](https://github.com/shrshi) -- Use exec_policy_nosync in write_json ([#17445](https://github.com/rapidsai/cudf/pull/17445)) [@karthikeyann](https://github.com/karthikeyann) -- Remove cudf._lib.json in favor of inlining pylibcudf ([#17443](https://github.com/rapidsai/cudf/pull/17443)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.null_mask in favor of inlining pylibcudf ([#17440](https://github.com/rapidsai/cudf/pull/17440)) [@mroeschke](https://github.com/mroeschke) -- Expose stream-ordering in replace API ([#17436](https://github.com/rapidsai/cudf/pull/17436)) [@shrshi](https://github.com/shrshi) -- Expose stream-ordering in copying APIs ([#17435](https://github.com/rapidsai/cudf/pull/17435)) [@shrshi](https://github.com/shrshi) -- Expose stream-ordering in column view APIs ([#17434](https://github.com/rapidsai/cudf/pull/17434)) [@shrshi](https://github.com/shrshi) -- Apply clang-tidy autofixes from new rules ([#17431](https://github.com/rapidsai/cudf/pull/17431)) [@vyasr](https://github.com/vyasr) -- Remove cudf._lib.round in favor of inlining pylibcudf ([#17430](https://github.com/rapidsai/cudf/pull/17430)) [@mroeschke](https://github.com/mroeschke) -- Update MurmurHash3_x86_32 to use the cuco equivalent implementation ([#17429](https://github.com/rapidsai/cudf/pull/17429)) [@PointKernel](https://github.com/PointKernel) -- Remove cudf._lib.replace in favor of inlining pylibcudf ([#17428](https://github.com/rapidsai/cudf/pull/17428)) [@mroeschke](https://github.com/mroeschke) -- Remove nvtx/ranges.hpp include from cuda.cuh ([#17427](https://github.com/rapidsai/cudf/pull/17427)) [@davidwendt](https://github.com/davidwendt) -- Remove the unused detail `int_fastdiv.h` header ([#17426](https://github.com/rapidsai/cudf/pull/17426)) [@PointKernel](https://github.com/PointKernel) -- Remove cudf._lib.lists in favor of inlining pylibcudf ([#17425](https://github.com/rapidsai/cudf/pull/17425)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.quantile ([#17424](https://github.com/rapidsai/cudf/pull/17424)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.rolling in favor of inlining pylibcudf ([#17423](https://github.com/rapidsai/cudf/pull/17423)) [@mroeschke](https://github.com/mroeschke) -- Avoid converting Decimal32/Decimal64 in `to_arrow` and `from_arrow` APIs ([#17422](https://github.com/rapidsai/cudf/pull/17422)) [@zeroshade](https://github.com/zeroshade) -- Rework minhash APIs for deprecation cycle ([#17421](https://github.com/rapidsai/cudf/pull/17421)) [@davidwendt](https://github.com/davidwendt) -- Use thread_index_type in binary-ops jit kernel.cu ([#17420](https://github.com/rapidsai/cudf/pull/17420)) [@davidwendt](https://github.com/davidwendt) -- Change binops for-each kernel to thrust::for_each_n ([#17419](https://github.com/rapidsai/cudf/pull/17419)) [@davidwendt](https://github.com/davidwendt) -- Move cudf._lib.search to cudf.core._internals ([#17411](https://github.com/rapidsai/cudf/pull/17411)) [@mroeschke](https://github.com/mroeschke) -- Use grid_1d utilities in copy_range.cuh ([#17409](https://github.com/rapidsai/cudf/pull/17409)) [@davidwendt](https://github.com/davidwendt) -- Remove cudf._lib.text in favor of inlining pylibcudf ([#17408](https://github.com/rapidsai/cudf/pull/17408)) [@mroeschke](https://github.com/mroeschke) -- Run clang-tidy checks in PR CI ([#17407](https://github.com/rapidsai/cudf/pull/17407)) [@bdice](https://github.com/bdice) -- Update strings/text source to use grid_1d for thread/block/stride calculations ([#17404](https://github.com/rapidsai/cudf/pull/17404)) [@davidwendt](https://github.com/davidwendt) -- Expose stream-ordering to strings attribute APIs ([#17398](https://github.com/rapidsai/cudf/pull/17398)) [@shrshi](https://github.com/shrshi) -- Expose stream-ordering to interop APIs ([#17397](https://github.com/rapidsai/cudf/pull/17397)) [@shrshi](https://github.com/shrshi) -- Remove unused type aliases ([#17396](https://github.com/rapidsai/cudf/pull/17396)) [@PointKernel](https://github.com/PointKernel) -- Remove some cudf._lib.strings files in favor of inlining pylibcudf ([#17394](https://github.com/rapidsai/cudf/pull/17394)) [@mroeschke](https://github.com/mroeschke) -- Update xxhash_64 to utilize the cuco equivalent implementation ([#17393](https://github.com/rapidsai/cudf/pull/17393)) [@PointKernel](https://github.com/PointKernel) -- Change indices for dictionary column to signed integer type ([#17390](https://github.com/rapidsai/cudf/pull/17390)) [@davidwendt](https://github.com/davidwendt) -- Return categorical values in to_numpy/to_cupy ([#17388](https://github.com/rapidsai/cudf/pull/17388)) [@mroeschke](https://github.com/mroeschke) -- Forward-merge branch-24.12 to branch-25.02 ([#17379](https://github.com/rapidsai/cudf/pull/17379)) [@bdice](https://github.com/bdice) -- Remove unused IO utilities from cudf python ([#17374](https://github.com/rapidsai/cudf/pull/17374)) [@Matt711](https://github.com/Matt711) -- Remove cudf._lib.datetime in favor of inlining pylibcudf ([#17372](https://github.com/rapidsai/cudf/pull/17372)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.join in favor of inlining pylibcudf ([#17371](https://github.com/rapidsai/cudf/pull/17371)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.merge in favor of inlining pylibcudf ([#17370](https://github.com/rapidsai/cudf/pull/17370)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.partitioning in favor of inlining pylibcudf ([#17369](https://github.com/rapidsai/cudf/pull/17369)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.reshape in favor of inlining pylibcudf ([#17368](https://github.com/rapidsai/cudf/pull/17368)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.timezone in favor of inlining pylibcudf ([#17366](https://github.com/rapidsai/cudf/pull/17366)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.transpose in favor of inlining pylibcudf ([#17365](https://github.com/rapidsai/cudf/pull/17365)) [@mroeschke](https://github.com/mroeschke) -- Move make_strings_column benchmark to nvbench ([#17340](https://github.com/rapidsai/cudf/pull/17340)) [@davidwendt](https://github.com/davidwendt) -- Improve strings contains/find performance for smaller strings ([#17330](https://github.com/rapidsai/cudf/pull/17330)) [@davidwendt](https://github.com/davidwendt) -- Use rapids-logger to generate the cudf logger ([#17307](https://github.com/rapidsai/cudf/pull/17307)) [@vyasr](https://github.com/vyasr) -- Mukernels strings ([#17286](https://github.com/rapidsai/cudf/pull/17286)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Add write_parquet to pylibcudf ([#17263](https://github.com/rapidsai/cudf/pull/17263)) [@mroeschke](https://github.com/mroeschke) -- Single-partition Dask executor for cuDF-Polars ([#17262](https://github.com/rapidsai/cudf/pull/17262)) [@rjzamora](https://github.com/rjzamora) -- Add breaking change workflow trigger ([#17248](https://github.com/rapidsai/cudf/pull/17248)) [@AyodeAwe](https://github.com/AyodeAwe) -- Precompute AST arity ([#17234](https://github.com/rapidsai/cudf/pull/17234)) [@bdice](https://github.com/bdice) -- Update to CCCL 2.7.0-rc2. ([#17233](https://github.com/rapidsai/cudf/pull/17233)) [@bdice](https://github.com/bdice) -- Make `column_empty` mask buffer creation consistent with libcudf ([#16715](https://github.com/rapidsai/cudf/pull/16715)) [@mroeschke](https://github.com/mroeschke) +- Remove pandas backend from `cudf.pandas` - ibis integration tests ([#17945](https://github.com/NVIDIA/cudf/pull/17945)) [@Matt711](https://github.com/Matt711) +- Revert CUDA 12.8 shared workflow branch changes ([#17879](https://github.com/NVIDIA/cudf/pull/17879)) [@vyasr](https://github.com/vyasr) +- Remove predicate param from `DataFrameScan` IR ([#17852](https://github.com/NVIDIA/cudf/pull/17852)) [@Matt711](https://github.com/Matt711) +- Remove cudf.Scalar from scatter APIs ([#17847](https://github.com/NVIDIA/cudf/pull/17847)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf.Scalar from interval_range ([#17844](https://github.com/NVIDIA/cudf/pull/17844)) [@mroeschke](https://github.com/mroeschke) +- Add `verify-codeowners` hook ([#17840](https://github.com/NVIDIA/cudf/pull/17840)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Build and test with CUDA 12.8.0 ([#17834](https://github.com/NVIDIA/cudf/pull/17834)) [@bdice](https://github.com/bdice) +- Increase timeout for recently added test ([#17829](https://github.com/NVIDIA/cudf/pull/17829)) [@galipremsagar](https://github.com/galipremsagar) +- Apply ruff everywhere (notebooks and scripts) ([#17820](https://github.com/NVIDIA/cudf/pull/17820)) [@bdice](https://github.com/bdice) +- Fix pre-commit.ci failures ([#17819](https://github.com/NVIDIA/cudf/pull/17819)) [@bdice](https://github.com/bdice) +- Remove incorrect calls to set architectures ([#17813](https://github.com/NVIDIA/cudf/pull/17813)) [@vyasr](https://github.com/vyasr) +- Fix typo in exception raised when attempting to convert a string column to cupy ([#17800](https://github.com/NVIDIA/cudf/pull/17800)) [@dagardner-nv](https://github.com/dagardner-nv) +- Add support for `pyarrow-19` ([#17794](https://github.com/NVIDIA/cudf/pull/17794)) [@galipremsagar](https://github.com/galipremsagar) +- increase parallelism in nightly builds ([#17792](https://github.com/NVIDIA/cudf/pull/17792)) [@jameslamb](https://github.com/jameslamb) +- Reduce libcudf memcheck tests output ([#17791](https://github.com/NVIDIA/cudf/pull/17791)) [@davidwendt](https://github.com/davidwendt) +- Make cudf build with latest CCCL ([#17788](https://github.com/NVIDIA/cudf/pull/17788)) [@miscco](https://github.com/miscco) +- Introduce some more rolling window benchmarks ([#17787](https://github.com/NVIDIA/cudf/pull/17787)) [@wence-](https://github.com/wence-) +- Add shellcheck to pre-commit and fix warnings ([#17778](https://github.com/NVIDIA/cudf/pull/17778)) [@gforsyth](https://github.com/gforsyth) +- Improve parquet reader very-long string performance ([#17773](https://github.com/NVIDIA/cudf/pull/17773)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Update how to manage host UDF instance ([#17770](https://github.com/NVIDIA/cudf/pull/17770)) [@res-life](https://github.com/res-life) +- Add getInts api for HostMemoryBuffer and UnsafeMemoryAccessor ([#17767](https://github.com/NVIDIA/cudf/pull/17767)) [@liurenjie1024](https://github.com/liurenjie1024) +- Expose stream-ordering in scalar and avro APIs ([#17766](https://github.com/NVIDIA/cudf/pull/17766)) [@shrshi](https://github.com/shrshi) +- Standarize methods used from `cudf.core._internals` ([#17765](https://github.com/NVIDIA/cudf/pull/17765)) [@mroeschke](https://github.com/mroeschke) +- Implement string join in cudf-polars ([#17755](https://github.com/NVIDIA/cudf/pull/17755)) [@wence-](https://github.com/wence-) +- Deprecate dataframe protocol ([#17736](https://github.com/NVIDIA/cudf/pull/17736)) [@vyasr](https://github.com/vyasr) +- Add parquet reader long row test ([#17735](https://github.com/NVIDIA/cudf/pull/17735)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Update kvikio call due to upstream changes ([#17733](https://github.com/NVIDIA/cudf/pull/17733)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Delay setting MultiIndex.level/codes until needed ([#17728](https://github.com/NVIDIA/cudf/pull/17728)) [@mroeschke](https://github.com/mroeschke) +- Bounding pool size in multi-batch JSON reader ([#17724](https://github.com/NVIDIA/cudf/pull/17724)) [@shrshi](https://github.com/shrshi) +- Use GCC 13 in CUDA 12 conda builds. ([#17721](https://github.com/NVIDIA/cudf/pull/17721)) [@bdice](https://github.com/bdice) +- Update minimal sphinx theme version so that we can use parallel doc builds ([#17719](https://github.com/NVIDIA/cudf/pull/17719)) [@vyasr](https://github.com/vyasr) +- Add more aggregation methods in pylibcudf ([#17717](https://github.com/NVIDIA/cudf/pull/17717)) [@mroeschke](https://github.com/mroeschke) +- Make cudf._lib.string_udf work with pylibcudf Columns instead of cudf._lib Columns ([#17715](https://github.com/NVIDIA/cudf/pull/17715)) [@mroeschke](https://github.com/mroeschke) +- Add special orc test data: timestamp interspersed with null values ([#17713](https://github.com/NVIDIA/cudf/pull/17713)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Add pylibcudf.null_mask.null_count ([#17711](https://github.com/NVIDIA/cudf/pull/17711)) [@mroeschke](https://github.com/mroeschke) +- Ensure pyarrow.Scalar to pylibcudf.Scalar is cached ([#17707](https://github.com/NVIDIA/cudf/pull/17707)) [@mroeschke](https://github.com/mroeschke) +- Adapt cudf numba config for numba 0.61 removal ([#17705](https://github.com/NVIDIA/cudf/pull/17705)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.scalar in favor of pylibcudf ([#17701](https://github.com/NVIDIA/cudf/pull/17701)) [@mroeschke](https://github.com/mroeschke) +- Fix parquet reader list bug ([#17699](https://github.com/NVIDIA/cudf/pull/17699)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Migrated Dynamic AST Expression Trees in Benchmarks and Tests to use AST Tree ([#17697](https://github.com/NVIDIA/cudf/pull/17697)) [@lamarrr](https://github.com/lamarrr) +- Skip polars test that can generate timezones that chrono_tz doesn't know ([#17694](https://github.com/NVIDIA/cudf/pull/17694)) [@wence-](https://github.com/wence-) +- Use 64-bit offsets only if the current strings column output chunk size exceeds threshold ([#17693](https://github.com/NVIDIA/cudf/pull/17693)) [@mhaseeb123](https://github.com/mhaseeb123) +- Use latest ci-conda images ([#17690](https://github.com/NVIDIA/cudf/pull/17690)) [@bdice](https://github.com/bdice) +- Add multi-source reading to JSON reader benchmarks ([#17688](https://github.com/NVIDIA/cudf/pull/17688)) [@shrshi](https://github.com/shrshi) +- Convert cudf.Scalar usage to pylibcudf and pyarrow usage ([#17686](https://github.com/NVIDIA/cudf/pull/17686)) [@mroeschke](https://github.com/mroeschke) +- remove find_package(Python) in libcudf build ([#17683](https://github.com/NVIDIA/cudf/pull/17683)) [@jameslamb](https://github.com/jameslamb) +- Fix build metrics report format with long placehold filenames ([#17679](https://github.com/NVIDIA/cudf/pull/17679)) [@davidwendt](https://github.com/davidwendt) +- Use rapids-cmake for the logger ([#17674](https://github.com/NVIDIA/cudf/pull/17674)) [@vyasr](https://github.com/vyasr) +- Java Parquet reads via multiple host buffers ([#17673](https://github.com/NVIDIA/cudf/pull/17673)) [@jlowe](https://github.com/jlowe) +- Remove cudf._libs.types.pyx ([#17665](https://github.com/NVIDIA/cudf/pull/17665)) [@mroeschke](https://github.com/mroeschke) +- Add support for `Groupby.cumprod` ([#17661](https://github.com/NVIDIA/cudf/pull/17661)) [@galipremsagar](https://github.com/galipremsagar) +- Implement `.dt.total_seconds` ([#17659](https://github.com/NVIDIA/cudf/pull/17659)) [@galipremsagar](https://github.com/galipremsagar) +- Avoid shallow copies in groupby methods ([#17646](https://github.com/NVIDIA/cudf/pull/17646)) [@mroeschke](https://github.com/mroeschke) +- Avoid double MultiIndex factorization in groupby index result ([#17644](https://github.com/NVIDIA/cudf/pull/17644)) [@mroeschke](https://github.com/mroeschke) +- Add seed parameter to hash_character_ngrams ([#17643](https://github.com/NVIDIA/cudf/pull/17643)) [@davidwendt](https://github.com/davidwendt) +- Fix possible overflow in WriteCoalescingCallbackWrapper::TearDown ([#17642](https://github.com/NVIDIA/cudf/pull/17642)) [@davidwendt](https://github.com/davidwendt) +- Remove pragma GCC diagnostic from source files ([#17637](https://github.com/NVIDIA/cudf/pull/17637)) [@davidwendt](https://github.com/davidwendt) +- Move unnecessary utilities from cudf._lib.scalar ([#17636](https://github.com/NVIDIA/cudf/pull/17636)) [@mroeschke](https://github.com/mroeschke) +- Support compression= in DataFrame.to_json ([#17634](https://github.com/NVIDIA/cudf/pull/17634)) [@mroeschke](https://github.com/mroeschke) +- Bump Polars version to <1.18 ([#17632](https://github.com/NVIDIA/cudf/pull/17632)) [@Matt711](https://github.com/Matt711) +- Add public APIs to Access Underlying `cudf` and `pandas` Objects from `cudf.pandas` Proxy Objects ([#17629](https://github.com/NVIDIA/cudf/pull/17629)) [@galipremsagar](https://github.com/galipremsagar) +- Use Numba Config to turn on Pynvjitlink Features ([#17628](https://github.com/NVIDIA/cudf/pull/17628)) [@isVoid](https://github.com/isVoid) +- Use PyNVML 12 ([#17627](https://github.com/NVIDIA/cudf/pull/17627)) [@jakirkham](https://github.com/jakirkham) +- Remove cudf._lib.utils in favor of python APIs ([#17625](https://github.com/NVIDIA/cudf/pull/17625)) [@mroeschke](https://github.com/mroeschke) +- Performance improvements and simplifications for fixed size row-based rolling windows ([#17623](https://github.com/NVIDIA/cudf/pull/17623)) [@wence-](https://github.com/wence-) +- Fix return types for MurmurHash3_x86_32 template specializations ([#17622](https://github.com/NVIDIA/cudf/pull/17622)) [@davidwendt](https://github.com/davidwendt) +- Clean up namespaces and improve compression-related headers ([#17621](https://github.com/NVIDIA/cudf/pull/17621)) [@vuule](https://github.com/vuule) +- Use more pylibcudf.types instead of cudf._lib.types ([#17619](https://github.com/NVIDIA/cudf/pull/17619)) [@mroeschke](https://github.com/mroeschke) +- Remove patch that is only needed for clang-tidy to run on test files ([#17618](https://github.com/NVIDIA/cudf/pull/17618)) [@vyasr](https://github.com/vyasr) +- update telemetry actions to fluent-bit friendly style ([#17615](https://github.com/NVIDIA/cudf/pull/17615)) [@msarahan](https://github.com/msarahan) +- Introduce some simple benchmarks for rolling window aggregations ([#17613](https://github.com/NVIDIA/cudf/pull/17613)) [@wence-](https://github.com/wence-) +- Bump the oldest `pyarrow` version to `14.0.2` in test matrix ([#17611](https://github.com/NVIDIA/cudf/pull/17611)) [@galipremsagar](https://github.com/galipremsagar) +- Use `[[nodiscard]]` attribute before `__device__` ([#17608](https://github.com/NVIDIA/cudf/pull/17608)) [@vuule](https://github.com/vuule) +- Use `host_vector` in `flatten_single_pass_aggs` ([#17605](https://github.com/NVIDIA/cudf/pull/17605)) [@vuule](https://github.com/vuule) +- Stop memory_resource.hpp from including itself ([#17603](https://github.com/NVIDIA/cudf/pull/17603)) [@vyasr](https://github.com/vyasr) +- Replace the outdated cuco window concept with buckets ([#17602](https://github.com/NVIDIA/cudf/pull/17602)) [@PointKernel](https://github.com/PointKernel) +- Check if nightlies have succeeded recently enough ([#17596](https://github.com/NVIDIA/cudf/pull/17596)) [@vyasr](https://github.com/vyasr) +- Deprecate cudf::grouped_time_range_rolling_window ([#17589](https://github.com/NVIDIA/cudf/pull/17589)) [@wence-](https://github.com/wence-) +- A couple of fixes in rapids-logger usage ([#17588](https://github.com/NVIDIA/cudf/pull/17588)) [@vyasr](https://github.com/vyasr) +- Simplify expression transformer in Parquet predicate pushdown with `ast::tree` ([#17587](https://github.com/NVIDIA/cudf/pull/17587)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove unused functionality in cudf._lib.utils.pyx ([#17586](https://github.com/NVIDIA/cudf/pull/17586)) [@mroeschke](https://github.com/mroeschke) +- Use cuda-python `cuda.bindings` import names. ([#17585](https://github.com/NVIDIA/cudf/pull/17585)) [@bdice](https://github.com/bdice) +- Use no-sync copy for fixed-width types in cudf::concatenate ([#17584](https://github.com/NVIDIA/cudf/pull/17584)) [@davidwendt](https://github.com/davidwendt) +- Remove cudf._lib.groupby in favor of inlining pylibcudf ([#17582](https://github.com/NVIDIA/cudf/pull/17582)) [@mroeschke](https://github.com/mroeschke) +- Remove unused code of json schema in JSON reader ([#17581](https://github.com/NVIDIA/cudf/pull/17581)) [@karthikeyann](https://github.com/karthikeyann) +- Expose Scalar's constructor and `Scalar#getScalarHandle()` to public ([#17580](https://github.com/NVIDIA/cudf/pull/17580)) [@ttnghia](https://github.com/ttnghia) +- Allow large strings in nvtext benchmarks ([#17579](https://github.com/NVIDIA/cudf/pull/17579)) [@davidwendt](https://github.com/davidwendt) +- Remove cudf._lib.reduce in favor of inlining pylibcudf ([#17574](https://github.com/NVIDIA/cudf/pull/17574)) [@mroeschke](https://github.com/mroeschke) +- Use batched memcpy when writing ORC statistics ([#17572](https://github.com/NVIDIA/cudf/pull/17572)) [@vuule](https://github.com/vuule) +- Allow large strings in nvbench strings benchmarks ([#17571](https://github.com/NVIDIA/cudf/pull/17571)) [@davidwendt](https://github.com/davidwendt) +- Update version references in workflow ([#17568](https://github.com/NVIDIA/cudf/pull/17568)) [@AyodeAwe](https://github.com/AyodeAwe) +- Enable all json reader options in pylibcudf read_json ([#17563](https://github.com/NVIDIA/cudf/pull/17563)) [@karthikeyann](https://github.com/karthikeyann) +- Remove cudf._lib.parquet in favor of inlining pylibcudf ([#17562](https://github.com/NVIDIA/cudf/pull/17562)) [@mroeschke](https://github.com/mroeschke) +- Fix CMake format in cudf/_lib/CMakeLists.txt ([#17559](https://github.com/NVIDIA/cudf/pull/17559)) [@mroeschke](https://github.com/mroeschke) +- Remove "legacy" Dask DataFrame support from Dask cuDF ([#17558](https://github.com/NVIDIA/cudf/pull/17558)) [@rjzamora](https://github.com/rjzamora) +- Replace direct `cudaMemcpyAsync` calls with utility functions (within `/include`) ([#17557](https://github.com/NVIDIA/cudf/pull/17557)) [@vuule](https://github.com/vuule) +- Remove cudf._lib.interop in favor of inlining pylibcudf ([#17555](https://github.com/NVIDIA/cudf/pull/17555)) [@mroeschke](https://github.com/mroeschke) +- gate telemetry dispatch calls on TELEMETRY_ENABLED env var ([#17551](https://github.com/NVIDIA/cudf/pull/17551)) [@msarahan](https://github.com/msarahan) +- Replace direct `cudaMemcpyAsync` calls with utility functions (within `/src`) ([#17550](https://github.com/NVIDIA/cudf/pull/17550)) [@vuule](https://github.com/vuule) +- Remove unused `BufferArrayFromVector` ([#17549](https://github.com/NVIDIA/cudf/pull/17549)) [@Matt711](https://github.com/Matt711) +- Move cudf._lib.copying to cudf.core._internals ([#17548](https://github.com/NVIDIA/cudf/pull/17548)) [@mroeschke](https://github.com/mroeschke) +- Update cuda-python lower bounds to 12.6.2 / 11.8.5 ([#17547](https://github.com/NVIDIA/cudf/pull/17547)) [@bdice](https://github.com/bdice) +- Fix typos, rename types, and add null_probability benchmark axis for distinct ([#17546](https://github.com/NVIDIA/cudf/pull/17546)) [@PointKernel](https://github.com/PointKernel) +- Mark more constexpr functions as device-available ([#17545](https://github.com/NVIDIA/cudf/pull/17545)) [@vyasr](https://github.com/vyasr) +- Use cooperative-groups instead of cub warp-reduce for strings contains ([#17540](https://github.com/NVIDIA/cudf/pull/17540)) [@davidwendt](https://github.com/davidwendt) +- Remove cudf._lib.nvtext in favor of inlining pylibcudf ([#17535](https://github.com/NVIDIA/cudf/pull/17535)) [@mroeschke](https://github.com/mroeschke) +- Add XXHash_32 hasher ([#17533](https://github.com/NVIDIA/cudf/pull/17533)) [@PointKernel](https://github.com/PointKernel) +- Remove unused masked keyword in column_empty ([#17530](https://github.com/NVIDIA/cudf/pull/17530)) [@mroeschke](https://github.com/mroeschke) +- Remove Thrust patch in favor of CMake definition for Thrust 32-bit offset types. ([#17527](https://github.com/NVIDIA/cudf/pull/17527)) [@bdice](https://github.com/bdice) +- [JNI] Enables fabric handles for CUDA async memory pools ([#17526](https://github.com/NVIDIA/cudf/pull/17526)) [@abellina](https://github.com/abellina) +- Force Thrust to use 32-bit offset type. ([#17523](https://github.com/NVIDIA/cudf/pull/17523)) [@bdice](https://github.com/bdice) +- Replace cudf::detail::copy_if logic with thrust::copy_if and gather ([#17520](https://github.com/NVIDIA/cudf/pull/17520)) [@davidwendt](https://github.com/davidwendt) +- Replaces uses of `cudf._lib.Column.from_unique_ptr` with `pylibcudf.Column.from_libcudf` ([#17517](https://github.com/NVIDIA/cudf/pull/17517)) [@Matt711](https://github.com/Matt711) +- Move cudf._lib.aggregation to cudf.core._internals ([#17516](https://github.com/NVIDIA/cudf/pull/17516)) [@mroeschke](https://github.com/mroeschke) +- Migrate copy_column and Column.from_scalar to pylibcudf ([#17513](https://github.com/NVIDIA/cudf/pull/17513)) [@Matt711](https://github.com/Matt711) +- Remove cudf._lib.transform in favor of inlining pylibcudf ([#17505](https://github.com/NVIDIA/cudf/pull/17505)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.string.convert/split in favor of inlining pylibcudf ([#17496](https://github.com/NVIDIA/cudf/pull/17496)) [@mroeschke](https://github.com/mroeschke) +- Move cudf._lib.sort to cudf.core._internals ([#17488](https://github.com/NVIDIA/cudf/pull/17488)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.csv in favor in inlining pylibcudf ([#17485](https://github.com/NVIDIA/cudf/pull/17485)) [@mroeschke](https://github.com/mroeschke) +- Update PyTorch to >=2.4.0 to get fix for CUDA array interface bug, and drop CUDA 11 PyTorch tests. ([#17475](https://github.com/NVIDIA/cudf/pull/17475)) [@bdice](https://github.com/bdice) +- Remove cudf._lib.binops in favor of inlining pylibcudf ([#17468](https://github.com/NVIDIA/cudf/pull/17468)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.orc in favor of inlining pylibcudf ([#17466](https://github.com/NVIDIA/cudf/pull/17466)) [@mroeschke](https://github.com/mroeschke) +- skip most CI on devcontainer-only changes ([#17465](https://github.com/NVIDIA/cudf/pull/17465)) [@jameslamb](https://github.com/jameslamb) +- Set build type for all examples ([#17463](https://github.com/NVIDIA/cudf/pull/17463)) [@vyasr](https://github.com/vyasr) +- Update the hook versions in pre-commit ([#17462](https://github.com/NVIDIA/cudf/pull/17462)) [@wence-](https://github.com/wence-) +- Remove cudf._lib.string_casting in favor of inlining pylibcudf ([#17460](https://github.com/NVIDIA/cudf/pull/17460)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.filling in favor of inlining pylibcudf ([#17459](https://github.com/NVIDIA/cudf/pull/17459)) [@mroeschke](https://github.com/mroeschke) +- Update MurmurHash3_x64_128 to use the cuco equivalent implementation ([#17457](https://github.com/NVIDIA/cudf/pull/17457)) [@PointKernel](https://github.com/PointKernel) +- Move cudf._lib.stream_compaction to cudf.core._internals ([#17456](https://github.com/NVIDIA/cudf/pull/17456)) [@mroeschke](https://github.com/mroeschke) +- Clean up xxhash_64 implementations ([#17455](https://github.com/NVIDIA/cudf/pull/17455)) [@PointKernel](https://github.com/PointKernel) +- Update Hadoop dependency in Java pom ([#17454](https://github.com/NVIDIA/cudf/pull/17454)) [@jlowe](https://github.com/jlowe) +- Adapt to rmm logger changes ([#17451](https://github.com/NVIDIA/cudf/pull/17451)) [@vyasr](https://github.com/vyasr) +- Require approval to run CI on draft PRs ([#17450](https://github.com/NVIDIA/cudf/pull/17450)) [@bdice](https://github.com/bdice) +- Expose stream-ordering in nvtext API ([#17446](https://github.com/NVIDIA/cudf/pull/17446)) [@shrshi](https://github.com/shrshi) +- Use exec_policy_nosync in write_json ([#17445](https://github.com/NVIDIA/cudf/pull/17445)) [@karthikeyann](https://github.com/karthikeyann) +- Remove cudf._lib.json in favor of inlining pylibcudf ([#17443](https://github.com/NVIDIA/cudf/pull/17443)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.null_mask in favor of inlining pylibcudf ([#17440](https://github.com/NVIDIA/cudf/pull/17440)) [@mroeschke](https://github.com/mroeschke) +- Expose stream-ordering in replace API ([#17436](https://github.com/NVIDIA/cudf/pull/17436)) [@shrshi](https://github.com/shrshi) +- Expose stream-ordering in copying APIs ([#17435](https://github.com/NVIDIA/cudf/pull/17435)) [@shrshi](https://github.com/shrshi) +- Expose stream-ordering in column view APIs ([#17434](https://github.com/NVIDIA/cudf/pull/17434)) [@shrshi](https://github.com/shrshi) +- Apply clang-tidy autofixes from new rules ([#17431](https://github.com/NVIDIA/cudf/pull/17431)) [@vyasr](https://github.com/vyasr) +- Remove cudf._lib.round in favor of inlining pylibcudf ([#17430](https://github.com/NVIDIA/cudf/pull/17430)) [@mroeschke](https://github.com/mroeschke) +- Update MurmurHash3_x86_32 to use the cuco equivalent implementation ([#17429](https://github.com/NVIDIA/cudf/pull/17429)) [@PointKernel](https://github.com/PointKernel) +- Remove cudf._lib.replace in favor of inlining pylibcudf ([#17428](https://github.com/NVIDIA/cudf/pull/17428)) [@mroeschke](https://github.com/mroeschke) +- Remove nvtx/ranges.hpp include from cuda.cuh ([#17427](https://github.com/NVIDIA/cudf/pull/17427)) [@davidwendt](https://github.com/davidwendt) +- Remove the unused detail `int_fastdiv.h` header ([#17426](https://github.com/NVIDIA/cudf/pull/17426)) [@PointKernel](https://github.com/PointKernel) +- Remove cudf._lib.lists in favor of inlining pylibcudf ([#17425](https://github.com/NVIDIA/cudf/pull/17425)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.quantile ([#17424](https://github.com/NVIDIA/cudf/pull/17424)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.rolling in favor of inlining pylibcudf ([#17423](https://github.com/NVIDIA/cudf/pull/17423)) [@mroeschke](https://github.com/mroeschke) +- Avoid converting Decimal32/Decimal64 in `to_arrow` and `from_arrow` APIs ([#17422](https://github.com/NVIDIA/cudf/pull/17422)) [@zeroshade](https://github.com/zeroshade) +- Rework minhash APIs for deprecation cycle ([#17421](https://github.com/NVIDIA/cudf/pull/17421)) [@davidwendt](https://github.com/davidwendt) +- Use thread_index_type in binary-ops jit kernel.cu ([#17420](https://github.com/NVIDIA/cudf/pull/17420)) [@davidwendt](https://github.com/davidwendt) +- Change binops for-each kernel to thrust::for_each_n ([#17419](https://github.com/NVIDIA/cudf/pull/17419)) [@davidwendt](https://github.com/davidwendt) +- Move cudf._lib.search to cudf.core._internals ([#17411](https://github.com/NVIDIA/cudf/pull/17411)) [@mroeschke](https://github.com/mroeschke) +- Use grid_1d utilities in copy_range.cuh ([#17409](https://github.com/NVIDIA/cudf/pull/17409)) [@davidwendt](https://github.com/davidwendt) +- Remove cudf._lib.text in favor of inlining pylibcudf ([#17408](https://github.com/NVIDIA/cudf/pull/17408)) [@mroeschke](https://github.com/mroeschke) +- Run clang-tidy checks in PR CI ([#17407](https://github.com/NVIDIA/cudf/pull/17407)) [@bdice](https://github.com/bdice) +- Update strings/text source to use grid_1d for thread/block/stride calculations ([#17404](https://github.com/NVIDIA/cudf/pull/17404)) [@davidwendt](https://github.com/davidwendt) +- Expose stream-ordering to strings attribute APIs ([#17398](https://github.com/NVIDIA/cudf/pull/17398)) [@shrshi](https://github.com/shrshi) +- Expose stream-ordering to interop APIs ([#17397](https://github.com/NVIDIA/cudf/pull/17397)) [@shrshi](https://github.com/shrshi) +- Remove unused type aliases ([#17396](https://github.com/NVIDIA/cudf/pull/17396)) [@PointKernel](https://github.com/PointKernel) +- Remove some cudf._lib.strings files in favor of inlining pylibcudf ([#17394](https://github.com/NVIDIA/cudf/pull/17394)) [@mroeschke](https://github.com/mroeschke) +- Update xxhash_64 to utilize the cuco equivalent implementation ([#17393](https://github.com/NVIDIA/cudf/pull/17393)) [@PointKernel](https://github.com/PointKernel) +- Change indices for dictionary column to signed integer type ([#17390](https://github.com/NVIDIA/cudf/pull/17390)) [@davidwendt](https://github.com/davidwendt) +- Return categorical values in to_numpy/to_cupy ([#17388](https://github.com/NVIDIA/cudf/pull/17388)) [@mroeschke](https://github.com/mroeschke) +- Forward-merge branch-24.12 to branch-25.02 ([#17379](https://github.com/NVIDIA/cudf/pull/17379)) [@bdice](https://github.com/bdice) +- Remove unused IO utilities from cudf python ([#17374](https://github.com/NVIDIA/cudf/pull/17374)) [@Matt711](https://github.com/Matt711) +- Remove cudf._lib.datetime in favor of inlining pylibcudf ([#17372](https://github.com/NVIDIA/cudf/pull/17372)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.join in favor of inlining pylibcudf ([#17371](https://github.com/NVIDIA/cudf/pull/17371)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.merge in favor of inlining pylibcudf ([#17370](https://github.com/NVIDIA/cudf/pull/17370)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.partitioning in favor of inlining pylibcudf ([#17369](https://github.com/NVIDIA/cudf/pull/17369)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.reshape in favor of inlining pylibcudf ([#17368](https://github.com/NVIDIA/cudf/pull/17368)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.timezone in favor of inlining pylibcudf ([#17366](https://github.com/NVIDIA/cudf/pull/17366)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.transpose in favor of inlining pylibcudf ([#17365](https://github.com/NVIDIA/cudf/pull/17365)) [@mroeschke](https://github.com/mroeschke) +- Move make_strings_column benchmark to nvbench ([#17340](https://github.com/NVIDIA/cudf/pull/17340)) [@davidwendt](https://github.com/davidwendt) +- Improve strings contains/find performance for smaller strings ([#17330](https://github.com/NVIDIA/cudf/pull/17330)) [@davidwendt](https://github.com/davidwendt) +- Use rapids-logger to generate the cudf logger ([#17307](https://github.com/NVIDIA/cudf/pull/17307)) [@vyasr](https://github.com/vyasr) +- Mukernels strings ([#17286](https://github.com/NVIDIA/cudf/pull/17286)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Add write_parquet to pylibcudf ([#17263](https://github.com/NVIDIA/cudf/pull/17263)) [@mroeschke](https://github.com/mroeschke) +- Single-partition Dask executor for cuDF-Polars ([#17262](https://github.com/NVIDIA/cudf/pull/17262)) [@rjzamora](https://github.com/rjzamora) +- Add breaking change workflow trigger ([#17248](https://github.com/NVIDIA/cudf/pull/17248)) [@AyodeAwe](https://github.com/AyodeAwe) +- Precompute AST arity ([#17234](https://github.com/NVIDIA/cudf/pull/17234)) [@bdice](https://github.com/bdice) +- Update to CCCL 2.7.0-rc2. ([#17233](https://github.com/NVIDIA/cudf/pull/17233)) [@bdice](https://github.com/bdice) +- Make `column_empty` mask buffer creation consistent with libcudf ([#16715](https://github.com/NVIDIA/cudf/pull/16715)) [@mroeschke](https://github.com/mroeschke) # cudf 24.12.00 (11 Dec 2024) ## 🚨 Breaking Changes -- Fix reading Parquet string cols when `nrows` and `input_pass_limit` > 0 ([#17321](https://github.com/rapidsai/cudf/pull/17321)) [@mhaseeb123](https://github.com/mhaseeb123) -- prefer wheel-provided libcudf.so in load_library(), use RTLD_LOCAL ([#17316](https://github.com/rapidsai/cudf/pull/17316)) [@jameslamb](https://github.com/jameslamb) -- Deprecate single component extraction methods in libcudf ([#17221](https://github.com/rapidsai/cudf/pull/17221)) [@Matt711](https://github.com/Matt711) -- Move detail header floating_conversion.hpp to detail subdirectory ([#17209](https://github.com/rapidsai/cudf/pull/17209)) [@davidwendt](https://github.com/davidwendt) -- Refactor Dask cuDF legacy code ([#17205](https://github.com/rapidsai/cudf/pull/17205)) [@rjzamora](https://github.com/rjzamora) -- Make HostMemoryBuffer call into the DefaultHostMemoryAllocator ([#17204](https://github.com/rapidsai/cudf/pull/17204)) [@revans2](https://github.com/revans2) -- Remove java reservation ([#17189](https://github.com/rapidsai/cudf/pull/17189)) [@revans2](https://github.com/revans2) -- Separate evaluation logic from `IR` objects in cudf-polars ([#17175](https://github.com/rapidsai/cudf/pull/17175)) [@rjzamora](https://github.com/rjzamora) -- Upgrade to polars 1.11 in cudf-polars ([#17154](https://github.com/rapidsai/cudf/pull/17154)) [@wence-](https://github.com/wence-) -- Remove the additional host register calls initially intended for performance improvement on Grace Hopper ([#17092](https://github.com/rapidsai/cudf/pull/17092)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Correctly set `is_device_accesible` when creating `host_span`s from other container/span types ([#17079](https://github.com/rapidsai/cudf/pull/17079)) [@vuule](https://github.com/vuule) -- Unify treatment of `Expr` and `IR` nodes in cudf-polars DSL ([#17016](https://github.com/rapidsai/cudf/pull/17016)) [@wence-](https://github.com/wence-) -- Deprecate support for directly accessing logger ([#16964](https://github.com/rapidsai/cudf/pull/16964)) [@vyasr](https://github.com/vyasr) -- Made cudftestutil header-only and removed GTest dependency ([#16839](https://github.com/rapidsai/cudf/pull/16839)) [@lamarrr](https://github.com/lamarrr) +- Fix reading Parquet string cols when `nrows` and `input_pass_limit` > 0 ([#17321](https://github.com/NVIDIA/cudf/pull/17321)) [@mhaseeb123](https://github.com/mhaseeb123) +- prefer wheel-provided libcudf.so in load_library(), use RTLD_LOCAL ([#17316](https://github.com/NVIDIA/cudf/pull/17316)) [@jameslamb](https://github.com/jameslamb) +- Deprecate single component extraction methods in libcudf ([#17221](https://github.com/NVIDIA/cudf/pull/17221)) [@Matt711](https://github.com/Matt711) +- Move detail header floating_conversion.hpp to detail subdirectory ([#17209](https://github.com/NVIDIA/cudf/pull/17209)) [@davidwendt](https://github.com/davidwendt) +- Refactor Dask cuDF legacy code ([#17205](https://github.com/NVIDIA/cudf/pull/17205)) [@rjzamora](https://github.com/rjzamora) +- Make HostMemoryBuffer call into the DefaultHostMemoryAllocator ([#17204](https://github.com/NVIDIA/cudf/pull/17204)) [@revans2](https://github.com/revans2) +- Remove java reservation ([#17189](https://github.com/NVIDIA/cudf/pull/17189)) [@revans2](https://github.com/revans2) +- Separate evaluation logic from `IR` objects in cudf-polars ([#17175](https://github.com/NVIDIA/cudf/pull/17175)) [@rjzamora](https://github.com/rjzamora) +- Upgrade to polars 1.11 in cudf-polars ([#17154](https://github.com/NVIDIA/cudf/pull/17154)) [@wence-](https://github.com/wence-) +- Remove the additional host register calls initially intended for performance improvement on Grace Hopper ([#17092](https://github.com/NVIDIA/cudf/pull/17092)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Correctly set `is_device_accesible` when creating `host_span`s from other container/span types ([#17079](https://github.com/NVIDIA/cudf/pull/17079)) [@vuule](https://github.com/vuule) +- Unify treatment of `Expr` and `IR` nodes in cudf-polars DSL ([#17016](https://github.com/NVIDIA/cudf/pull/17016)) [@wence-](https://github.com/wence-) +- Deprecate support for directly accessing logger ([#16964](https://github.com/NVIDIA/cudf/pull/16964)) [@vyasr](https://github.com/vyasr) +- Made cudftestutil header-only and removed GTest dependency ([#16839](https://github.com/NVIDIA/cudf/pull/16839)) [@lamarrr](https://github.com/lamarrr) ## 🐛 Bug Fixes -- Turn off cudf.pandas 3rd party integrations tests for 24.12 ([#17500](https://github.com/rapidsai/cudf/pull/17500)) [@Matt711](https://github.com/Matt711) -- Ignore errors when testing glibc versions ([#17389](https://github.com/rapidsai/cudf/pull/17389)) [@vyasr](https://github.com/vyasr) -- Adapt to KvikIO API change in the compatibility mode ([#17377](https://github.com/rapidsai/cudf/pull/17377)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Support pivot with index or column arguments as lists ([#17373](https://github.com/rapidsai/cudf/pull/17373)) [@mroeschke](https://github.com/mroeschke) -- Deselect failing polars tests ([#17362](https://github.com/rapidsai/cudf/pull/17362)) [@pentschev](https://github.com/pentschev) -- Fix integer overflow in compiled binaryop ([#17354](https://github.com/rapidsai/cudf/pull/17354)) [@wence-](https://github.com/wence-) -- Update cmake to 3.28.6 in JNI Dockerfile ([#17342](https://github.com/rapidsai/cudf/pull/17342)) [@jlowe](https://github.com/jlowe) -- fix library-loading issues in editable installs ([#17338](https://github.com/rapidsai/cudf/pull/17338)) [@jameslamb](https://github.com/jameslamb) -- Bug fix: restrict lines=True to JSON format in Kafka read_gdf method ([#17333](https://github.com/rapidsai/cudf/pull/17333)) [@a-hirota](https://github.com/a-hirota) -- Fix various issues with `replace` API and add support in `datetime` and `timedelta` columns ([#17331](https://github.com/rapidsai/cudf/pull/17331)) [@galipremsagar](https://github.com/galipremsagar) -- Do not exclude nanoarrow and flatbuffers from installation if statically linked ([#17322](https://github.com/rapidsai/cudf/pull/17322)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix reading Parquet string cols when `nrows` and `input_pass_limit` > 0 ([#17321](https://github.com/rapidsai/cudf/pull/17321)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove another reference to `FindcuFile` ([#17315](https://github.com/rapidsai/cudf/pull/17315)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Fix reading of single-row unterminated CSV files ([#17305](https://github.com/rapidsai/cudf/pull/17305)) [@vuule](https://github.com/vuule) -- Fixed lifetime issue in ast transform tests ([#17292](https://github.com/rapidsai/cudf/pull/17292)) [@lamarrr](https://github.com/lamarrr) -- Switch to using `TaskSpec` ([#17285](https://github.com/rapidsai/cudf/pull/17285)) [@galipremsagar](https://github.com/galipremsagar) -- Fix data_type ctor call in JSON_TEST ([#17273](https://github.com/rapidsai/cudf/pull/17273)) [@davidwendt](https://github.com/davidwendt) -- Expose delimiter character in JSON reader options to JSON reader APIs ([#17266](https://github.com/rapidsai/cudf/pull/17266)) [@shrshi](https://github.com/shrshi) -- Fix extract-datetime deprecation warning in ndsh benchmark ([#17254](https://github.com/rapidsai/cudf/pull/17254)) [@davidwendt](https://github.com/davidwendt) -- Disallow cuda-python 12.6.1 and 11.8.4 ([#17253](https://github.com/rapidsai/cudf/pull/17253)) [@bdice](https://github.com/bdice) -- Wrap custom iterator result ([#17251](https://github.com/rapidsai/cudf/pull/17251)) [@galipremsagar](https://github.com/galipremsagar) -- Fix binop with LHS numpy datetimelike scalar ([#17226](https://github.com/rapidsai/cudf/pull/17226)) [@mroeschke](https://github.com/mroeschke) -- Fix `Dataframe.__setitem__` slow-downs ([#17222](https://github.com/rapidsai/cudf/pull/17222)) [@galipremsagar](https://github.com/galipremsagar) -- Fix groupby.get_group with length-1 tuple with list-like grouper ([#17216](https://github.com/rapidsai/cudf/pull/17216)) [@mroeschke](https://github.com/mroeschke) -- Fix discoverability of submodules inside `pd.util` ([#17215](https://github.com/rapidsai/cudf/pull/17215)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `Schema.Builder` does not propagate precision value to `Builder` instance ([#17214](https://github.com/rapidsai/cudf/pull/17214)) [@ttnghia](https://github.com/ttnghia) -- Mark column chunks in a PQ reader `pass` as large strings when the cumulative `offsets` exceeds the large strings threshold. ([#17207](https://github.com/rapidsai/cudf/pull/17207)) [@mhaseeb123](https://github.com/mhaseeb123) -- [BUG] Replace `repo_token` with `github_token` in Auto Assign PR GHA ([#17203](https://github.com/rapidsai/cudf/pull/17203)) [@Matt711](https://github.com/Matt711) -- Remove unsanitized nulls from input strings columns in reduction gtests ([#17202](https://github.com/rapidsai/cudf/pull/17202)) [@davidwendt](https://github.com/davidwendt) -- Fix ``to_parquet`` append behavior with global metadata file ([#17198](https://github.com/rapidsai/cudf/pull/17198)) [@rjzamora](https://github.com/rjzamora) -- Check `num_children() == 0` in `Column.from_column_view` ([#17193](https://github.com/rapidsai/cudf/pull/17193)) [@cwharris](https://github.com/cwharris) -- Fix host-to-device copy missing sync in strings/duration convert ([#17149](https://github.com/rapidsai/cudf/pull/17149)) [@davidwendt](https://github.com/davidwendt) -- Add JNI Support for Multi-line Delimiters and Include Test ([#17139](https://github.com/rapidsai/cudf/pull/17139)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Ignore loud dask warnings about legacy dataframe implementation ([#17137](https://github.com/rapidsai/cudf/pull/17137)) [@galipremsagar](https://github.com/galipremsagar) -- Fix the GDS read/write segfault/bus error when the cuFile policy is set to GDS or ALWAYS ([#17122](https://github.com/rapidsai/cudf/pull/17122)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Fix `DataFrame._from_arrays` and introduce validations ([#17112](https://github.com/rapidsai/cudf/pull/17112)) [@galipremsagar](https://github.com/galipremsagar) -- [Bug] Fix Arrow-FS parquet reader for larger files ([#17099](https://github.com/rapidsai/cudf/pull/17099)) [@rjzamora](https://github.com/rjzamora) -- Fix bug in recovering invalid lines in JSONL inputs ([#17098](https://github.com/rapidsai/cudf/pull/17098)) [@shrshi](https://github.com/shrshi) -- Reenable huge pages for arrow host copying ([#17097](https://github.com/rapidsai/cudf/pull/17097)) [@vyasr](https://github.com/vyasr) -- Correctly set `is_device_accesible` when creating `host_span`s from other container/span types ([#17079](https://github.com/rapidsai/cudf/pull/17079)) [@vuule](https://github.com/vuule) -- Fix ORC reader when using `device_read_async` while the destination device buffers are not ready ([#17074](https://github.com/rapidsai/cudf/pull/17074)) [@ttnghia](https://github.com/ttnghia) -- Fix regex handling of fixed quantifier with 0 range ([#17067](https://github.com/rapidsai/cudf/pull/17067)) [@davidwendt](https://github.com/davidwendt) -- Limit the number of keys to calculate column sizes and page starts in PQ reader to 1B ([#17059](https://github.com/rapidsai/cudf/pull/17059)) [@mhaseeb123](https://github.com/mhaseeb123) -- Adding assertion to check for regular JSON inputs of size greater than `INT_MAX` bytes ([#17057](https://github.com/rapidsai/cudf/pull/17057)) [@shrshi](https://github.com/shrshi) -- bug fix: use `self.ck_consumer` in `poll` method of kafka.py to align with `__init__` ([#17044](https://github.com/rapidsai/cudf/pull/17044)) [@a-hirota](https://github.com/a-hirota) -- Disable kvikio remote I/O to avoid openssl dependencies in JNI build ([#17026](https://github.com/rapidsai/cudf/pull/17026)) [@pxLi](https://github.com/pxLi) -- Fix `host_span` constructor to correctly copy `is_device_accessible` ([#17020](https://github.com/rapidsai/cudf/pull/17020)) [@vuule](https://github.com/vuule) -- Add pinning for pyarrow in wheels ([#17018](https://github.com/rapidsai/cudf/pull/17018)) [@vyasr](https://github.com/vyasr) -- Use std::optional for host types ([#17015](https://github.com/rapidsai/cudf/pull/17015)) [@robertmaynard](https://github.com/robertmaynard) -- Fix write_json to handle empty string column ([#16995](https://github.com/rapidsai/cudf/pull/16995)) [@karthikeyann](https://github.com/karthikeyann) -- Restore export of nvcomp outside of wheel builds ([#16988](https://github.com/rapidsai/cudf/pull/16988)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Allow melt(var_name=) to be a falsy label ([#16981](https://github.com/rapidsai/cudf/pull/16981)) [@mroeschke](https://github.com/mroeschke) -- Fix astype from tz-aware type to tz-aware type ([#16980](https://github.com/rapidsai/cudf/pull/16980)) [@mroeschke](https://github.com/mroeschke) -- Use `libcudf` wheel from PR rather than nightly for `polars-polars` CI test job ([#16975](https://github.com/rapidsai/cudf/pull/16975)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix order-preservation in pandas-compat unsorted groupby ([#16942](https://github.com/rapidsai/cudf/pull/16942)) [@wence-](https://github.com/wence-) -- Fix cudf::strings::findall error with empty input ([#16928](https://github.com/rapidsai/cudf/pull/16928)) [@davidwendt](https://github.com/davidwendt) -- Fix JsonLargeReaderTest.MultiBatch use of LIBCUDF_JSON_BATCH_SIZE env var ([#16927](https://github.com/rapidsai/cudf/pull/16927)) [@davidwendt](https://github.com/davidwendt) -- Parse newline as whitespace character while tokenizing JSONL inputs with non-newline delimiter ([#16923](https://github.com/rapidsai/cudf/pull/16923)) [@shrshi](https://github.com/shrshi) -- Respect groupby.nunique(dropna=False) ([#16921](https://github.com/rapidsai/cudf/pull/16921)) [@mroeschke](https://github.com/mroeschke) -- Update all rmm imports to use pylibrmm/librmm ([#16913](https://github.com/rapidsai/cudf/pull/16913)) [@Matt711](https://github.com/Matt711) -- Fix order-preservation in cudf-polars groupby ([#16907](https://github.com/rapidsai/cudf/pull/16907)) [@wence-](https://github.com/wence-) -- Add a shortcut for when the input clusters are all empty for the tdigest merge ([#16897](https://github.com/rapidsai/cudf/pull/16897)) [@jihoonson](https://github.com/jihoonson) -- Properly handle the mapped and registered regions in `memory_mapped_source` ([#16865](https://github.com/rapidsai/cudf/pull/16865)) [@vuule](https://github.com/vuule) -- Fix performance regression for generate_character_ngrams ([#16849](https://github.com/rapidsai/cudf/pull/16849)) [@davidwendt](https://github.com/davidwendt) -- Fix regex parsing logic handling of nested quantifiers ([#16798](https://github.com/rapidsai/cudf/pull/16798)) [@davidwendt](https://github.com/davidwendt) -- Compute whole column variance using numerically stable approach ([#16448](https://github.com/rapidsai/cudf/pull/16448)) [@wence-](https://github.com/wence-) +- Turn off cudf.pandas 3rd party integrations tests for 24.12 ([#17500](https://github.com/NVIDIA/cudf/pull/17500)) [@Matt711](https://github.com/Matt711) +- Ignore errors when testing glibc versions ([#17389](https://github.com/NVIDIA/cudf/pull/17389)) [@vyasr](https://github.com/vyasr) +- Adapt to KvikIO API change in the compatibility mode ([#17377](https://github.com/NVIDIA/cudf/pull/17377)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Support pivot with index or column arguments as lists ([#17373](https://github.com/NVIDIA/cudf/pull/17373)) [@mroeschke](https://github.com/mroeschke) +- Deselect failing polars tests ([#17362](https://github.com/NVIDIA/cudf/pull/17362)) [@pentschev](https://github.com/pentschev) +- Fix integer overflow in compiled binaryop ([#17354](https://github.com/NVIDIA/cudf/pull/17354)) [@wence-](https://github.com/wence-) +- Update cmake to 3.28.6 in JNI Dockerfile ([#17342](https://github.com/NVIDIA/cudf/pull/17342)) [@jlowe](https://github.com/jlowe) +- fix library-loading issues in editable installs ([#17338](https://github.com/NVIDIA/cudf/pull/17338)) [@jameslamb](https://github.com/jameslamb) +- Bug fix: restrict lines=True to JSON format in Kafka read_gdf method ([#17333](https://github.com/NVIDIA/cudf/pull/17333)) [@a-hirota](https://github.com/a-hirota) +- Fix various issues with `replace` API and add support in `datetime` and `timedelta` columns ([#17331](https://github.com/NVIDIA/cudf/pull/17331)) [@galipremsagar](https://github.com/galipremsagar) +- Do not exclude nanoarrow and flatbuffers from installation if statically linked ([#17322](https://github.com/NVIDIA/cudf/pull/17322)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix reading Parquet string cols when `nrows` and `input_pass_limit` > 0 ([#17321](https://github.com/NVIDIA/cudf/pull/17321)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove another reference to `FindcuFile` ([#17315](https://github.com/NVIDIA/cudf/pull/17315)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Fix reading of single-row unterminated CSV files ([#17305](https://github.com/NVIDIA/cudf/pull/17305)) [@vuule](https://github.com/vuule) +- Fixed lifetime issue in ast transform tests ([#17292](https://github.com/NVIDIA/cudf/pull/17292)) [@lamarrr](https://github.com/lamarrr) +- Switch to using `TaskSpec` ([#17285](https://github.com/NVIDIA/cudf/pull/17285)) [@galipremsagar](https://github.com/galipremsagar) +- Fix data_type ctor call in JSON_TEST ([#17273](https://github.com/NVIDIA/cudf/pull/17273)) [@davidwendt](https://github.com/davidwendt) +- Expose delimiter character in JSON reader options to JSON reader APIs ([#17266](https://github.com/NVIDIA/cudf/pull/17266)) [@shrshi](https://github.com/shrshi) +- Fix extract-datetime deprecation warning in ndsh benchmark ([#17254](https://github.com/NVIDIA/cudf/pull/17254)) [@davidwendt](https://github.com/davidwendt) +- Disallow cuda-python 12.6.1 and 11.8.4 ([#17253](https://github.com/NVIDIA/cudf/pull/17253)) [@bdice](https://github.com/bdice) +- Wrap custom iterator result ([#17251](https://github.com/NVIDIA/cudf/pull/17251)) [@galipremsagar](https://github.com/galipremsagar) +- Fix binop with LHS numpy datetimelike scalar ([#17226](https://github.com/NVIDIA/cudf/pull/17226)) [@mroeschke](https://github.com/mroeschke) +- Fix `Dataframe.__setitem__` slow-downs ([#17222](https://github.com/NVIDIA/cudf/pull/17222)) [@galipremsagar](https://github.com/galipremsagar) +- Fix groupby.get_group with length-1 tuple with list-like grouper ([#17216](https://github.com/NVIDIA/cudf/pull/17216)) [@mroeschke](https://github.com/mroeschke) +- Fix discoverability of submodules inside `pd.util` ([#17215](https://github.com/NVIDIA/cudf/pull/17215)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `Schema.Builder` does not propagate precision value to `Builder` instance ([#17214](https://github.com/NVIDIA/cudf/pull/17214)) [@ttnghia](https://github.com/ttnghia) +- Mark column chunks in a PQ reader `pass` as large strings when the cumulative `offsets` exceeds the large strings threshold. ([#17207](https://github.com/NVIDIA/cudf/pull/17207)) [@mhaseeb123](https://github.com/mhaseeb123) +- [BUG] Replace `repo_token` with `github_token` in Auto Assign PR GHA ([#17203](https://github.com/NVIDIA/cudf/pull/17203)) [@Matt711](https://github.com/Matt711) +- Remove unsanitized nulls from input strings columns in reduction gtests ([#17202](https://github.com/NVIDIA/cudf/pull/17202)) [@davidwendt](https://github.com/davidwendt) +- Fix ``to_parquet`` append behavior with global metadata file ([#17198](https://github.com/NVIDIA/cudf/pull/17198)) [@rjzamora](https://github.com/rjzamora) +- Check `num_children() == 0` in `Column.from_column_view` ([#17193](https://github.com/NVIDIA/cudf/pull/17193)) [@cwharris](https://github.com/cwharris) +- Fix host-to-device copy missing sync in strings/duration convert ([#17149](https://github.com/NVIDIA/cudf/pull/17149)) [@davidwendt](https://github.com/davidwendt) +- Add JNI Support for Multi-line Delimiters and Include Test ([#17139](https://github.com/NVIDIA/cudf/pull/17139)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Ignore loud dask warnings about legacy dataframe implementation ([#17137](https://github.com/NVIDIA/cudf/pull/17137)) [@galipremsagar](https://github.com/galipremsagar) +- Fix the GDS read/write segfault/bus error when the cuFile policy is set to GDS or ALWAYS ([#17122](https://github.com/NVIDIA/cudf/pull/17122)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Fix `DataFrame._from_arrays` and introduce validations ([#17112](https://github.com/NVIDIA/cudf/pull/17112)) [@galipremsagar](https://github.com/galipremsagar) +- [Bug] Fix Arrow-FS parquet reader for larger files ([#17099](https://github.com/NVIDIA/cudf/pull/17099)) [@rjzamora](https://github.com/rjzamora) +- Fix bug in recovering invalid lines in JSONL inputs ([#17098](https://github.com/NVIDIA/cudf/pull/17098)) [@shrshi](https://github.com/shrshi) +- Reenable huge pages for arrow host copying ([#17097](https://github.com/NVIDIA/cudf/pull/17097)) [@vyasr](https://github.com/vyasr) +- Correctly set `is_device_accesible` when creating `host_span`s from other container/span types ([#17079](https://github.com/NVIDIA/cudf/pull/17079)) [@vuule](https://github.com/vuule) +- Fix ORC reader when using `device_read_async` while the destination device buffers are not ready ([#17074](https://github.com/NVIDIA/cudf/pull/17074)) [@ttnghia](https://github.com/ttnghia) +- Fix regex handling of fixed quantifier with 0 range ([#17067](https://github.com/NVIDIA/cudf/pull/17067)) [@davidwendt](https://github.com/davidwendt) +- Limit the number of keys to calculate column sizes and page starts in PQ reader to 1B ([#17059](https://github.com/NVIDIA/cudf/pull/17059)) [@mhaseeb123](https://github.com/mhaseeb123) +- Adding assertion to check for regular JSON inputs of size greater than `INT_MAX` bytes ([#17057](https://github.com/NVIDIA/cudf/pull/17057)) [@shrshi](https://github.com/shrshi) +- bug fix: use `self.ck_consumer` in `poll` method of kafka.py to align with `__init__` ([#17044](https://github.com/NVIDIA/cudf/pull/17044)) [@a-hirota](https://github.com/a-hirota) +- Disable kvikio remote I/O to avoid openssl dependencies in JNI build ([#17026](https://github.com/NVIDIA/cudf/pull/17026)) [@pxLi](https://github.com/pxLi) +- Fix `host_span` constructor to correctly copy `is_device_accessible` ([#17020](https://github.com/NVIDIA/cudf/pull/17020)) [@vuule](https://github.com/vuule) +- Add pinning for pyarrow in wheels ([#17018](https://github.com/NVIDIA/cudf/pull/17018)) [@vyasr](https://github.com/vyasr) +- Use std::optional for host types ([#17015](https://github.com/NVIDIA/cudf/pull/17015)) [@robertmaynard](https://github.com/robertmaynard) +- Fix write_json to handle empty string column ([#16995](https://github.com/NVIDIA/cudf/pull/16995)) [@karthikeyann](https://github.com/karthikeyann) +- Restore export of nvcomp outside of wheel builds ([#16988](https://github.com/NVIDIA/cudf/pull/16988)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Allow melt(var_name=) to be a falsy label ([#16981](https://github.com/NVIDIA/cudf/pull/16981)) [@mroeschke](https://github.com/mroeschke) +- Fix astype from tz-aware type to tz-aware type ([#16980](https://github.com/NVIDIA/cudf/pull/16980)) [@mroeschke](https://github.com/mroeschke) +- Use `libcudf` wheel from PR rather than nightly for `polars-polars` CI test job ([#16975](https://github.com/NVIDIA/cudf/pull/16975)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix order-preservation in pandas-compat unsorted groupby ([#16942](https://github.com/NVIDIA/cudf/pull/16942)) [@wence-](https://github.com/wence-) +- Fix cudf::strings::findall error with empty input ([#16928](https://github.com/NVIDIA/cudf/pull/16928)) [@davidwendt](https://github.com/davidwendt) +- Fix JsonLargeReaderTest.MultiBatch use of LIBCUDF_JSON_BATCH_SIZE env var ([#16927](https://github.com/NVIDIA/cudf/pull/16927)) [@davidwendt](https://github.com/davidwendt) +- Parse newline as whitespace character while tokenizing JSONL inputs with non-newline delimiter ([#16923](https://github.com/NVIDIA/cudf/pull/16923)) [@shrshi](https://github.com/shrshi) +- Respect groupby.nunique(dropna=False) ([#16921](https://github.com/NVIDIA/cudf/pull/16921)) [@mroeschke](https://github.com/mroeschke) +- Update all rmm imports to use pylibrmm/librmm ([#16913](https://github.com/NVIDIA/cudf/pull/16913)) [@Matt711](https://github.com/Matt711) +- Fix order-preservation in cudf-polars groupby ([#16907](https://github.com/NVIDIA/cudf/pull/16907)) [@wence-](https://github.com/wence-) +- Add a shortcut for when the input clusters are all empty for the tdigest merge ([#16897](https://github.com/NVIDIA/cudf/pull/16897)) [@jihoonson](https://github.com/jihoonson) +- Properly handle the mapped and registered regions in `memory_mapped_source` ([#16865](https://github.com/NVIDIA/cudf/pull/16865)) [@vuule](https://github.com/vuule) +- Fix performance regression for generate_character_ngrams ([#16849](https://github.com/NVIDIA/cudf/pull/16849)) [@davidwendt](https://github.com/davidwendt) +- Fix regex parsing logic handling of nested quantifiers ([#16798](https://github.com/NVIDIA/cudf/pull/16798)) [@davidwendt](https://github.com/davidwendt) +- Compute whole column variance using numerically stable approach ([#16448](https://github.com/NVIDIA/cudf/pull/16448)) [@wence-](https://github.com/wence-) ## 📖 Documentation -- Add documentation for low memory readers ([#17314](https://github.com/rapidsai/cudf/pull/17314)) [@btepera](https://github.com/btepera) -- Fix the example in documentation for `get_dremel_data()` ([#17242](https://github.com/rapidsai/cudf/pull/17242)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix some documentation rendering for pylibcudf ([#17217](https://github.com/rapidsai/cudf/pull/17217)) [@mroeschke](https://github.com/mroeschke) -- Move detail header floating_conversion.hpp to detail subdirectory ([#17209](https://github.com/rapidsai/cudf/pull/17209)) [@davidwendt](https://github.com/davidwendt) -- Add TokenizeVocabulary to api docs ([#17208](https://github.com/rapidsai/cudf/pull/17208)) [@davidwendt](https://github.com/davidwendt) -- Add jaccard_index to generated cuDF docs ([#17199](https://github.com/rapidsai/cudf/pull/17199)) [@davidwendt](https://github.com/davidwendt) -- [no ci] Add empty-columns section to the libcudf developer guide ([#17183](https://github.com/rapidsai/cudf/pull/17183)) [@davidwendt](https://github.com/davidwendt) -- Add 2-cpp approvers text to contributing guide [no ci] ([#17182](https://github.com/rapidsai/cudf/pull/17182)) [@davidwendt](https://github.com/davidwendt) -- Changing developer guide int_64_t to int64_t ([#17130](https://github.com/rapidsai/cudf/pull/17130)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- docs: change 'CSV' to 'csv' in python/custreamz/README.md to match kafka.py ([#17041](https://github.com/rapidsai/cudf/pull/17041)) [@a-hirota](https://github.com/a-hirota) -- [DOC] Document limitation using `cudf.pandas` proxy arrays ([#16955](https://github.com/rapidsai/cudf/pull/16955)) [@Matt711](https://github.com/Matt711) -- [DOC] Document environment variable for failing on fallback in `cudf.pandas` ([#16932](https://github.com/rapidsai/cudf/pull/16932)) [@Matt711](https://github.com/Matt711) +- Add documentation for low memory readers ([#17314](https://github.com/NVIDIA/cudf/pull/17314)) [@btepera](https://github.com/btepera) +- Fix the example in documentation for `get_dremel_data()` ([#17242](https://github.com/NVIDIA/cudf/pull/17242)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix some documentation rendering for pylibcudf ([#17217](https://github.com/NVIDIA/cudf/pull/17217)) [@mroeschke](https://github.com/mroeschke) +- Move detail header floating_conversion.hpp to detail subdirectory ([#17209](https://github.com/NVIDIA/cudf/pull/17209)) [@davidwendt](https://github.com/davidwendt) +- Add TokenizeVocabulary to api docs ([#17208](https://github.com/NVIDIA/cudf/pull/17208)) [@davidwendt](https://github.com/davidwendt) +- Add jaccard_index to generated cuDF docs ([#17199](https://github.com/NVIDIA/cudf/pull/17199)) [@davidwendt](https://github.com/davidwendt) +- [no ci] Add empty-columns section to the libcudf developer guide ([#17183](https://github.com/NVIDIA/cudf/pull/17183)) [@davidwendt](https://github.com/davidwendt) +- Add 2-cpp approvers text to contributing guide [no ci] ([#17182](https://github.com/NVIDIA/cudf/pull/17182)) [@davidwendt](https://github.com/davidwendt) +- Changing developer guide int_64_t to int64_t ([#17130](https://github.com/NVIDIA/cudf/pull/17130)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- docs: change 'CSV' to 'csv' in python/custreamz/README.md to match kafka.py ([#17041](https://github.com/NVIDIA/cudf/pull/17041)) [@a-hirota](https://github.com/a-hirota) +- [DOC] Document limitation using `cudf.pandas` proxy arrays ([#16955](https://github.com/NVIDIA/cudf/pull/16955)) [@Matt711](https://github.com/Matt711) +- [DOC] Document environment variable for failing on fallback in `cudf.pandas` ([#16932](https://github.com/NVIDIA/cudf/pull/16932)) [@Matt711](https://github.com/Matt711) ## 🚀 New Features -- Add version config ([#17312](https://github.com/rapidsai/cudf/pull/17312)) [@vyasr](https://github.com/vyasr) -- Java JNI for Multiple contains ([#17281](https://github.com/rapidsai/cudf/pull/17281)) [@res-life](https://github.com/res-life) -- Add `cudf::calendrical_month_sequence` to pylibcudf ([#17277](https://github.com/rapidsai/cudf/pull/17277)) [@Matt711](https://github.com/Matt711) -- Raise errors on specific types of fallback in `cudf.pandas` ([#17268](https://github.com/rapidsai/cudf/pull/17268)) [@Matt711](https://github.com/Matt711) -- Add `catboost` to the third-party integration tests ([#17267](https://github.com/rapidsai/cudf/pull/17267)) [@Matt711](https://github.com/Matt711) -- Add type stubs for pylibcudf ([#17258](https://github.com/rapidsai/cudf/pull/17258)) [@wence-](https://github.com/wence-) -- Use pylibcudf contiguous split APIs in cudf python ([#17246](https://github.com/rapidsai/cudf/pull/17246)) [@Matt711](https://github.com/Matt711) -- Upgrade nvcomp to 4.1.0.6 ([#17201](https://github.com/rapidsai/cudf/pull/17201)) [@bdice](https://github.com/bdice) -- Added Arrow Interop Benchmarks ([#17194](https://github.com/rapidsai/cudf/pull/17194)) [@lamarrr](https://github.com/lamarrr) -- Rewrite Java API `Table.readJSON` to return the output from libcudf `read_json` directly ([#17180](https://github.com/rapidsai/cudf/pull/17180)) [@ttnghia](https://github.com/ttnghia) -- Support storing `precision` of decimal types in `Schema` class ([#17176](https://github.com/rapidsai/cudf/pull/17176)) [@ttnghia](https://github.com/ttnghia) -- Migrate CSV writer to pylibcudf ([#17163](https://github.com/rapidsai/cudf/pull/17163)) [@Matt711](https://github.com/Matt711) -- Add compute_shared_memory_aggs used by shared memory groupby ([#17162](https://github.com/rapidsai/cudf/pull/17162)) [@PointKernel](https://github.com/PointKernel) -- Added ast tree to simplify expression lifetime management ([#17156](https://github.com/rapidsai/cudf/pull/17156)) [@lamarrr](https://github.com/lamarrr) -- Add compute_mapping_indices used by shared memory groupby ([#17147](https://github.com/rapidsai/cudf/pull/17147)) [@PointKernel](https://github.com/PointKernel) -- Add remaining datetime APIs to pylibcudf ([#17143](https://github.com/rapidsai/cudf/pull/17143)) [@Matt711](https://github.com/Matt711) -- Added strings AST vs BINARY_OP benchmarks ([#17128](https://github.com/rapidsai/cudf/pull/17128)) [@lamarrr](https://github.com/lamarrr) -- Use `libcudf_exception_handler` throughout `pylibcudf.libcudf` ([#17109](https://github.com/rapidsai/cudf/pull/17109)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Include timezone file path in error message ([#17102](https://github.com/rapidsai/cudf/pull/17102)) [@bdice](https://github.com/bdice) -- Migrate NVText Byte Pair Encoding APIs to pylibcudf ([#17101](https://github.com/rapidsai/cudf/pull/17101)) [@Matt711](https://github.com/Matt711) -- Migrate NVText Tokenizing APIs to pylibcudf ([#17100](https://github.com/rapidsai/cudf/pull/17100)) [@Matt711](https://github.com/Matt711) -- Migrate NVtext subword tokenizing APIs to pylibcudf ([#17096](https://github.com/rapidsai/cudf/pull/17096)) [@Matt711](https://github.com/Matt711) -- Migrate NVText Stemming APIs to pylibcudf ([#17085](https://github.com/rapidsai/cudf/pull/17085)) [@Matt711](https://github.com/Matt711) -- Migrate NVText Replacing APIs to pylibcudf ([#17084](https://github.com/rapidsai/cudf/pull/17084)) [@Matt711](https://github.com/Matt711) -- Add IWYU to CI ([#17078](https://github.com/rapidsai/cudf/pull/17078)) [@vyasr](https://github.com/vyasr) -- `cudf-polars` string/numeric casting ([#17076](https://github.com/rapidsai/cudf/pull/17076)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate NVText Normalizing APIs to Pylibcudf ([#17072](https://github.com/rapidsai/cudf/pull/17072)) [@Matt711](https://github.com/Matt711) -- Migrate remaining nvtext NGrams APIs to pylibcudf ([#17070](https://github.com/rapidsai/cudf/pull/17070)) [@Matt711](https://github.com/Matt711) -- Add profilers to CUDA 12 conda devcontainers ([#17066](https://github.com/rapidsai/cudf/pull/17066)) [@vyasr](https://github.com/vyasr) -- Add conda recipe for cudf-polars ([#17037](https://github.com/rapidsai/cudf/pull/17037)) [@bdice](https://github.com/bdice) -- Implement batch construction for strings columns ([#17035](https://github.com/rapidsai/cudf/pull/17035)) [@ttnghia](https://github.com/ttnghia) -- Add device aggregators used by shared memory groupby ([#17031](https://github.com/rapidsai/cudf/pull/17031)) [@PointKernel](https://github.com/PointKernel) -- Add optional column_order in JSON reader ([#17029](https://github.com/rapidsai/cudf/pull/17029)) [@karthikeyann](https://github.com/karthikeyann) -- Migrate Min Hashing APIs to pylibcudf ([#17021](https://github.com/rapidsai/cudf/pull/17021)) [@Matt711](https://github.com/Matt711) -- Reorganize `cudf_polars` expression code ([#17014](https://github.com/rapidsai/cudf/pull/17014)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate nvtext jaccard API to pylibcudf ([#17007](https://github.com/rapidsai/cudf/pull/17007)) [@Matt711](https://github.com/Matt711) -- Migrate nvtext generate_ngrams APIs to pylibcudf ([#17006](https://github.com/rapidsai/cudf/pull/17006)) [@Matt711](https://github.com/Matt711) -- Control whether a file data source memory-maps the file with an environment variable ([#17004](https://github.com/rapidsai/cudf/pull/17004)) [@vuule](https://github.com/vuule) -- Switched BINARY_OP Benchmarks from GoogleBench to NVBench ([#16963](https://github.com/rapidsai/cudf/pull/16963)) [@lamarrr](https://github.com/lamarrr) -- [FEA] Report all unsupported operations for a query in cudf.polars ([#16960](https://github.com/rapidsai/cudf/pull/16960)) [@Matt711](https://github.com/Matt711) -- [FEA] Migrate nvtext/edit_distance APIs to pylibcudf ([#16957](https://github.com/rapidsai/cudf/pull/16957)) [@Matt711](https://github.com/Matt711) -- Switched AST benchmarks from GoogleBench to NVBench ([#16952](https://github.com/rapidsai/cudf/pull/16952)) [@lamarrr](https://github.com/lamarrr) -- Extend `device_scalar` to optionally use pinned bounce buffer ([#16947](https://github.com/rapidsai/cudf/pull/16947)) [@vuule](https://github.com/vuule) -- Implement `cudf-polars` chunked parquet reading ([#16944](https://github.com/rapidsai/cudf/pull/16944)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Expose streams in public round APIs ([#16925](https://github.com/rapidsai/cudf/pull/16925)) [@Matt711](https://github.com/Matt711) -- add telemetry setup to test ([#16924](https://github.com/rapidsai/cudf/pull/16924)) [@msarahan](https://github.com/msarahan) -- Add cudf::strings::contains_multiple ([#16900](https://github.com/rapidsai/cudf/pull/16900)) [@davidwendt](https://github.com/davidwendt) -- Made cudftestutil header-only and removed GTest dependency ([#16839](https://github.com/rapidsai/cudf/pull/16839)) [@lamarrr](https://github.com/lamarrr) -- Add an example to demonstrate multithreaded `read_parquet` pipelines ([#16828](https://github.com/rapidsai/cudf/pull/16828)) [@mhaseeb123](https://github.com/mhaseeb123) -- Implement `extract_datetime_component` in `libcudf`/`pylibcudf` ([#16776](https://github.com/rapidsai/cudf/pull/16776)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add cudf::strings::find_re API ([#16742](https://github.com/rapidsai/cudf/pull/16742)) [@davidwendt](https://github.com/davidwendt) -- Migrate hashing operations to `pylibcudf` ([#15418](https://github.com/rapidsai/cudf/pull/15418)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add version config ([#17312](https://github.com/NVIDIA/cudf/pull/17312)) [@vyasr](https://github.com/vyasr) +- Java JNI for Multiple contains ([#17281](https://github.com/NVIDIA/cudf/pull/17281)) [@res-life](https://github.com/res-life) +- Add `cudf::calendrical_month_sequence` to pylibcudf ([#17277](https://github.com/NVIDIA/cudf/pull/17277)) [@Matt711](https://github.com/Matt711) +- Raise errors on specific types of fallback in `cudf.pandas` ([#17268](https://github.com/NVIDIA/cudf/pull/17268)) [@Matt711](https://github.com/Matt711) +- Add `catboost` to the third-party integration tests ([#17267](https://github.com/NVIDIA/cudf/pull/17267)) [@Matt711](https://github.com/Matt711) +- Add type stubs for pylibcudf ([#17258](https://github.com/NVIDIA/cudf/pull/17258)) [@wence-](https://github.com/wence-) +- Use pylibcudf contiguous split APIs in cudf python ([#17246](https://github.com/NVIDIA/cudf/pull/17246)) [@Matt711](https://github.com/Matt711) +- Upgrade nvcomp to 4.1.0.6 ([#17201](https://github.com/NVIDIA/cudf/pull/17201)) [@bdice](https://github.com/bdice) +- Added Arrow Interop Benchmarks ([#17194](https://github.com/NVIDIA/cudf/pull/17194)) [@lamarrr](https://github.com/lamarrr) +- Rewrite Java API `Table.readJSON` to return the output from libcudf `read_json` directly ([#17180](https://github.com/NVIDIA/cudf/pull/17180)) [@ttnghia](https://github.com/ttnghia) +- Support storing `precision` of decimal types in `Schema` class ([#17176](https://github.com/NVIDIA/cudf/pull/17176)) [@ttnghia](https://github.com/ttnghia) +- Migrate CSV writer to pylibcudf ([#17163](https://github.com/NVIDIA/cudf/pull/17163)) [@Matt711](https://github.com/Matt711) +- Add compute_shared_memory_aggs used by shared memory groupby ([#17162](https://github.com/NVIDIA/cudf/pull/17162)) [@PointKernel](https://github.com/PointKernel) +- Added ast tree to simplify expression lifetime management ([#17156](https://github.com/NVIDIA/cudf/pull/17156)) [@lamarrr](https://github.com/lamarrr) +- Add compute_mapping_indices used by shared memory groupby ([#17147](https://github.com/NVIDIA/cudf/pull/17147)) [@PointKernel](https://github.com/PointKernel) +- Add remaining datetime APIs to pylibcudf ([#17143](https://github.com/NVIDIA/cudf/pull/17143)) [@Matt711](https://github.com/Matt711) +- Added strings AST vs BINARY_OP benchmarks ([#17128](https://github.com/NVIDIA/cudf/pull/17128)) [@lamarrr](https://github.com/lamarrr) +- Use `libcudf_exception_handler` throughout `pylibcudf.libcudf` ([#17109](https://github.com/NVIDIA/cudf/pull/17109)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Include timezone file path in error message ([#17102](https://github.com/NVIDIA/cudf/pull/17102)) [@bdice](https://github.com/bdice) +- Migrate NVText Byte Pair Encoding APIs to pylibcudf ([#17101](https://github.com/NVIDIA/cudf/pull/17101)) [@Matt711](https://github.com/Matt711) +- Migrate NVText Tokenizing APIs to pylibcudf ([#17100](https://github.com/NVIDIA/cudf/pull/17100)) [@Matt711](https://github.com/Matt711) +- Migrate NVtext subword tokenizing APIs to pylibcudf ([#17096](https://github.com/NVIDIA/cudf/pull/17096)) [@Matt711](https://github.com/Matt711) +- Migrate NVText Stemming APIs to pylibcudf ([#17085](https://github.com/NVIDIA/cudf/pull/17085)) [@Matt711](https://github.com/Matt711) +- Migrate NVText Replacing APIs to pylibcudf ([#17084](https://github.com/NVIDIA/cudf/pull/17084)) [@Matt711](https://github.com/Matt711) +- Add IWYU to CI ([#17078](https://github.com/NVIDIA/cudf/pull/17078)) [@vyasr](https://github.com/vyasr) +- `cudf-polars` string/numeric casting ([#17076](https://github.com/NVIDIA/cudf/pull/17076)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate NVText Normalizing APIs to Pylibcudf ([#17072](https://github.com/NVIDIA/cudf/pull/17072)) [@Matt711](https://github.com/Matt711) +- Migrate remaining nvtext NGrams APIs to pylibcudf ([#17070](https://github.com/NVIDIA/cudf/pull/17070)) [@Matt711](https://github.com/Matt711) +- Add profilers to CUDA 12 conda devcontainers ([#17066](https://github.com/NVIDIA/cudf/pull/17066)) [@vyasr](https://github.com/vyasr) +- Add conda recipe for cudf-polars ([#17037](https://github.com/NVIDIA/cudf/pull/17037)) [@bdice](https://github.com/bdice) +- Implement batch construction for strings columns ([#17035](https://github.com/NVIDIA/cudf/pull/17035)) [@ttnghia](https://github.com/ttnghia) +- Add device aggregators used by shared memory groupby ([#17031](https://github.com/NVIDIA/cudf/pull/17031)) [@PointKernel](https://github.com/PointKernel) +- Add optional column_order in JSON reader ([#17029](https://github.com/NVIDIA/cudf/pull/17029)) [@karthikeyann](https://github.com/karthikeyann) +- Migrate Min Hashing APIs to pylibcudf ([#17021](https://github.com/NVIDIA/cudf/pull/17021)) [@Matt711](https://github.com/Matt711) +- Reorganize `cudf_polars` expression code ([#17014](https://github.com/NVIDIA/cudf/pull/17014)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate nvtext jaccard API to pylibcudf ([#17007](https://github.com/NVIDIA/cudf/pull/17007)) [@Matt711](https://github.com/Matt711) +- Migrate nvtext generate_ngrams APIs to pylibcudf ([#17006](https://github.com/NVIDIA/cudf/pull/17006)) [@Matt711](https://github.com/Matt711) +- Control whether a file data source memory-maps the file with an environment variable ([#17004](https://github.com/NVIDIA/cudf/pull/17004)) [@vuule](https://github.com/vuule) +- Switched BINARY_OP Benchmarks from GoogleBench to NVBench ([#16963](https://github.com/NVIDIA/cudf/pull/16963)) [@lamarrr](https://github.com/lamarrr) +- [FEA] Report all unsupported operations for a query in cudf.polars ([#16960](https://github.com/NVIDIA/cudf/pull/16960)) [@Matt711](https://github.com/Matt711) +- [FEA] Migrate nvtext/edit_distance APIs to pylibcudf ([#16957](https://github.com/NVIDIA/cudf/pull/16957)) [@Matt711](https://github.com/Matt711) +- Switched AST benchmarks from GoogleBench to NVBench ([#16952](https://github.com/NVIDIA/cudf/pull/16952)) [@lamarrr](https://github.com/lamarrr) +- Extend `device_scalar` to optionally use pinned bounce buffer ([#16947](https://github.com/NVIDIA/cudf/pull/16947)) [@vuule](https://github.com/vuule) +- Implement `cudf-polars` chunked parquet reading ([#16944](https://github.com/NVIDIA/cudf/pull/16944)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Expose streams in public round APIs ([#16925](https://github.com/NVIDIA/cudf/pull/16925)) [@Matt711](https://github.com/Matt711) +- add telemetry setup to test ([#16924](https://github.com/NVIDIA/cudf/pull/16924)) [@msarahan](https://github.com/msarahan) +- Add cudf::strings::contains_multiple ([#16900](https://github.com/NVIDIA/cudf/pull/16900)) [@davidwendt](https://github.com/davidwendt) +- Made cudftestutil header-only and removed GTest dependency ([#16839](https://github.com/NVIDIA/cudf/pull/16839)) [@lamarrr](https://github.com/lamarrr) +- Add an example to demonstrate multithreaded `read_parquet` pipelines ([#16828](https://github.com/NVIDIA/cudf/pull/16828)) [@mhaseeb123](https://github.com/mhaseeb123) +- Implement `extract_datetime_component` in `libcudf`/`pylibcudf` ([#16776](https://github.com/NVIDIA/cudf/pull/16776)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add cudf::strings::find_re API ([#16742](https://github.com/NVIDIA/cudf/pull/16742)) [@davidwendt](https://github.com/davidwendt) +- Migrate hashing operations to `pylibcudf` ([#15418](https://github.com/NVIDIA/cudf/pull/15418)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🛠️ Improvements -- Simplify serialization protocols ([#17552](https://github.com/rapidsai/cudf/pull/17552)) [@vyasr](https://github.com/vyasr) -- Add `pynvml` as a dependency for `dask-cudf` ([#17386](https://github.com/rapidsai/cudf/pull/17386)) [@pentschev](https://github.com/pentschev) -- Enable unified memory by default in `cudf_polars` ([#17375](https://github.com/rapidsai/cudf/pull/17375)) [@galipremsagar](https://github.com/galipremsagar) -- Support polars 1.14 ([#17355](https://github.com/rapidsai/cudf/pull/17355)) [@wence-](https://github.com/wence-) -- Remove cudf._lib.quantiles in favor of inlining pylibcudf ([#17347](https://github.com/rapidsai/cudf/pull/17347)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.labeling in favor of inlining pylibcudf ([#17346](https://github.com/rapidsai/cudf/pull/17346)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.hash in favor of inlining pylibcudf ([#17345](https://github.com/rapidsai/cudf/pull/17345)) [@mroeschke](https://github.com/mroeschke) -- Remove cudf._lib.concat in favor of inlining pylibcudf ([#17344](https://github.com/rapidsai/cudf/pull/17344)) [@mroeschke](https://github.com/mroeschke) -- Extract ``GPUEngine`` config options at translation time ([#17339](https://github.com/rapidsai/cudf/pull/17339)) [@rjzamora](https://github.com/rjzamora) -- Update java datetime APIs to match CUDF. ([#17329](https://github.com/rapidsai/cudf/pull/17329)) [@revans2](https://github.com/revans2) -- Move strings url_decode benchmarks to nvbench ([#17328](https://github.com/rapidsai/cudf/pull/17328)) [@davidwendt](https://github.com/davidwendt) -- Move strings translate benchmarks to nvbench ([#17325](https://github.com/rapidsai/cudf/pull/17325)) [@davidwendt](https://github.com/davidwendt) -- Writing compressed output using JSON writer ([#17323](https://github.com/rapidsai/cudf/pull/17323)) [@shrshi](https://github.com/shrshi) -- Test the full matrix for polars and dask wheels on nightlies ([#17320](https://github.com/rapidsai/cudf/pull/17320)) [@vyasr](https://github.com/vyasr) -- Remove cudf._lib.avro in favor of inlining pylicudf ([#17319](https://github.com/rapidsai/cudf/pull/17319)) [@mroeschke](https://github.com/mroeschke) -- Move cudf._lib.unary to cudf.core._internals ([#17318](https://github.com/rapidsai/cudf/pull/17318)) [@mroeschke](https://github.com/mroeschke) -- prefer wheel-provided libcudf.so in load_library(), use RTLD_LOCAL ([#17316](https://github.com/rapidsai/cudf/pull/17316)) [@jameslamb](https://github.com/jameslamb) -- Clean up misc, unneeded pylibcudf.libcudf in cudf._lib ([#17309](https://github.com/rapidsai/cudf/pull/17309)) [@mroeschke](https://github.com/mroeschke) -- Exclude nanoarrow and flatbuffers from installation ([#17308](https://github.com/rapidsai/cudf/pull/17308)) [@vyasr](https://github.com/vyasr) -- Update CI jobs to include Polars in nightlies and improve IWYU ([#17306](https://github.com/rapidsai/cudf/pull/17306)) [@vyasr](https://github.com/vyasr) -- Move strings repeat benchmarks to nvbench ([#17304](https://github.com/rapidsai/cudf/pull/17304)) [@davidwendt](https://github.com/davidwendt) -- Fix synchronization bug in bool parquet mukernels ([#17302](https://github.com/rapidsai/cudf/pull/17302)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Move strings replace benchmarks to nvbench ([#17301](https://github.com/rapidsai/cudf/pull/17301)) [@davidwendt](https://github.com/davidwendt) -- Support polars 1.13 ([#17299](https://github.com/rapidsai/cudf/pull/17299)) [@wence-](https://github.com/wence-) -- Replace FindcuFile with upstream FindCUDAToolkit support ([#17298](https://github.com/rapidsai/cudf/pull/17298)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Expose stream-ordering in public transpose API ([#17294](https://github.com/rapidsai/cudf/pull/17294)) [@shrshi](https://github.com/shrshi) -- Replace workaround of JNI build with CUDF_KVIKIO_REMOTE_IO=OFF ([#17293](https://github.com/rapidsai/cudf/pull/17293)) [@pxLi](https://github.com/pxLi) -- cmake option: `CUDF_KVIKIO_REMOTE_IO` ([#17291](https://github.com/rapidsai/cudf/pull/17291)) [@madsbk](https://github.com/madsbk) -- Use more pylibcudf Python enums in cudf._lib ([#17288](https://github.com/rapidsai/cudf/pull/17288)) [@mroeschke](https://github.com/mroeschke) -- Use pylibcudf enums in cudf Python quantile ([#17287](https://github.com/rapidsai/cudf/pull/17287)) [@mroeschke](https://github.com/mroeschke) -- enforce wheel size limits, README formatting in CI ([#17284](https://github.com/rapidsai/cudf/pull/17284)) [@jameslamb](https://github.com/jameslamb) -- Use numba-cuda<0.0.18 ([#17280](https://github.com/rapidsai/cudf/pull/17280)) [@gmarkall](https://github.com/gmarkall) -- Add compute_column_expression to pylibcudf for transform.compute_column ([#17279](https://github.com/rapidsai/cudf/pull/17279)) [@mroeschke](https://github.com/mroeschke) -- Optimize distinct inner join to use set `find` instead of `retrieve` ([#17278](https://github.com/rapidsai/cudf/pull/17278)) [@PointKernel](https://github.com/PointKernel) -- remove WheelHelpers.cmake ([#17276](https://github.com/rapidsai/cudf/pull/17276)) [@jameslamb](https://github.com/jameslamb) -- Plumb pylibcudf datetime APIs through cudf python ([#17275](https://github.com/rapidsai/cudf/pull/17275)) [@Matt711](https://github.com/Matt711) -- Follow up making Python tests more deterministic ([#17272](https://github.com/rapidsai/cudf/pull/17272)) [@mroeschke](https://github.com/mroeschke) -- Use pylibcudf.search APIs in cudf python ([#17271](https://github.com/rapidsai/cudf/pull/17271)) [@Matt711](https://github.com/Matt711) -- Use `pylibcudf.strings.convert.convert_integers.is_integer` in cudf python ([#17270](https://github.com/rapidsai/cudf/pull/17270)) [@Matt711](https://github.com/Matt711) -- Move strings filter benchmarks to nvbench ([#17269](https://github.com/rapidsai/cudf/pull/17269)) [@davidwendt](https://github.com/davidwendt) -- Make constructor of DeviceMemoryBufferView public ([#17265](https://github.com/rapidsai/cudf/pull/17265)) [@liurenjie1024](https://github.com/liurenjie1024) -- Put a ceiling on cuda-python ([#17264](https://github.com/rapidsai/cudf/pull/17264)) [@jameslamb](https://github.com/jameslamb) -- Always prefer `device_read`s and `device_write`s when kvikIO is enabled ([#17260](https://github.com/rapidsai/cudf/pull/17260)) [@vuule](https://github.com/vuule) -- Expose streams in public quantile APIs ([#17257](https://github.com/rapidsai/cudf/pull/17257)) [@shrshi](https://github.com/shrshi) -- Add support for `pyarrow-18` ([#17256](https://github.com/rapidsai/cudf/pull/17256)) [@galipremsagar](https://github.com/galipremsagar) -- Move strings/numeric convert benchmarks to nvbench ([#17255](https://github.com/rapidsai/cudf/pull/17255)) [@davidwendt](https://github.com/davidwendt) -- Add new ``dask_cudf.read_parquet`` API ([#17250](https://github.com/rapidsai/cudf/pull/17250)) [@rjzamora](https://github.com/rjzamora) -- Add read_parquet_metadata to pylibcudf ([#17245](https://github.com/rapidsai/cudf/pull/17245)) [@mroeschke](https://github.com/mroeschke) -- Search for kvikio with lowercase ([#17243](https://github.com/rapidsai/cudf/pull/17243)) [@vyasr](https://github.com/vyasr) -- KvikIO shared library ([#17239](https://github.com/rapidsai/cudf/pull/17239)) [@madsbk](https://github.com/madsbk) -- Use more pylibcudf.io.types enums in cudf._libs ([#17237](https://github.com/rapidsai/cudf/pull/17237)) [@mroeschke](https://github.com/mroeschke) -- Expose mixed and conditional joins in pylibcudf ([#17235](https://github.com/rapidsai/cudf/pull/17235)) [@wence-](https://github.com/wence-) -- Add io.text APIs to pylibcudf ([#17232](https://github.com/rapidsai/cudf/pull/17232)) [@mroeschke](https://github.com/mroeschke) -- Add `num_iterations` axis to the multi-threaded Parquet benchmarks ([#17231](https://github.com/rapidsai/cudf/pull/17231)) [@vuule](https://github.com/vuule) -- Move strings to date/time types benchmarks to nvbench ([#17229](https://github.com/rapidsai/cudf/pull/17229)) [@davidwendt](https://github.com/davidwendt) -- Support for polars 1.12 in cudf-polars ([#17227](https://github.com/rapidsai/cudf/pull/17227)) [@wence-](https://github.com/wence-) -- Allow generating large strings in benchmarks ([#17224](https://github.com/rapidsai/cudf/pull/17224)) [@davidwendt](https://github.com/davidwendt) -- Refactor gather/scatter benchmarks for strings ([#17223](https://github.com/rapidsai/cudf/pull/17223)) [@davidwendt](https://github.com/davidwendt) -- Deprecate single component extraction methods in libcudf ([#17221](https://github.com/rapidsai/cudf/pull/17221)) [@Matt711](https://github.com/Matt711) -- Remove `nvtext::load_vocabulary` from pylibcudf ([#17220](https://github.com/rapidsai/cudf/pull/17220)) [@Matt711](https://github.com/Matt711) -- Benchmarking JSON reader for compressed inputs ([#17219](https://github.com/rapidsai/cudf/pull/17219)) [@shrshi](https://github.com/shrshi) -- Expose stream-ordering in partitioning API ([#17213](https://github.com/rapidsai/cudf/pull/17213)) [@shrshi](https://github.com/shrshi) -- Move strings::concatenate benchmark to nvbench ([#17211](https://github.com/rapidsai/cudf/pull/17211)) [@davidwendt](https://github.com/davidwendt) -- Expose stream-ordering in subword tokenizer API ([#17206](https://github.com/rapidsai/cudf/pull/17206)) [@shrshi](https://github.com/shrshi) -- Refactor Dask cuDF legacy code ([#17205](https://github.com/rapidsai/cudf/pull/17205)) [@rjzamora](https://github.com/rjzamora) -- Make HostMemoryBuffer call into the DefaultHostMemoryAllocator ([#17204](https://github.com/rapidsai/cudf/pull/17204)) [@revans2](https://github.com/revans2) -- Unified binary_ops and ast benchmarks parameter names ([#17200](https://github.com/rapidsai/cudf/pull/17200)) [@lamarrr](https://github.com/lamarrr) -- Add in new java API for raw host memory allocation ([#17197](https://github.com/rapidsai/cudf/pull/17197)) [@revans2](https://github.com/revans2) -- Remove java reservation ([#17189](https://github.com/rapidsai/cudf/pull/17189)) [@revans2](https://github.com/revans2) -- Fixed unused attribute compilation error for GCC 13 ([#17188](https://github.com/rapidsai/cudf/pull/17188)) [@lamarrr](https://github.com/lamarrr) -- Change default KvikIO parameters in cuDF: set the thread pool size to 4, and compatibility mode to ON ([#17185](https://github.com/rapidsai/cudf/pull/17185)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Use make_device_uvector instead of cudaMemcpyAsync in inplace_bitmask_binop ([#17181](https://github.com/rapidsai/cudf/pull/17181)) [@davidwendt](https://github.com/davidwendt) -- Make ai.rapids.cudf.HostMemoryBuffer#copyFromStream public. ([#17179](https://github.com/rapidsai/cudf/pull/17179)) [@liurenjie1024](https://github.com/liurenjie1024) -- Separate evaluation logic from `IR` objects in cudf-polars ([#17175](https://github.com/rapidsai/cudf/pull/17175)) [@rjzamora](https://github.com/rjzamora) -- Move nvtext ngrams benchmarks to nvbench ([#17173](https://github.com/rapidsai/cudf/pull/17173)) [@davidwendt](https://github.com/davidwendt) -- Remove includes suggested by include-what-you-use ([#17170](https://github.com/rapidsai/cudf/pull/17170)) [@vyasr](https://github.com/vyasr) -- Reading multi-source compressed JSONL files ([#17161](https://github.com/rapidsai/cudf/pull/17161)) [@shrshi](https://github.com/shrshi) -- Process parquet bools with microkernels ([#17157](https://github.com/rapidsai/cudf/pull/17157)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Upgrade to polars 1.11 in cudf-polars ([#17154](https://github.com/rapidsai/cudf/pull/17154)) [@wence-](https://github.com/wence-) -- Deprecate current libcudf nvtext minhash functions ([#17152](https://github.com/rapidsai/cudf/pull/17152)) [@davidwendt](https://github.com/davidwendt) -- Remove unused variable in internal merge_tdigests utility ([#17151](https://github.com/rapidsai/cudf/pull/17151)) [@davidwendt](https://github.com/davidwendt) -- Use the full ref name of `rmm.DeviceBuffer` in the sphinx config file ([#17150](https://github.com/rapidsai/cudf/pull/17150)) [@Matt711](https://github.com/Matt711) -- Move `segmented_gather` function from the copying module to the lists module ([#17148](https://github.com/rapidsai/cudf/pull/17148)) [@Matt711](https://github.com/Matt711) -- Use async execution policy for true_if ([#17146](https://github.com/rapidsai/cudf/pull/17146)) [@PointKernel](https://github.com/PointKernel) -- Add conversion from cudf-polars expressions to libcudf ast for parquet filters ([#17141](https://github.com/rapidsai/cudf/pull/17141)) [@wence-](https://github.com/wence-) -- devcontainer: replace `VAULT_HOST` with `AWS_ROLE_ARN` ([#17134](https://github.com/rapidsai/cudf/pull/17134)) [@jjacobelli](https://github.com/jjacobelli) -- Replace direct `cudaMemcpyAsync` calls with utility functions (limited to `cudf::io`) ([#17132](https://github.com/rapidsai/cudf/pull/17132)) [@vuule](https://github.com/vuule) -- use rapids-generate-pip-constraints to pin to oldest dependencies in CI ([#17131](https://github.com/rapidsai/cudf/pull/17131)) [@jameslamb](https://github.com/jameslamb) -- Set the default number of threads in KvikIO thread pool to 8 ([#17126](https://github.com/rapidsai/cudf/pull/17126)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Fix clang-tidy violations for span.hpp and hostdevice_vector.hpp ([#17124](https://github.com/rapidsai/cudf/pull/17124)) [@davidwendt](https://github.com/davidwendt) -- Disable the Parquet reader's wide lists tables GTest by default ([#17120](https://github.com/rapidsai/cudf/pull/17120)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add compile time check to ensure the `counting_iterator` type in `counting_transform_iterator` fits in `size_type` ([#17118](https://github.com/rapidsai/cudf/pull/17118)) [@mhaseeb123](https://github.com/mhaseeb123) -- Minor I/O code quality improvements ([#17105](https://github.com/rapidsai/cudf/pull/17105)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Remove the additional host register calls initially intended for performance improvement on Grace Hopper ([#17092](https://github.com/rapidsai/cudf/pull/17092)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Split hash-based groupby into multiple smaller files to reduce build time ([#17089](https://github.com/rapidsai/cudf/pull/17089)) [@PointKernel](https://github.com/PointKernel) -- build wheels without build isolation ([#17088](https://github.com/rapidsai/cudf/pull/17088)) [@jameslamb](https://github.com/jameslamb) -- Polars: DataFrame Serialization ([#17062](https://github.com/rapidsai/cudf/pull/17062)) [@madsbk](https://github.com/madsbk) -- Remove unused hash helper functions ([#17056](https://github.com/rapidsai/cudf/pull/17056)) [@PointKernel](https://github.com/PointKernel) -- Add to_dlpack/from_dlpack APIs to pylibcudf ([#17055](https://github.com/rapidsai/cudf/pull/17055)) [@mroeschke](https://github.com/mroeschke) -- Move `flatten_single_pass_aggs` to its own TU ([#17053](https://github.com/rapidsai/cudf/pull/17053)) [@PointKernel](https://github.com/PointKernel) -- Replace deprecated cuco APIs with updated versions ([#17052](https://github.com/rapidsai/cudf/pull/17052)) [@PointKernel](https://github.com/PointKernel) -- Refactor ORC dictionary encoding to migrate to the new `cuco::static_map` ([#17049](https://github.com/rapidsai/cudf/pull/17049)) [@mhaseeb123](https://github.com/mhaseeb123) -- Move pylibcudf/libcudf/wrappers/decimals to pylibcudf/libcudf/fixed_point ([#17048](https://github.com/rapidsai/cudf/pull/17048)) [@mroeschke](https://github.com/mroeschke) -- make conda installs in CI stricter (part 2) ([#17042](https://github.com/rapidsai/cudf/pull/17042)) [@jameslamb](https://github.com/jameslamb) -- Use managed memory for NDSH benchmarks ([#17039](https://github.com/rapidsai/cudf/pull/17039)) [@karthikeyann](https://github.com/karthikeyann) -- Clean up hash-groupby `var_hash_functor` ([#17034](https://github.com/rapidsai/cudf/pull/17034)) [@PointKernel](https://github.com/PointKernel) -- Add json APIs to pylibcudf ([#17025](https://github.com/rapidsai/cudf/pull/17025)) [@mroeschke](https://github.com/mroeschke) -- Add string.replace_re APIs to pylibcudf ([#17023](https://github.com/rapidsai/cudf/pull/17023)) [@mroeschke](https://github.com/mroeschke) -- Replace old host tree algorithm with new algorithm in JSON reader ([#17019](https://github.com/rapidsai/cudf/pull/17019)) [@karthikeyann](https://github.com/karthikeyann) -- Unify treatment of `Expr` and `IR` nodes in cudf-polars DSL ([#17016](https://github.com/rapidsai/cudf/pull/17016)) [@wence-](https://github.com/wence-) -- make conda installs in CI stricter ([#17013](https://github.com/rapidsai/cudf/pull/17013)) [@jameslamb](https://github.com/jameslamb) -- Pylibcudf: pack and unpack ([#17012](https://github.com/rapidsai/cudf/pull/17012)) [@madsbk](https://github.com/madsbk) -- Remove unneeded pylibcudf.libcudf.wrappers.duration usage in cudf ([#17010](https://github.com/rapidsai/cudf/pull/17010)) [@mroeschke](https://github.com/mroeschke) -- Add custom "fused" groupby aggregation to Dask cuDF ([#17009](https://github.com/rapidsai/cudf/pull/17009)) [@rjzamora](https://github.com/rjzamora) -- Make tests more deterministic ([#17008](https://github.com/rapidsai/cudf/pull/17008)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unused import ([#17005](https://github.com/rapidsai/cudf/pull/17005)) [@Matt711](https://github.com/Matt711) -- Add string.convert.convert_urls APIs to pylibcudf ([#17003](https://github.com/rapidsai/cudf/pull/17003)) [@mroeschke](https://github.com/mroeschke) -- Add release tracking to project automation scripts ([#17001](https://github.com/rapidsai/cudf/pull/17001)) [@jarmak-nv](https://github.com/jarmak-nv) -- Implement inequality joins by translation to conditional joins ([#17000](https://github.com/rapidsai/cudf/pull/17000)) [@wence-](https://github.com/wence-) -- Add string.convert.convert_lists APIs to pylibcudf ([#16997](https://github.com/rapidsai/cudf/pull/16997)) [@mroeschke](https://github.com/mroeschke) -- Performance optimization of JSON validation ([#16996](https://github.com/rapidsai/cudf/pull/16996)) [@karthikeyann](https://github.com/karthikeyann) -- Add string.convert.convert_ipv4 APIs to pylibcudf ([#16994](https://github.com/rapidsai/cudf/pull/16994)) [@mroeschke](https://github.com/mroeschke) -- Add string.convert.convert_integers APIs to pylibcudf ([#16991](https://github.com/rapidsai/cudf/pull/16991)) [@mroeschke](https://github.com/mroeschke) -- Add string.convert_floats APIs to pylibcudf ([#16990](https://github.com/rapidsai/cudf/pull/16990)) [@mroeschke](https://github.com/mroeschke) -- Add string.convert.convert_fixed_type APIs to pylibcudf ([#16984](https://github.com/rapidsai/cudf/pull/16984)) [@mroeschke](https://github.com/mroeschke) -- Remove unnecessary `std::move`'s in pylibcudf ([#16983](https://github.com/rapidsai/cudf/pull/16983)) [@Matt711](https://github.com/Matt711) -- Add docstrings and test for strings.convert_durations APIs for pylibcudf ([#16982](https://github.com/rapidsai/cudf/pull/16982)) [@mroeschke](https://github.com/mroeschke) -- JSON tokenizer memory optimizations ([#16978](https://github.com/rapidsai/cudf/pull/16978)) [@shrshi](https://github.com/shrshi) -- Turn on `xfail_strict = true` for all python packages ([#16977](https://github.com/rapidsai/cudf/pull/16977)) [@wence-](https://github.com/wence-) -- Add string.convert.convert_datetime/convert_booleans APIs to pylibcudf ([#16971](https://github.com/rapidsai/cudf/pull/16971)) [@mroeschke](https://github.com/mroeschke) -- Auto assign PR to author ([#16969](https://github.com/rapidsai/cudf/pull/16969)) [@Matt711](https://github.com/Matt711) -- Deprecate support for directly accessing logger ([#16964](https://github.com/rapidsai/cudf/pull/16964)) [@vyasr](https://github.com/vyasr) -- Expunge NamedColumn ([#16962](https://github.com/rapidsai/cudf/pull/16962)) [@wence-](https://github.com/wence-) -- Add clang-tidy to CI ([#16958](https://github.com/rapidsai/cudf/pull/16958)) [@vyasr](https://github.com/vyasr) -- Address all remaining clang-tidy errors ([#16956](https://github.com/rapidsai/cudf/pull/16956)) [@vyasr](https://github.com/vyasr) -- Apply clang-tidy autofixes ([#16949](https://github.com/rapidsai/cudf/pull/16949)) [@vyasr](https://github.com/vyasr) -- Use nvcomp wheel instead of bundling nvcomp ([#16946](https://github.com/rapidsai/cudf/pull/16946)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Refactor the `cuda_memcpy` functions to make them more usable ([#16945](https://github.com/rapidsai/cudf/pull/16945)) [@vuule](https://github.com/vuule) -- Add string.split APIs to pylibcudf ([#16940](https://github.com/rapidsai/cudf/pull/16940)) [@mroeschke](https://github.com/mroeschke) -- clang-tidy fixes part 3 ([#16939](https://github.com/rapidsai/cudf/pull/16939)) [@vyasr](https://github.com/vyasr) -- clang-tidy fixes part 2 ([#16938](https://github.com/rapidsai/cudf/pull/16938)) [@vyasr](https://github.com/vyasr) -- clang-tidy fixes part 1 ([#16937](https://github.com/rapidsai/cudf/pull/16937)) [@vyasr](https://github.com/vyasr) -- Add string.wrap APIs to pylibcudf ([#16935](https://github.com/rapidsai/cudf/pull/16935)) [@mroeschke](https://github.com/mroeschke) -- Add string.translate APIs to pylibcudf ([#16934](https://github.com/rapidsai/cudf/pull/16934)) [@mroeschke](https://github.com/mroeschke) -- Add string.find_multiple APIs to pylibcudf ([#16920](https://github.com/rapidsai/cudf/pull/16920)) [@mroeschke](https://github.com/mroeschke) -- Batch memcpy the last offsets for output buffers of str and list cols in PQ reader ([#16905](https://github.com/rapidsai/cudf/pull/16905)) [@mhaseeb123](https://github.com/mhaseeb123) -- reduce wheel build verbosity, narrow deprecation warning filter ([#16896](https://github.com/rapidsai/cudf/pull/16896)) [@jameslamb](https://github.com/jameslamb) -- Improve aggregation device functors ([#16884](https://github.com/rapidsai/cudf/pull/16884)) [@PointKernel](https://github.com/PointKernel) -- Upgrade pandas pinnings to support `2.2.3` ([#16882](https://github.com/rapidsai/cudf/pull/16882)) [@galipremsagar](https://github.com/galipremsagar) -- Fix 24.10 to 24.12 forward merge ([#16876](https://github.com/rapidsai/cudf/pull/16876)) [@bdice](https://github.com/bdice) -- Manually resolve conflicts in between branch-24.12 and branch-24.10 ([#16871](https://github.com/rapidsai/cudf/pull/16871)) [@galipremsagar](https://github.com/galipremsagar) -- Add in support for setting delim when parsing JSON through java ([#16867](https://github.com/rapidsai/cudf/pull/16867)) [@revans2](https://github.com/revans2) -- Reapply `mixed_semi_join` refactoring and bug fixes ([#16859](https://github.com/rapidsai/cudf/pull/16859)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add string padding and side_type APIs to pylibcudf ([#16833](https://github.com/rapidsai/cudf/pull/16833)) [@mroeschke](https://github.com/mroeschke) -- Organize parquet reader mukernel non-nullable code, introduce manual block scans ([#16830](https://github.com/rapidsai/cudf/pull/16830)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Remove superfluous use of std::vector for std::future ([#16829](https://github.com/rapidsai/cudf/pull/16829)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Rework `read_csv` IO to avoid reading whole input with a single `host_read` ([#16826](https://github.com/rapidsai/cudf/pull/16826)) [@vuule](https://github.com/vuule) -- Add strings.combine APIs to pylibcudf ([#16790](https://github.com/rapidsai/cudf/pull/16790)) [@mroeschke](https://github.com/mroeschke) -- Add remaining string.char_types APIs to pylibcudf ([#16788](https://github.com/rapidsai/cudf/pull/16788)) [@mroeschke](https://github.com/mroeschke) -- Add new nvtext minhash_permuted API ([#16756](https://github.com/rapidsai/cudf/pull/16756)) [@davidwendt](https://github.com/davidwendt) -- Avoid public constructors when called with columns to avoid unnecessary validation ([#16747](https://github.com/rapidsai/cudf/pull/16747)) [@mroeschke](https://github.com/mroeschke) -- Use `changed-files` shared workflow ([#16713](https://github.com/rapidsai/cudf/pull/16713)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- lint: replace `isort` with Ruff's rule I ([#16685](https://github.com/rapidsai/cudf/pull/16685)) [@Borda](https://github.com/Borda) -- Improve the performance of low cardinality groupby ([#16619](https://github.com/rapidsai/cudf/pull/16619)) [@PointKernel](https://github.com/PointKernel) -- Parquet reader list microkernel ([#16538](https://github.com/rapidsai/cudf/pull/16538)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- AWS S3 IO through KvikIO ([#16499](https://github.com/rapidsai/cudf/pull/16499)) [@madsbk](https://github.com/madsbk) -- Refactor `histogram` reduction using `cuco::static_set::insert_and_find` ([#16485](https://github.com/rapidsai/cudf/pull/16485)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Use numba-cuda>=0.0.13 ([#16474](https://github.com/rapidsai/cudf/pull/16474)) [@gmarkall](https://github.com/gmarkall) +- Simplify serialization protocols ([#17552](https://github.com/NVIDIA/cudf/pull/17552)) [@vyasr](https://github.com/vyasr) +- Add `pynvml` as a dependency for `dask-cudf` ([#17386](https://github.com/NVIDIA/cudf/pull/17386)) [@pentschev](https://github.com/pentschev) +- Enable unified memory by default in `cudf_polars` ([#17375](https://github.com/NVIDIA/cudf/pull/17375)) [@galipremsagar](https://github.com/galipremsagar) +- Support polars 1.14 ([#17355](https://github.com/NVIDIA/cudf/pull/17355)) [@wence-](https://github.com/wence-) +- Remove cudf._lib.quantiles in favor of inlining pylibcudf ([#17347](https://github.com/NVIDIA/cudf/pull/17347)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.labeling in favor of inlining pylibcudf ([#17346](https://github.com/NVIDIA/cudf/pull/17346)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.hash in favor of inlining pylibcudf ([#17345](https://github.com/NVIDIA/cudf/pull/17345)) [@mroeschke](https://github.com/mroeschke) +- Remove cudf._lib.concat in favor of inlining pylibcudf ([#17344](https://github.com/NVIDIA/cudf/pull/17344)) [@mroeschke](https://github.com/mroeschke) +- Extract ``GPUEngine`` config options at translation time ([#17339](https://github.com/NVIDIA/cudf/pull/17339)) [@rjzamora](https://github.com/rjzamora) +- Update java datetime APIs to match CUDF. ([#17329](https://github.com/NVIDIA/cudf/pull/17329)) [@revans2](https://github.com/revans2) +- Move strings url_decode benchmarks to nvbench ([#17328](https://github.com/NVIDIA/cudf/pull/17328)) [@davidwendt](https://github.com/davidwendt) +- Move strings translate benchmarks to nvbench ([#17325](https://github.com/NVIDIA/cudf/pull/17325)) [@davidwendt](https://github.com/davidwendt) +- Writing compressed output using JSON writer ([#17323](https://github.com/NVIDIA/cudf/pull/17323)) [@shrshi](https://github.com/shrshi) +- Test the full matrix for polars and dask wheels on nightlies ([#17320](https://github.com/NVIDIA/cudf/pull/17320)) [@vyasr](https://github.com/vyasr) +- Remove cudf._lib.avro in favor of inlining pylicudf ([#17319](https://github.com/NVIDIA/cudf/pull/17319)) [@mroeschke](https://github.com/mroeschke) +- Move cudf._lib.unary to cudf.core._internals ([#17318](https://github.com/NVIDIA/cudf/pull/17318)) [@mroeschke](https://github.com/mroeschke) +- prefer wheel-provided libcudf.so in load_library(), use RTLD_LOCAL ([#17316](https://github.com/NVIDIA/cudf/pull/17316)) [@jameslamb](https://github.com/jameslamb) +- Clean up misc, unneeded pylibcudf.libcudf in cudf._lib ([#17309](https://github.com/NVIDIA/cudf/pull/17309)) [@mroeschke](https://github.com/mroeschke) +- Exclude nanoarrow and flatbuffers from installation ([#17308](https://github.com/NVIDIA/cudf/pull/17308)) [@vyasr](https://github.com/vyasr) +- Update CI jobs to include Polars in nightlies and improve IWYU ([#17306](https://github.com/NVIDIA/cudf/pull/17306)) [@vyasr](https://github.com/vyasr) +- Move strings repeat benchmarks to nvbench ([#17304](https://github.com/NVIDIA/cudf/pull/17304)) [@davidwendt](https://github.com/davidwendt) +- Fix synchronization bug in bool parquet mukernels ([#17302](https://github.com/NVIDIA/cudf/pull/17302)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Move strings replace benchmarks to nvbench ([#17301](https://github.com/NVIDIA/cudf/pull/17301)) [@davidwendt](https://github.com/davidwendt) +- Support polars 1.13 ([#17299](https://github.com/NVIDIA/cudf/pull/17299)) [@wence-](https://github.com/wence-) +- Replace FindcuFile with upstream FindCUDAToolkit support ([#17298](https://github.com/NVIDIA/cudf/pull/17298)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Expose stream-ordering in public transpose API ([#17294](https://github.com/NVIDIA/cudf/pull/17294)) [@shrshi](https://github.com/shrshi) +- Replace workaround of JNI build with CUDF_KVIKIO_REMOTE_IO=OFF ([#17293](https://github.com/NVIDIA/cudf/pull/17293)) [@pxLi](https://github.com/pxLi) +- cmake option: `CUDF_KVIKIO_REMOTE_IO` ([#17291](https://github.com/NVIDIA/cudf/pull/17291)) [@madsbk](https://github.com/madsbk) +- Use more pylibcudf Python enums in cudf._lib ([#17288](https://github.com/NVIDIA/cudf/pull/17288)) [@mroeschke](https://github.com/mroeschke) +- Use pylibcudf enums in cudf Python quantile ([#17287](https://github.com/NVIDIA/cudf/pull/17287)) [@mroeschke](https://github.com/mroeschke) +- enforce wheel size limits, README formatting in CI ([#17284](https://github.com/NVIDIA/cudf/pull/17284)) [@jameslamb](https://github.com/jameslamb) +- Use numba-cuda<0.0.18 ([#17280](https://github.com/NVIDIA/cudf/pull/17280)) [@gmarkall](https://github.com/gmarkall) +- Add compute_column_expression to pylibcudf for transform.compute_column ([#17279](https://github.com/NVIDIA/cudf/pull/17279)) [@mroeschke](https://github.com/mroeschke) +- Optimize distinct inner join to use set `find` instead of `retrieve` ([#17278](https://github.com/NVIDIA/cudf/pull/17278)) [@PointKernel](https://github.com/PointKernel) +- remove WheelHelpers.cmake ([#17276](https://github.com/NVIDIA/cudf/pull/17276)) [@jameslamb](https://github.com/jameslamb) +- Plumb pylibcudf datetime APIs through cudf python ([#17275](https://github.com/NVIDIA/cudf/pull/17275)) [@Matt711](https://github.com/Matt711) +- Follow up making Python tests more deterministic ([#17272](https://github.com/NVIDIA/cudf/pull/17272)) [@mroeschke](https://github.com/mroeschke) +- Use pylibcudf.search APIs in cudf python ([#17271](https://github.com/NVIDIA/cudf/pull/17271)) [@Matt711](https://github.com/Matt711) +- Use `pylibcudf.strings.convert.convert_integers.is_integer` in cudf python ([#17270](https://github.com/NVIDIA/cudf/pull/17270)) [@Matt711](https://github.com/Matt711) +- Move strings filter benchmarks to nvbench ([#17269](https://github.com/NVIDIA/cudf/pull/17269)) [@davidwendt](https://github.com/davidwendt) +- Make constructor of DeviceMemoryBufferView public ([#17265](https://github.com/NVIDIA/cudf/pull/17265)) [@liurenjie1024](https://github.com/liurenjie1024) +- Put a ceiling on cuda-python ([#17264](https://github.com/NVIDIA/cudf/pull/17264)) [@jameslamb](https://github.com/jameslamb) +- Always prefer `device_read`s and `device_write`s when kvikIO is enabled ([#17260](https://github.com/NVIDIA/cudf/pull/17260)) [@vuule](https://github.com/vuule) +- Expose streams in public quantile APIs ([#17257](https://github.com/NVIDIA/cudf/pull/17257)) [@shrshi](https://github.com/shrshi) +- Add support for `pyarrow-18` ([#17256](https://github.com/NVIDIA/cudf/pull/17256)) [@galipremsagar](https://github.com/galipremsagar) +- Move strings/numeric convert benchmarks to nvbench ([#17255](https://github.com/NVIDIA/cudf/pull/17255)) [@davidwendt](https://github.com/davidwendt) +- Add new ``dask_cudf.read_parquet`` API ([#17250](https://github.com/NVIDIA/cudf/pull/17250)) [@rjzamora](https://github.com/rjzamora) +- Add read_parquet_metadata to pylibcudf ([#17245](https://github.com/NVIDIA/cudf/pull/17245)) [@mroeschke](https://github.com/mroeschke) +- Search for kvikio with lowercase ([#17243](https://github.com/NVIDIA/cudf/pull/17243)) [@vyasr](https://github.com/vyasr) +- KvikIO shared library ([#17239](https://github.com/NVIDIA/cudf/pull/17239)) [@madsbk](https://github.com/madsbk) +- Use more pylibcudf.io.types enums in cudf._libs ([#17237](https://github.com/NVIDIA/cudf/pull/17237)) [@mroeschke](https://github.com/mroeschke) +- Expose mixed and conditional joins in pylibcudf ([#17235](https://github.com/NVIDIA/cudf/pull/17235)) [@wence-](https://github.com/wence-) +- Add io.text APIs to pylibcudf ([#17232](https://github.com/NVIDIA/cudf/pull/17232)) [@mroeschke](https://github.com/mroeschke) +- Add `num_iterations` axis to the multi-threaded Parquet benchmarks ([#17231](https://github.com/NVIDIA/cudf/pull/17231)) [@vuule](https://github.com/vuule) +- Move strings to date/time types benchmarks to nvbench ([#17229](https://github.com/NVIDIA/cudf/pull/17229)) [@davidwendt](https://github.com/davidwendt) +- Support for polars 1.12 in cudf-polars ([#17227](https://github.com/NVIDIA/cudf/pull/17227)) [@wence-](https://github.com/wence-) +- Allow generating large strings in benchmarks ([#17224](https://github.com/NVIDIA/cudf/pull/17224)) [@davidwendt](https://github.com/davidwendt) +- Refactor gather/scatter benchmarks for strings ([#17223](https://github.com/NVIDIA/cudf/pull/17223)) [@davidwendt](https://github.com/davidwendt) +- Deprecate single component extraction methods in libcudf ([#17221](https://github.com/NVIDIA/cudf/pull/17221)) [@Matt711](https://github.com/Matt711) +- Remove `nvtext::load_vocabulary` from pylibcudf ([#17220](https://github.com/NVIDIA/cudf/pull/17220)) [@Matt711](https://github.com/Matt711) +- Benchmarking JSON reader for compressed inputs ([#17219](https://github.com/NVIDIA/cudf/pull/17219)) [@shrshi](https://github.com/shrshi) +- Expose stream-ordering in partitioning API ([#17213](https://github.com/NVIDIA/cudf/pull/17213)) [@shrshi](https://github.com/shrshi) +- Move strings::concatenate benchmark to nvbench ([#17211](https://github.com/NVIDIA/cudf/pull/17211)) [@davidwendt](https://github.com/davidwendt) +- Expose stream-ordering in subword tokenizer API ([#17206](https://github.com/NVIDIA/cudf/pull/17206)) [@shrshi](https://github.com/shrshi) +- Refactor Dask cuDF legacy code ([#17205](https://github.com/NVIDIA/cudf/pull/17205)) [@rjzamora](https://github.com/rjzamora) +- Make HostMemoryBuffer call into the DefaultHostMemoryAllocator ([#17204](https://github.com/NVIDIA/cudf/pull/17204)) [@revans2](https://github.com/revans2) +- Unified binary_ops and ast benchmarks parameter names ([#17200](https://github.com/NVIDIA/cudf/pull/17200)) [@lamarrr](https://github.com/lamarrr) +- Add in new java API for raw host memory allocation ([#17197](https://github.com/NVIDIA/cudf/pull/17197)) [@revans2](https://github.com/revans2) +- Remove java reservation ([#17189](https://github.com/NVIDIA/cudf/pull/17189)) [@revans2](https://github.com/revans2) +- Fixed unused attribute compilation error for GCC 13 ([#17188](https://github.com/NVIDIA/cudf/pull/17188)) [@lamarrr](https://github.com/lamarrr) +- Change default KvikIO parameters in cuDF: set the thread pool size to 4, and compatibility mode to ON ([#17185](https://github.com/NVIDIA/cudf/pull/17185)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Use make_device_uvector instead of cudaMemcpyAsync in inplace_bitmask_binop ([#17181](https://github.com/NVIDIA/cudf/pull/17181)) [@davidwendt](https://github.com/davidwendt) +- Make ai.rapids.cudf.HostMemoryBuffer#copyFromStream public. ([#17179](https://github.com/NVIDIA/cudf/pull/17179)) [@liurenjie1024](https://github.com/liurenjie1024) +- Separate evaluation logic from `IR` objects in cudf-polars ([#17175](https://github.com/NVIDIA/cudf/pull/17175)) [@rjzamora](https://github.com/rjzamora) +- Move nvtext ngrams benchmarks to nvbench ([#17173](https://github.com/NVIDIA/cudf/pull/17173)) [@davidwendt](https://github.com/davidwendt) +- Remove includes suggested by include-what-you-use ([#17170](https://github.com/NVIDIA/cudf/pull/17170)) [@vyasr](https://github.com/vyasr) +- Reading multi-source compressed JSONL files ([#17161](https://github.com/NVIDIA/cudf/pull/17161)) [@shrshi](https://github.com/shrshi) +- Process parquet bools with microkernels ([#17157](https://github.com/NVIDIA/cudf/pull/17157)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Upgrade to polars 1.11 in cudf-polars ([#17154](https://github.com/NVIDIA/cudf/pull/17154)) [@wence-](https://github.com/wence-) +- Deprecate current libcudf nvtext minhash functions ([#17152](https://github.com/NVIDIA/cudf/pull/17152)) [@davidwendt](https://github.com/davidwendt) +- Remove unused variable in internal merge_tdigests utility ([#17151](https://github.com/NVIDIA/cudf/pull/17151)) [@davidwendt](https://github.com/davidwendt) +- Use the full ref name of `rmm.DeviceBuffer` in the sphinx config file ([#17150](https://github.com/NVIDIA/cudf/pull/17150)) [@Matt711](https://github.com/Matt711) +- Move `segmented_gather` function from the copying module to the lists module ([#17148](https://github.com/NVIDIA/cudf/pull/17148)) [@Matt711](https://github.com/Matt711) +- Use async execution policy for true_if ([#17146](https://github.com/NVIDIA/cudf/pull/17146)) [@PointKernel](https://github.com/PointKernel) +- Add conversion from cudf-polars expressions to libcudf ast for parquet filters ([#17141](https://github.com/NVIDIA/cudf/pull/17141)) [@wence-](https://github.com/wence-) +- devcontainer: replace `VAULT_HOST` with `AWS_ROLE_ARN` ([#17134](https://github.com/NVIDIA/cudf/pull/17134)) [@jjacobelli](https://github.com/jjacobelli) +- Replace direct `cudaMemcpyAsync` calls with utility functions (limited to `cudf::io`) ([#17132](https://github.com/NVIDIA/cudf/pull/17132)) [@vuule](https://github.com/vuule) +- use rapids-generate-pip-constraints to pin to oldest dependencies in CI ([#17131](https://github.com/NVIDIA/cudf/pull/17131)) [@jameslamb](https://github.com/jameslamb) +- Set the default number of threads in KvikIO thread pool to 8 ([#17126](https://github.com/NVIDIA/cudf/pull/17126)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Fix clang-tidy violations for span.hpp and hostdevice_vector.hpp ([#17124](https://github.com/NVIDIA/cudf/pull/17124)) [@davidwendt](https://github.com/davidwendt) +- Disable the Parquet reader's wide lists tables GTest by default ([#17120](https://github.com/NVIDIA/cudf/pull/17120)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add compile time check to ensure the `counting_iterator` type in `counting_transform_iterator` fits in `size_type` ([#17118](https://github.com/NVIDIA/cudf/pull/17118)) [@mhaseeb123](https://github.com/mhaseeb123) +- Minor I/O code quality improvements ([#17105](https://github.com/NVIDIA/cudf/pull/17105)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Remove the additional host register calls initially intended for performance improvement on Grace Hopper ([#17092](https://github.com/NVIDIA/cudf/pull/17092)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Split hash-based groupby into multiple smaller files to reduce build time ([#17089](https://github.com/NVIDIA/cudf/pull/17089)) [@PointKernel](https://github.com/PointKernel) +- build wheels without build isolation ([#17088](https://github.com/NVIDIA/cudf/pull/17088)) [@jameslamb](https://github.com/jameslamb) +- Polars: DataFrame Serialization ([#17062](https://github.com/NVIDIA/cudf/pull/17062)) [@madsbk](https://github.com/madsbk) +- Remove unused hash helper functions ([#17056](https://github.com/NVIDIA/cudf/pull/17056)) [@PointKernel](https://github.com/PointKernel) +- Add to_dlpack/from_dlpack APIs to pylibcudf ([#17055](https://github.com/NVIDIA/cudf/pull/17055)) [@mroeschke](https://github.com/mroeschke) +- Move `flatten_single_pass_aggs` to its own TU ([#17053](https://github.com/NVIDIA/cudf/pull/17053)) [@PointKernel](https://github.com/PointKernel) +- Replace deprecated cuco APIs with updated versions ([#17052](https://github.com/NVIDIA/cudf/pull/17052)) [@PointKernel](https://github.com/PointKernel) +- Refactor ORC dictionary encoding to migrate to the new `cuco::static_map` ([#17049](https://github.com/NVIDIA/cudf/pull/17049)) [@mhaseeb123](https://github.com/mhaseeb123) +- Move pylibcudf/libcudf/wrappers/decimals to pylibcudf/libcudf/fixed_point ([#17048](https://github.com/NVIDIA/cudf/pull/17048)) [@mroeschke](https://github.com/mroeschke) +- make conda installs in CI stricter (part 2) ([#17042](https://github.com/NVIDIA/cudf/pull/17042)) [@jameslamb](https://github.com/jameslamb) +- Use managed memory for NDSH benchmarks ([#17039](https://github.com/NVIDIA/cudf/pull/17039)) [@karthikeyann](https://github.com/karthikeyann) +- Clean up hash-groupby `var_hash_functor` ([#17034](https://github.com/NVIDIA/cudf/pull/17034)) [@PointKernel](https://github.com/PointKernel) +- Add json APIs to pylibcudf ([#17025](https://github.com/NVIDIA/cudf/pull/17025)) [@mroeschke](https://github.com/mroeschke) +- Add string.replace_re APIs to pylibcudf ([#17023](https://github.com/NVIDIA/cudf/pull/17023)) [@mroeschke](https://github.com/mroeschke) +- Replace old host tree algorithm with new algorithm in JSON reader ([#17019](https://github.com/NVIDIA/cudf/pull/17019)) [@karthikeyann](https://github.com/karthikeyann) +- Unify treatment of `Expr` and `IR` nodes in cudf-polars DSL ([#17016](https://github.com/NVIDIA/cudf/pull/17016)) [@wence-](https://github.com/wence-) +- make conda installs in CI stricter ([#17013](https://github.com/NVIDIA/cudf/pull/17013)) [@jameslamb](https://github.com/jameslamb) +- Pylibcudf: pack and unpack ([#17012](https://github.com/NVIDIA/cudf/pull/17012)) [@madsbk](https://github.com/madsbk) +- Remove unneeded pylibcudf.libcudf.wrappers.duration usage in cudf ([#17010](https://github.com/NVIDIA/cudf/pull/17010)) [@mroeschke](https://github.com/mroeschke) +- Add custom "fused" groupby aggregation to Dask cuDF ([#17009](https://github.com/NVIDIA/cudf/pull/17009)) [@rjzamora](https://github.com/rjzamora) +- Make tests more deterministic ([#17008](https://github.com/NVIDIA/cudf/pull/17008)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unused import ([#17005](https://github.com/NVIDIA/cudf/pull/17005)) [@Matt711](https://github.com/Matt711) +- Add string.convert.convert_urls APIs to pylibcudf ([#17003](https://github.com/NVIDIA/cudf/pull/17003)) [@mroeschke](https://github.com/mroeschke) +- Add release tracking to project automation scripts ([#17001](https://github.com/NVIDIA/cudf/pull/17001)) [@jarmak-nv](https://github.com/jarmak-nv) +- Implement inequality joins by translation to conditional joins ([#17000](https://github.com/NVIDIA/cudf/pull/17000)) [@wence-](https://github.com/wence-) +- Add string.convert.convert_lists APIs to pylibcudf ([#16997](https://github.com/NVIDIA/cudf/pull/16997)) [@mroeschke](https://github.com/mroeschke) +- Performance optimization of JSON validation ([#16996](https://github.com/NVIDIA/cudf/pull/16996)) [@karthikeyann](https://github.com/karthikeyann) +- Add string.convert.convert_ipv4 APIs to pylibcudf ([#16994](https://github.com/NVIDIA/cudf/pull/16994)) [@mroeschke](https://github.com/mroeschke) +- Add string.convert.convert_integers APIs to pylibcudf ([#16991](https://github.com/NVIDIA/cudf/pull/16991)) [@mroeschke](https://github.com/mroeschke) +- Add string.convert_floats APIs to pylibcudf ([#16990](https://github.com/NVIDIA/cudf/pull/16990)) [@mroeschke](https://github.com/mroeschke) +- Add string.convert.convert_fixed_type APIs to pylibcudf ([#16984](https://github.com/NVIDIA/cudf/pull/16984)) [@mroeschke](https://github.com/mroeschke) +- Remove unnecessary `std::move`'s in pylibcudf ([#16983](https://github.com/NVIDIA/cudf/pull/16983)) [@Matt711](https://github.com/Matt711) +- Add docstrings and test for strings.convert_durations APIs for pylibcudf ([#16982](https://github.com/NVIDIA/cudf/pull/16982)) [@mroeschke](https://github.com/mroeschke) +- JSON tokenizer memory optimizations ([#16978](https://github.com/NVIDIA/cudf/pull/16978)) [@shrshi](https://github.com/shrshi) +- Turn on `xfail_strict = true` for all python packages ([#16977](https://github.com/NVIDIA/cudf/pull/16977)) [@wence-](https://github.com/wence-) +- Add string.convert.convert_datetime/convert_booleans APIs to pylibcudf ([#16971](https://github.com/NVIDIA/cudf/pull/16971)) [@mroeschke](https://github.com/mroeschke) +- Auto assign PR to author ([#16969](https://github.com/NVIDIA/cudf/pull/16969)) [@Matt711](https://github.com/Matt711) +- Deprecate support for directly accessing logger ([#16964](https://github.com/NVIDIA/cudf/pull/16964)) [@vyasr](https://github.com/vyasr) +- Expunge NamedColumn ([#16962](https://github.com/NVIDIA/cudf/pull/16962)) [@wence-](https://github.com/wence-) +- Add clang-tidy to CI ([#16958](https://github.com/NVIDIA/cudf/pull/16958)) [@vyasr](https://github.com/vyasr) +- Address all remaining clang-tidy errors ([#16956](https://github.com/NVIDIA/cudf/pull/16956)) [@vyasr](https://github.com/vyasr) +- Apply clang-tidy autofixes ([#16949](https://github.com/NVIDIA/cudf/pull/16949)) [@vyasr](https://github.com/vyasr) +- Use nvcomp wheel instead of bundling nvcomp ([#16946](https://github.com/NVIDIA/cudf/pull/16946)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Refactor the `cuda_memcpy` functions to make them more usable ([#16945](https://github.com/NVIDIA/cudf/pull/16945)) [@vuule](https://github.com/vuule) +- Add string.split APIs to pylibcudf ([#16940](https://github.com/NVIDIA/cudf/pull/16940)) [@mroeschke](https://github.com/mroeschke) +- clang-tidy fixes part 3 ([#16939](https://github.com/NVIDIA/cudf/pull/16939)) [@vyasr](https://github.com/vyasr) +- clang-tidy fixes part 2 ([#16938](https://github.com/NVIDIA/cudf/pull/16938)) [@vyasr](https://github.com/vyasr) +- clang-tidy fixes part 1 ([#16937](https://github.com/NVIDIA/cudf/pull/16937)) [@vyasr](https://github.com/vyasr) +- Add string.wrap APIs to pylibcudf ([#16935](https://github.com/NVIDIA/cudf/pull/16935)) [@mroeschke](https://github.com/mroeschke) +- Add string.translate APIs to pylibcudf ([#16934](https://github.com/NVIDIA/cudf/pull/16934)) [@mroeschke](https://github.com/mroeschke) +- Add string.find_multiple APIs to pylibcudf ([#16920](https://github.com/NVIDIA/cudf/pull/16920)) [@mroeschke](https://github.com/mroeschke) +- Batch memcpy the last offsets for output buffers of str and list cols in PQ reader ([#16905](https://github.com/NVIDIA/cudf/pull/16905)) [@mhaseeb123](https://github.com/mhaseeb123) +- reduce wheel build verbosity, narrow deprecation warning filter ([#16896](https://github.com/NVIDIA/cudf/pull/16896)) [@jameslamb](https://github.com/jameslamb) +- Improve aggregation device functors ([#16884](https://github.com/NVIDIA/cudf/pull/16884)) [@PointKernel](https://github.com/PointKernel) +- Upgrade pandas pinnings to support `2.2.3` ([#16882](https://github.com/NVIDIA/cudf/pull/16882)) [@galipremsagar](https://github.com/galipremsagar) +- Fix 24.10 to 24.12 forward merge ([#16876](https://github.com/NVIDIA/cudf/pull/16876)) [@bdice](https://github.com/bdice) +- Manually resolve conflicts in between branch-24.12 and branch-24.10 ([#16871](https://github.com/NVIDIA/cudf/pull/16871)) [@galipremsagar](https://github.com/galipremsagar) +- Add in support for setting delim when parsing JSON through java ([#16867](https://github.com/NVIDIA/cudf/pull/16867)) [@revans2](https://github.com/revans2) +- Reapply `mixed_semi_join` refactoring and bug fixes ([#16859](https://github.com/NVIDIA/cudf/pull/16859)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add string padding and side_type APIs to pylibcudf ([#16833](https://github.com/NVIDIA/cudf/pull/16833)) [@mroeschke](https://github.com/mroeschke) +- Organize parquet reader mukernel non-nullable code, introduce manual block scans ([#16830](https://github.com/NVIDIA/cudf/pull/16830)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Remove superfluous use of std::vector for std::future ([#16829](https://github.com/NVIDIA/cudf/pull/16829)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Rework `read_csv` IO to avoid reading whole input with a single `host_read` ([#16826](https://github.com/NVIDIA/cudf/pull/16826)) [@vuule](https://github.com/vuule) +- Add strings.combine APIs to pylibcudf ([#16790](https://github.com/NVIDIA/cudf/pull/16790)) [@mroeschke](https://github.com/mroeschke) +- Add remaining string.char_types APIs to pylibcudf ([#16788](https://github.com/NVIDIA/cudf/pull/16788)) [@mroeschke](https://github.com/mroeschke) +- Add new nvtext minhash_permuted API ([#16756](https://github.com/NVIDIA/cudf/pull/16756)) [@davidwendt](https://github.com/davidwendt) +- Avoid public constructors when called with columns to avoid unnecessary validation ([#16747](https://github.com/NVIDIA/cudf/pull/16747)) [@mroeschke](https://github.com/mroeschke) +- Use `changed-files` shared workflow ([#16713](https://github.com/NVIDIA/cudf/pull/16713)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- lint: replace `isort` with Ruff's rule I ([#16685](https://github.com/NVIDIA/cudf/pull/16685)) [@Borda](https://github.com/Borda) +- Improve the performance of low cardinality groupby ([#16619](https://github.com/NVIDIA/cudf/pull/16619)) [@PointKernel](https://github.com/PointKernel) +- Parquet reader list microkernel ([#16538](https://github.com/NVIDIA/cudf/pull/16538)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- AWS S3 IO through KvikIO ([#16499](https://github.com/NVIDIA/cudf/pull/16499)) [@madsbk](https://github.com/madsbk) +- Refactor `histogram` reduction using `cuco::static_set::insert_and_find` ([#16485](https://github.com/NVIDIA/cudf/pull/16485)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Use numba-cuda>=0.0.13 ([#16474](https://github.com/NVIDIA/cudf/pull/16474)) [@gmarkall](https://github.com/gmarkall) # cudf 24.10.00 (9 Oct 2024) ## 🚨 Breaking Changes -- Whitespace normalization of nested column coerced as string column in JSONL inputs ([#16759](https://github.com/rapidsai/cudf/pull/16759)) [@shrshi](https://github.com/shrshi) -- Add libcudf wrappers around current_device_resource functions. ([#16679](https://github.com/rapidsai/cudf/pull/16679)) [@harrism](https://github.com/harrism) -- Fix empty cluster handling in tdigest merge ([#16675](https://github.com/rapidsai/cudf/pull/16675)) [@jihoonson](https://github.com/jihoonson) -- Remove java ColumnView.copyWithBooleanColumnAsValidity ([#16660](https://github.com/rapidsai/cudf/pull/16660)) [@revans2](https://github.com/revans2) -- Support reading multiple PQ sources with mismatching nullability for columns ([#16639](https://github.com/rapidsai/cudf/pull/16639)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove arrow_io_source ([#16607](https://github.com/rapidsai/cudf/pull/16607)) [@vyasr](https://github.com/vyasr) -- Remove legacy Arrow interop APIs ([#16590](https://github.com/rapidsai/cudf/pull/16590)) [@vyasr](https://github.com/vyasr) -- Remove NativeFile support from cudf Python ([#16589](https://github.com/rapidsai/cudf/pull/16589)) [@vyasr](https://github.com/vyasr) -- Revert "Make proxy NumPy arrays pass isinstance check in `cudf.pandas`" ([#16586](https://github.com/rapidsai/cudf/pull/16586)) [@Matt711](https://github.com/Matt711) -- Align public utility function signatures with pandas 2.x ([#16565](https://github.com/rapidsai/cudf/pull/16565)) [@mroeschke](https://github.com/mroeschke) -- Disallow cudf.Index accepting column in favor of ._from_column ([#16549](https://github.com/rapidsai/cudf/pull/16549)) [@mroeschke](https://github.com/mroeschke) -- Refactor dictionary encoding in PQ writer to migrate to the new `cuco::static_map` ([#16541](https://github.com/rapidsai/cudf/pull/16541)) [@mhaseeb123](https://github.com/mhaseeb123) -- Change IPv4 convert APIs to support UINT32 instead of INT64 ([#16489](https://github.com/rapidsai/cudf/pull/16489)) [@davidwendt](https://github.com/davidwendt) -- enable list to be forced as string in JSON reader. ([#16472](https://github.com/rapidsai/cudf/pull/16472)) [@karthikeyann](https://github.com/karthikeyann) -- Disallow cudf.Series to accept column in favor of `._from_column` ([#16454](https://github.com/rapidsai/cudf/pull/16454)) [@mroeschke](https://github.com/mroeschke) -- Align groupby APIs with pandas 2.x ([#16403](https://github.com/rapidsai/cudf/pull/16403)) [@mroeschke](https://github.com/mroeschke) -- Align misc DataFrame and MultiIndex methods with pandas 2.x ([#16402](https://github.com/rapidsai/cudf/pull/16402)) [@mroeschke](https://github.com/mroeschke) -- Align Index APIs with pandas 2.x ([#16361](https://github.com/rapidsai/cudf/pull/16361)) [@mroeschke](https://github.com/mroeschke) -- Add `stream` param to stream compaction APIs ([#16295](https://github.com/rapidsai/cudf/pull/16295)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Whitespace normalization of nested column coerced as string column in JSONL inputs ([#16759](https://github.com/NVIDIA/cudf/pull/16759)) [@shrshi](https://github.com/shrshi) +- Add libcudf wrappers around current_device_resource functions. ([#16679](https://github.com/NVIDIA/cudf/pull/16679)) [@harrism](https://github.com/harrism) +- Fix empty cluster handling in tdigest merge ([#16675](https://github.com/NVIDIA/cudf/pull/16675)) [@jihoonson](https://github.com/jihoonson) +- Remove java ColumnView.copyWithBooleanColumnAsValidity ([#16660](https://github.com/NVIDIA/cudf/pull/16660)) [@revans2](https://github.com/revans2) +- Support reading multiple PQ sources with mismatching nullability for columns ([#16639](https://github.com/NVIDIA/cudf/pull/16639)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove arrow_io_source ([#16607](https://github.com/NVIDIA/cudf/pull/16607)) [@vyasr](https://github.com/vyasr) +- Remove legacy Arrow interop APIs ([#16590](https://github.com/NVIDIA/cudf/pull/16590)) [@vyasr](https://github.com/vyasr) +- Remove NativeFile support from cudf Python ([#16589](https://github.com/NVIDIA/cudf/pull/16589)) [@vyasr](https://github.com/vyasr) +- Revert "Make proxy NumPy arrays pass isinstance check in `cudf.pandas`" ([#16586](https://github.com/NVIDIA/cudf/pull/16586)) [@Matt711](https://github.com/Matt711) +- Align public utility function signatures with pandas 2.x ([#16565](https://github.com/NVIDIA/cudf/pull/16565)) [@mroeschke](https://github.com/mroeschke) +- Disallow cudf.Index accepting column in favor of ._from_column ([#16549](https://github.com/NVIDIA/cudf/pull/16549)) [@mroeschke](https://github.com/mroeschke) +- Refactor dictionary encoding in PQ writer to migrate to the new `cuco::static_map` ([#16541](https://github.com/NVIDIA/cudf/pull/16541)) [@mhaseeb123](https://github.com/mhaseeb123) +- Change IPv4 convert APIs to support UINT32 instead of INT64 ([#16489](https://github.com/NVIDIA/cudf/pull/16489)) [@davidwendt](https://github.com/davidwendt) +- enable list to be forced as string in JSON reader. ([#16472](https://github.com/NVIDIA/cudf/pull/16472)) [@karthikeyann](https://github.com/karthikeyann) +- Disallow cudf.Series to accept column in favor of `._from_column` ([#16454](https://github.com/NVIDIA/cudf/pull/16454)) [@mroeschke](https://github.com/mroeschke) +- Align groupby APIs with pandas 2.x ([#16403](https://github.com/NVIDIA/cudf/pull/16403)) [@mroeschke](https://github.com/mroeschke) +- Align misc DataFrame and MultiIndex methods with pandas 2.x ([#16402](https://github.com/NVIDIA/cudf/pull/16402)) [@mroeschke](https://github.com/mroeschke) +- Align Index APIs with pandas 2.x ([#16361](https://github.com/NVIDIA/cudf/pull/16361)) [@mroeschke](https://github.com/mroeschke) +- Add `stream` param to stream compaction APIs ([#16295](https://github.com/NVIDIA/cudf/pull/16295)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) ## 🐛 Bug Fixes -- Add license to the pylibcudf wheel ([#16976](https://github.com/rapidsai/cudf/pull/16976)) [@raydouglass](https://github.com/raydouglass) -- Parse newline as whitespace character while tokenizing JSONL inputs with non-newline delimiter ([#16950](https://github.com/rapidsai/cudf/pull/16950)) [@shrshi](https://github.com/shrshi) -- Add dask-cudf workaround for missing `rename_axis` support in cudf ([#16899](https://github.com/rapidsai/cudf/pull/16899)) [@rjzamora](https://github.com/rjzamora) -- Update oldest deps for `pyarrow` & `numpy` ([#16883](https://github.com/rapidsai/cudf/pull/16883)) [@galipremsagar](https://github.com/galipremsagar) -- Update labeler for pylibcudf ([#16868](https://github.com/rapidsai/cudf/pull/16868)) [@vyasr](https://github.com/vyasr) -- Revert "Refactor mixed_semi_join using cuco::static_set" ([#16855](https://github.com/rapidsai/cudf/pull/16855)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix metadata after implicit array conversion from Dask cuDF ([#16842](https://github.com/rapidsai/cudf/pull/16842)) [@rjzamora](https://github.com/rjzamora) -- Add cudf.pandas dependencies.yaml to update-version.sh ([#16840](https://github.com/rapidsai/cudf/pull/16840)) [@raydouglass](https://github.com/raydouglass) -- Use cupy 12.2.0 as oldest dependency pinning on CUDA 12 ARM ([#16808](https://github.com/rapidsai/cudf/pull/16808)) [@bdice](https://github.com/bdice) -- Revert "Fix empty cluster handling in tdigest merge ([#16675)" (#16800](https://github.com/rapidsai/cudf/pull/16675)" (#16800)) [@jihoonson](https://github.com/jihoonson) -- Intentionally leak thread_local CUDA resources to avoid crash (part 1) ([#16787](https://github.com/rapidsai/cudf/pull/16787)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) -- Fix `cov`/`corr` bug in dask-cudf ([#16786](https://github.com/rapidsai/cudf/pull/16786)) [@rjzamora](https://github.com/rjzamora) -- Fix slice_strings wide strings logic with multi-byte characters ([#16777](https://github.com/rapidsai/cudf/pull/16777)) [@davidwendt](https://github.com/davidwendt) -- Fix nvbench output for sha512 ([#16773](https://github.com/rapidsai/cudf/pull/16773)) [@davidwendt](https://github.com/davidwendt) -- Allow read_csv(header=None) to return int column labels in `mode.pandas_compatible` ([#16769](https://github.com/rapidsai/cudf/pull/16769)) [@mroeschke](https://github.com/mroeschke) -- Whitespace normalization of nested column coerced as string column in JSONL inputs ([#16759](https://github.com/rapidsai/cudf/pull/16759)) [@shrshi](https://github.com/shrshi) -- Fix DataFrame.drop(columns=cudf.Series/Index, axis=1) ([#16712](https://github.com/rapidsai/cudf/pull/16712)) [@mroeschke](https://github.com/mroeschke) -- Use merge base when calculating changed files ([#16709](https://github.com/rapidsai/cudf/pull/16709)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Ensure we pass the has_nulls tparam to mixed_join kernels ([#16708](https://github.com/rapidsai/cudf/pull/16708)) [@abellina](https://github.com/abellina) -- Add boost-devel to Java CI Docker image ([#16707](https://github.com/rapidsai/cudf/pull/16707)) [@jlowe](https://github.com/jlowe) -- [BUG] Add gpu node type to cudf-pandas 3rd-party integration nightly CI job ([#16704](https://github.com/rapidsai/cudf/pull/16704)) [@Matt711](https://github.com/Matt711) -- Fix typo in column_factories.hpp comment from 'depth 1' to 'depth 2' ([#16700](https://github.com/rapidsai/cudf/pull/16700)) [@a-hirota](https://github.com/a-hirota) -- Fix Series.to_frame(name=None) setting a None name ([#16698](https://github.com/rapidsai/cudf/pull/16698)) [@mroeschke](https://github.com/mroeschke) -- Disable gtests/ERROR_TEST during compute-sanitizer memcheck test ([#16691](https://github.com/rapidsai/cudf/pull/16691)) [@davidwendt](https://github.com/davidwendt) -- Enable batched multi-source reading of JSONL files with large records ([#16687](https://github.com/rapidsai/cudf/pull/16687)) [@shrshi](https://github.com/shrshi) -- Handle `ordered` parameter in `CategoricalIndex.__repr__` ([#16683](https://github.com/rapidsai/cudf/pull/16683)) [@galipremsagar](https://github.com/galipremsagar) -- Fix loc/iloc.__setitem__[:, loc] with non cupy types ([#16677](https://github.com/rapidsai/cudf/pull/16677)) [@mroeschke](https://github.com/mroeschke) -- Fix empty cluster handling in tdigest merge ([#16675](https://github.com/rapidsai/cudf/pull/16675)) [@jihoonson](https://github.com/jihoonson) -- Fix `cudf::rank` not getting enough params ([#16666](https://github.com/rapidsai/cudf/pull/16666)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Fix slowdown in `CategoricalIndex.__repr__` ([#16665](https://github.com/rapidsai/cudf/pull/16665)) [@galipremsagar](https://github.com/galipremsagar) -- Remove java ColumnView.copyWithBooleanColumnAsValidity ([#16660](https://github.com/rapidsai/cudf/pull/16660)) [@revans2](https://github.com/revans2) -- Fix slowdown in DataFrame repr in jupyter notebook ([#16656](https://github.com/rapidsai/cudf/pull/16656)) [@galipremsagar](https://github.com/galipremsagar) -- Preserve Series name in duplicated method. ([#16655](https://github.com/rapidsai/cudf/pull/16655)) [@bdice](https://github.com/bdice) -- Fix interval_range right child non-zero offset ([#16651](https://github.com/rapidsai/cudf/pull/16651)) [@mroeschke](https://github.com/mroeschke) -- fix libcudf wheel publishing, make package-type explicit in wheel publishing ([#16650](https://github.com/rapidsai/cudf/pull/16650)) [@jameslamb](https://github.com/jameslamb) -- Revert "Hide all gtest symbols in cudftestutil ([#16546)" (#16644](https://github.com/rapidsai/cudf/pull/16546)" (#16644)) [@robertmaynard](https://github.com/robertmaynard) -- Fix integer overflow in indexalator pointer logic ([#16643](https://github.com/rapidsai/cudf/pull/16643)) [@davidwendt](https://github.com/davidwendt) -- Allow for binops between two differently sized DecimalDtypes ([#16638](https://github.com/rapidsai/cudf/pull/16638)) [@mroeschke](https://github.com/mroeschke) -- Move pragma once in rolling/jit/operation.hpp. ([#16636](https://github.com/rapidsai/cudf/pull/16636)) [@bdice](https://github.com/bdice) -- Fix overflow bug in low-memory JSON reader ([#16632](https://github.com/rapidsai/cudf/pull/16632)) [@shrshi](https://github.com/shrshi) -- Add the missing `num_aggregations` axis for `groupby_max_cardinality` ([#16630](https://github.com/rapidsai/cudf/pull/16630)) [@PointKernel](https://github.com/PointKernel) -- Fix strings::detail::copy_range when target contains nulls ([#16626](https://github.com/rapidsai/cudf/pull/16626)) [@davidwendt](https://github.com/davidwendt) -- Fix function parameters with common dependency modified during their evaluation ([#16620](https://github.com/rapidsai/cudf/pull/16620)) [@ttnghia](https://github.com/ttnghia) -- bug-fix: Don't enable the CUDA language if testing was requested when finding cudf ([#16615](https://github.com/rapidsai/cudf/pull/16615)) [@cryos](https://github.com/cryos) -- bug-fix: cudf/io/json.hpp use after move ([#16609](https://github.com/rapidsai/cudf/pull/16609)) [@NicolasDenoyelle](https://github.com/NicolasDenoyelle) -- Remove CUDA whole compilation ODR violations ([#16603](https://github.com/rapidsai/cudf/pull/16603)) [@robertmaynard](https://github.com/robertmaynard) -- MAINT: Adapt to numpy hiding flagsobject away ([#16593](https://github.com/rapidsai/cudf/pull/16593)) [@seberg](https://github.com/seberg) -- Revert "Make proxy NumPy arrays pass isinstance check in `cudf.pandas`" ([#16586](https://github.com/rapidsai/cudf/pull/16586)) [@Matt711](https://github.com/Matt711) -- Switch python version to `3.10` in `cudf.pandas` pandas test scripts ([#16559](https://github.com/rapidsai/cudf/pull/16559)) [@galipremsagar](https://github.com/galipremsagar) -- Hide all gtest symbols in cudftestutil ([#16546](https://github.com/rapidsai/cudf/pull/16546)) [@robertmaynard](https://github.com/robertmaynard) -- Update the java code to properly deal with lists being returned as strings ([#16536](https://github.com/rapidsai/cudf/pull/16536)) [@revans2](https://github.com/revans2) -- Register `read_parquet` and `read_csv` with dask-expr ([#16535](https://github.com/rapidsai/cudf/pull/16535)) [@rjzamora](https://github.com/rjzamora) -- Change cudf::empty_like to not include offsets for empty strings columns ([#16529](https://github.com/rapidsai/cudf/pull/16529)) [@davidwendt](https://github.com/davidwendt) -- Fix DataFrame reductions with median returning scalar instead of Series ([#16527](https://github.com/rapidsai/cudf/pull/16527)) [@mroeschke](https://github.com/mroeschke) -- Allow DataFrame.sort_values(by=) to select an index level ([#16519](https://github.com/rapidsai/cudf/pull/16519)) [@mroeschke](https://github.com/mroeschke) -- Fix `date_range(start, end, freq)` when end-start is divisible by freq ([#16516](https://github.com/rapidsai/cudf/pull/16516)) [@mroeschke](https://github.com/mroeschke) -- Preserve array name in MultiIndex.from_arrays ([#16515](https://github.com/rapidsai/cudf/pull/16515)) [@mroeschke](https://github.com/mroeschke) -- Disallow indexing by selecting duplicate labels ([#16514](https://github.com/rapidsai/cudf/pull/16514)) [@mroeschke](https://github.com/mroeschke) -- Fix `.replace(Index, Index)` raising a TypeError ([#16513](https://github.com/rapidsai/cudf/pull/16513)) [@mroeschke](https://github.com/mroeschke) -- Check index bounds in compact protocol reader. ([#16493](https://github.com/rapidsai/cudf/pull/16493)) [@bdice](https://github.com/bdice) -- Fix build failures with GCC 13 ([#16488](https://github.com/rapidsai/cudf/pull/16488)) [@PointKernel](https://github.com/PointKernel) -- Fix all-empty input column for strings split APIs ([#16466](https://github.com/rapidsai/cudf/pull/16466)) [@davidwendt](https://github.com/davidwendt) -- Fix segmented-sort overlapped input/output indices ([#16463](https://github.com/rapidsai/cudf/pull/16463)) [@davidwendt](https://github.com/davidwendt) -- Fix merge conflict for auto merge 16447 ([#16449](https://github.com/rapidsai/cudf/pull/16449)) [@davidwendt](https://github.com/davidwendt) +- Add license to the pylibcudf wheel ([#16976](https://github.com/NVIDIA/cudf/pull/16976)) [@raydouglass](https://github.com/raydouglass) +- Parse newline as whitespace character while tokenizing JSONL inputs with non-newline delimiter ([#16950](https://github.com/NVIDIA/cudf/pull/16950)) [@shrshi](https://github.com/shrshi) +- Add dask-cudf workaround for missing `rename_axis` support in cudf ([#16899](https://github.com/NVIDIA/cudf/pull/16899)) [@rjzamora](https://github.com/rjzamora) +- Update oldest deps for `pyarrow` & `numpy` ([#16883](https://github.com/NVIDIA/cudf/pull/16883)) [@galipremsagar](https://github.com/galipremsagar) +- Update labeler for pylibcudf ([#16868](https://github.com/NVIDIA/cudf/pull/16868)) [@vyasr](https://github.com/vyasr) +- Revert "Refactor mixed_semi_join using cuco::static_set" ([#16855](https://github.com/NVIDIA/cudf/pull/16855)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix metadata after implicit array conversion from Dask cuDF ([#16842](https://github.com/NVIDIA/cudf/pull/16842)) [@rjzamora](https://github.com/rjzamora) +- Add cudf.pandas dependencies.yaml to update-version.sh ([#16840](https://github.com/NVIDIA/cudf/pull/16840)) [@raydouglass](https://github.com/raydouglass) +- Use cupy 12.2.0 as oldest dependency pinning on CUDA 12 ARM ([#16808](https://github.com/NVIDIA/cudf/pull/16808)) [@bdice](https://github.com/bdice) +- Revert "Fix empty cluster handling in tdigest merge ([#16675)" (#16800](https://github.com/NVIDIA/cudf/pull/16675)" (#16800)) [@jihoonson](https://github.com/jihoonson) +- Intentionally leak thread_local CUDA resources to avoid crash (part 1) ([#16787](https://github.com/NVIDIA/cudf/pull/16787)) [@kingcrimsontianyu](https://github.com/kingcrimsontianyu) +- Fix `cov`/`corr` bug in dask-cudf ([#16786](https://github.com/NVIDIA/cudf/pull/16786)) [@rjzamora](https://github.com/rjzamora) +- Fix slice_strings wide strings logic with multi-byte characters ([#16777](https://github.com/NVIDIA/cudf/pull/16777)) [@davidwendt](https://github.com/davidwendt) +- Fix nvbench output for sha512 ([#16773](https://github.com/NVIDIA/cudf/pull/16773)) [@davidwendt](https://github.com/davidwendt) +- Allow read_csv(header=None) to return int column labels in `mode.pandas_compatible` ([#16769](https://github.com/NVIDIA/cudf/pull/16769)) [@mroeschke](https://github.com/mroeschke) +- Whitespace normalization of nested column coerced as string column in JSONL inputs ([#16759](https://github.com/NVIDIA/cudf/pull/16759)) [@shrshi](https://github.com/shrshi) +- Fix DataFrame.drop(columns=cudf.Series/Index, axis=1) ([#16712](https://github.com/NVIDIA/cudf/pull/16712)) [@mroeschke](https://github.com/mroeschke) +- Use merge base when calculating changed files ([#16709](https://github.com/NVIDIA/cudf/pull/16709)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Ensure we pass the has_nulls tparam to mixed_join kernels ([#16708](https://github.com/NVIDIA/cudf/pull/16708)) [@abellina](https://github.com/abellina) +- Add boost-devel to Java CI Docker image ([#16707](https://github.com/NVIDIA/cudf/pull/16707)) [@jlowe](https://github.com/jlowe) +- [BUG] Add gpu node type to cudf-pandas 3rd-party integration nightly CI job ([#16704](https://github.com/NVIDIA/cudf/pull/16704)) [@Matt711](https://github.com/Matt711) +- Fix typo in column_factories.hpp comment from 'depth 1' to 'depth 2' ([#16700](https://github.com/NVIDIA/cudf/pull/16700)) [@a-hirota](https://github.com/a-hirota) +- Fix Series.to_frame(name=None) setting a None name ([#16698](https://github.com/NVIDIA/cudf/pull/16698)) [@mroeschke](https://github.com/mroeschke) +- Disable gtests/ERROR_TEST during compute-sanitizer memcheck test ([#16691](https://github.com/NVIDIA/cudf/pull/16691)) [@davidwendt](https://github.com/davidwendt) +- Enable batched multi-source reading of JSONL files with large records ([#16687](https://github.com/NVIDIA/cudf/pull/16687)) [@shrshi](https://github.com/shrshi) +- Handle `ordered` parameter in `CategoricalIndex.__repr__` ([#16683](https://github.com/NVIDIA/cudf/pull/16683)) [@galipremsagar](https://github.com/galipremsagar) +- Fix loc/iloc.__setitem__[:, loc] with non cupy types ([#16677](https://github.com/NVIDIA/cudf/pull/16677)) [@mroeschke](https://github.com/mroeschke) +- Fix empty cluster handling in tdigest merge ([#16675](https://github.com/NVIDIA/cudf/pull/16675)) [@jihoonson](https://github.com/jihoonson) +- Fix `cudf::rank` not getting enough params ([#16666](https://github.com/NVIDIA/cudf/pull/16666)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Fix slowdown in `CategoricalIndex.__repr__` ([#16665](https://github.com/NVIDIA/cudf/pull/16665)) [@galipremsagar](https://github.com/galipremsagar) +- Remove java ColumnView.copyWithBooleanColumnAsValidity ([#16660](https://github.com/NVIDIA/cudf/pull/16660)) [@revans2](https://github.com/revans2) +- Fix slowdown in DataFrame repr in jupyter notebook ([#16656](https://github.com/NVIDIA/cudf/pull/16656)) [@galipremsagar](https://github.com/galipremsagar) +- Preserve Series name in duplicated method. ([#16655](https://github.com/NVIDIA/cudf/pull/16655)) [@bdice](https://github.com/bdice) +- Fix interval_range right child non-zero offset ([#16651](https://github.com/NVIDIA/cudf/pull/16651)) [@mroeschke](https://github.com/mroeschke) +- fix libcudf wheel publishing, make package-type explicit in wheel publishing ([#16650](https://github.com/NVIDIA/cudf/pull/16650)) [@jameslamb](https://github.com/jameslamb) +- Revert "Hide all gtest symbols in cudftestutil ([#16546)" (#16644](https://github.com/NVIDIA/cudf/pull/16546)" (#16644)) [@robertmaynard](https://github.com/robertmaynard) +- Fix integer overflow in indexalator pointer logic ([#16643](https://github.com/NVIDIA/cudf/pull/16643)) [@davidwendt](https://github.com/davidwendt) +- Allow for binops between two differently sized DecimalDtypes ([#16638](https://github.com/NVIDIA/cudf/pull/16638)) [@mroeschke](https://github.com/mroeschke) +- Move pragma once in rolling/jit/operation.hpp. ([#16636](https://github.com/NVIDIA/cudf/pull/16636)) [@bdice](https://github.com/bdice) +- Fix overflow bug in low-memory JSON reader ([#16632](https://github.com/NVIDIA/cudf/pull/16632)) [@shrshi](https://github.com/shrshi) +- Add the missing `num_aggregations` axis for `groupby_max_cardinality` ([#16630](https://github.com/NVIDIA/cudf/pull/16630)) [@PointKernel](https://github.com/PointKernel) +- Fix strings::detail::copy_range when target contains nulls ([#16626](https://github.com/NVIDIA/cudf/pull/16626)) [@davidwendt](https://github.com/davidwendt) +- Fix function parameters with common dependency modified during their evaluation ([#16620](https://github.com/NVIDIA/cudf/pull/16620)) [@ttnghia](https://github.com/ttnghia) +- bug-fix: Don't enable the CUDA language if testing was requested when finding cudf ([#16615](https://github.com/NVIDIA/cudf/pull/16615)) [@cryos](https://github.com/cryos) +- bug-fix: cudf/io/json.hpp use after move ([#16609](https://github.com/NVIDIA/cudf/pull/16609)) [@NicolasDenoyelle](https://github.com/NicolasDenoyelle) +- Remove CUDA whole compilation ODR violations ([#16603](https://github.com/NVIDIA/cudf/pull/16603)) [@robertmaynard](https://github.com/robertmaynard) +- MAINT: Adapt to numpy hiding flagsobject away ([#16593](https://github.com/NVIDIA/cudf/pull/16593)) [@seberg](https://github.com/seberg) +- Revert "Make proxy NumPy arrays pass isinstance check in `cudf.pandas`" ([#16586](https://github.com/NVIDIA/cudf/pull/16586)) [@Matt711](https://github.com/Matt711) +- Switch python version to `3.10` in `cudf.pandas` pandas test scripts ([#16559](https://github.com/NVIDIA/cudf/pull/16559)) [@galipremsagar](https://github.com/galipremsagar) +- Hide all gtest symbols in cudftestutil ([#16546](https://github.com/NVIDIA/cudf/pull/16546)) [@robertmaynard](https://github.com/robertmaynard) +- Update the java code to properly deal with lists being returned as strings ([#16536](https://github.com/NVIDIA/cudf/pull/16536)) [@revans2](https://github.com/revans2) +- Register `read_parquet` and `read_csv` with dask-expr ([#16535](https://github.com/NVIDIA/cudf/pull/16535)) [@rjzamora](https://github.com/rjzamora) +- Change cudf::empty_like to not include offsets for empty strings columns ([#16529](https://github.com/NVIDIA/cudf/pull/16529)) [@davidwendt](https://github.com/davidwendt) +- Fix DataFrame reductions with median returning scalar instead of Series ([#16527](https://github.com/NVIDIA/cudf/pull/16527)) [@mroeschke](https://github.com/mroeschke) +- Allow DataFrame.sort_values(by=) to select an index level ([#16519](https://github.com/NVIDIA/cudf/pull/16519)) [@mroeschke](https://github.com/mroeschke) +- Fix `date_range(start, end, freq)` when end-start is divisible by freq ([#16516](https://github.com/NVIDIA/cudf/pull/16516)) [@mroeschke](https://github.com/mroeschke) +- Preserve array name in MultiIndex.from_arrays ([#16515](https://github.com/NVIDIA/cudf/pull/16515)) [@mroeschke](https://github.com/mroeschke) +- Disallow indexing by selecting duplicate labels ([#16514](https://github.com/NVIDIA/cudf/pull/16514)) [@mroeschke](https://github.com/mroeschke) +- Fix `.replace(Index, Index)` raising a TypeError ([#16513](https://github.com/NVIDIA/cudf/pull/16513)) [@mroeschke](https://github.com/mroeschke) +- Check index bounds in compact protocol reader. ([#16493](https://github.com/NVIDIA/cudf/pull/16493)) [@bdice](https://github.com/bdice) +- Fix build failures with GCC 13 ([#16488](https://github.com/NVIDIA/cudf/pull/16488)) [@PointKernel](https://github.com/PointKernel) +- Fix all-empty input column for strings split APIs ([#16466](https://github.com/NVIDIA/cudf/pull/16466)) [@davidwendt](https://github.com/davidwendt) +- Fix segmented-sort overlapped input/output indices ([#16463](https://github.com/NVIDIA/cudf/pull/16463)) [@davidwendt](https://github.com/davidwendt) +- Fix merge conflict for auto merge 16447 ([#16449](https://github.com/NVIDIA/cudf/pull/16449)) [@davidwendt](https://github.com/davidwendt) ## 📖 Documentation -- Fix links in Dask cuDF documentation ([#16929](https://github.com/rapidsai/cudf/pull/16929)) [@rjzamora](https://github.com/rjzamora) -- Improve aggregation documentation ([#16822](https://github.com/rapidsai/cudf/pull/16822)) [@PointKernel](https://github.com/PointKernel) -- Add best practices page to Dask cuDF docs ([#16821](https://github.com/rapidsai/cudf/pull/16821)) [@rjzamora](https://github.com/rjzamora) -- [DOC] Update Pylibcudf doc strings ([#16810](https://github.com/rapidsai/cudf/pull/16810)) [@Matt711](https://github.com/Matt711) -- Recommending `miniforge` for conda install ([#16782](https://github.com/rapidsai/cudf/pull/16782)) [@mmccarty](https://github.com/mmccarty) -- Add labeling pylibcudf doc pages ([#16779](https://github.com/rapidsai/cudf/pull/16779)) [@mroeschke](https://github.com/mroeschke) -- Migrate dask-cudf README improvements to dask-cudf sphinx docs ([#16765](https://github.com/rapidsai/cudf/pull/16765)) [@rjzamora](https://github.com/rjzamora) -- [DOC] Remove out of date section from cudf.pandas docs ([#16697](https://github.com/rapidsai/cudf/pull/16697)) [@Matt711](https://github.com/Matt711) -- Add performance tips to cudf.pandas FAQ. ([#16693](https://github.com/rapidsai/cudf/pull/16693)) [@bdice](https://github.com/bdice) -- Update documentation for Dask cuDF ([#16671](https://github.com/rapidsai/cudf/pull/16671)) [@rjzamora](https://github.com/rjzamora) -- Add missing pylibcudf strings docs ([#16471](https://github.com/rapidsai/cudf/pull/16471)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- DOC: Refresh pylibcudf guide ([#15856](https://github.com/rapidsai/cudf/pull/15856)) [@lithomas1](https://github.com/lithomas1) +- Fix links in Dask cuDF documentation ([#16929](https://github.com/NVIDIA/cudf/pull/16929)) [@rjzamora](https://github.com/rjzamora) +- Improve aggregation documentation ([#16822](https://github.com/NVIDIA/cudf/pull/16822)) [@PointKernel](https://github.com/PointKernel) +- Add best practices page to Dask cuDF docs ([#16821](https://github.com/NVIDIA/cudf/pull/16821)) [@rjzamora](https://github.com/rjzamora) +- [DOC] Update Pylibcudf doc strings ([#16810](https://github.com/NVIDIA/cudf/pull/16810)) [@Matt711](https://github.com/Matt711) +- Recommending `miniforge` for conda install ([#16782](https://github.com/NVIDIA/cudf/pull/16782)) [@mmccarty](https://github.com/mmccarty) +- Add labeling pylibcudf doc pages ([#16779](https://github.com/NVIDIA/cudf/pull/16779)) [@mroeschke](https://github.com/mroeschke) +- Migrate dask-cudf README improvements to dask-cudf sphinx docs ([#16765](https://github.com/NVIDIA/cudf/pull/16765)) [@rjzamora](https://github.com/rjzamora) +- [DOC] Remove out of date section from cudf.pandas docs ([#16697](https://github.com/NVIDIA/cudf/pull/16697)) [@Matt711](https://github.com/Matt711) +- Add performance tips to cudf.pandas FAQ. ([#16693](https://github.com/NVIDIA/cudf/pull/16693)) [@bdice](https://github.com/bdice) +- Update documentation for Dask cuDF ([#16671](https://github.com/NVIDIA/cudf/pull/16671)) [@rjzamora](https://github.com/rjzamora) +- Add missing pylibcudf strings docs ([#16471](https://github.com/NVIDIA/cudf/pull/16471)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- DOC: Refresh pylibcudf guide ([#15856](https://github.com/NVIDIA/cudf/pull/15856)) [@lithomas1](https://github.com/lithomas1) ## 🚀 New Features -- Build `cudf-polars` with `build.sh` ([#16898](https://github.com/rapidsai/cudf/pull/16898)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add polars to "all" dependency list. ([#16875](https://github.com/rapidsai/cudf/pull/16875)) [@bdice](https://github.com/bdice) -- nvCOMP GZIP integration ([#16770](https://github.com/rapidsai/cudf/pull/16770)) [@vuule](https://github.com/vuule) -- [FEA] Add support for `cudf.NamedAgg` ([#16744](https://github.com/rapidsai/cudf/pull/16744)) [@Matt711](https://github.com/Matt711) -- Add experimental `filesystem="arrow"` support in `dask_cudf.read_parquet` ([#16684](https://github.com/rapidsai/cudf/pull/16684)) [@rjzamora](https://github.com/rjzamora) -- Relax Arrow pin ([#16681](https://github.com/rapidsai/cudf/pull/16681)) [@vyasr](https://github.com/vyasr) -- Add libcudf wrappers around current_device_resource functions. ([#16679](https://github.com/rapidsai/cudf/pull/16679)) [@harrism](https://github.com/harrism) -- Move NDS-H examples into benchmarks ([#16663](https://github.com/rapidsai/cudf/pull/16663)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- [FEA] Add third-party library integration testing of cudf.pandas to cudf ([#16645](https://github.com/rapidsai/cudf/pull/16645)) [@Matt711](https://github.com/Matt711) -- Make isinstance check pass for proxy ndarrays ([#16601](https://github.com/rapidsai/cudf/pull/16601)) [@Matt711](https://github.com/Matt711) -- [FEA] Add an environment variable to fail on fallback in `cudf.pandas` ([#16562](https://github.com/rapidsai/cudf/pull/16562)) [@Matt711](https://github.com/Matt711) -- [FEA] Add support for `cudf.unique` ([#16554](https://github.com/rapidsai/cudf/pull/16554)) [@Matt711](https://github.com/Matt711) -- [FEA] Support named aggregations in `df.groupby().agg()` ([#16528](https://github.com/rapidsai/cudf/pull/16528)) [@Matt711](https://github.com/Matt711) -- Change IPv4 convert APIs to support UINT32 instead of INT64 ([#16489](https://github.com/rapidsai/cudf/pull/16489)) [@davidwendt](https://github.com/davidwendt) -- enable list to be forced as string in JSON reader. ([#16472](https://github.com/rapidsai/cudf/pull/16472)) [@karthikeyann](https://github.com/karthikeyann) -- Remove cuDF dependency from pylibcudf column from_device tests ([#16441](https://github.com/rapidsai/cudf/pull/16441)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Enable cudf.pandas REPL and -c command support ([#16428](https://github.com/rapidsai/cudf/pull/16428)) [@bdice](https://github.com/bdice) -- Setup pylibcudf package ([#16299](https://github.com/rapidsai/cudf/pull/16299)) [@lithomas1](https://github.com/lithomas1) -- Add a libcudf/thrust-based TPC-H derived datagen ([#16294](https://github.com/rapidsai/cudf/pull/16294)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Make proxy NumPy arrays pass isinstance check in `cudf.pandas` ([#16286](https://github.com/rapidsai/cudf/pull/16286)) [@Matt711](https://github.com/Matt711) -- Add skiprows and nrows to parquet reader ([#16214](https://github.com/rapidsai/cudf/pull/16214)) [@lithomas1](https://github.com/lithomas1) -- Upgrade to nvcomp 4.0.1 ([#16076](https://github.com/rapidsai/cudf/pull/16076)) [@vuule](https://github.com/vuule) -- Migrate ORC reader to pylibcudf ([#16042](https://github.com/rapidsai/cudf/pull/16042)) [@lithomas1](https://github.com/lithomas1) -- JSON reader validation of values ([#15968](https://github.com/rapidsai/cudf/pull/15968)) [@karthikeyann](https://github.com/karthikeyann) -- Implement exposed null mask APIs in pylibcudf ([#15908](https://github.com/rapidsai/cudf/pull/15908)) [@charlesbluca](https://github.com/charlesbluca) -- Word-based nvtext::minhash function ([#15368](https://github.com/rapidsai/cudf/pull/15368)) [@davidwendt](https://github.com/davidwendt) +- Build `cudf-polars` with `build.sh` ([#16898](https://github.com/NVIDIA/cudf/pull/16898)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add polars to "all" dependency list. ([#16875](https://github.com/NVIDIA/cudf/pull/16875)) [@bdice](https://github.com/bdice) +- nvCOMP GZIP integration ([#16770](https://github.com/NVIDIA/cudf/pull/16770)) [@vuule](https://github.com/vuule) +- [FEA] Add support for `cudf.NamedAgg` ([#16744](https://github.com/NVIDIA/cudf/pull/16744)) [@Matt711](https://github.com/Matt711) +- Add experimental `filesystem="arrow"` support in `dask_cudf.read_parquet` ([#16684](https://github.com/NVIDIA/cudf/pull/16684)) [@rjzamora](https://github.com/rjzamora) +- Relax Arrow pin ([#16681](https://github.com/NVIDIA/cudf/pull/16681)) [@vyasr](https://github.com/vyasr) +- Add libcudf wrappers around current_device_resource functions. ([#16679](https://github.com/NVIDIA/cudf/pull/16679)) [@harrism](https://github.com/harrism) +- Move NDS-H examples into benchmarks ([#16663](https://github.com/NVIDIA/cudf/pull/16663)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- [FEA] Add third-party library integration testing of cudf.pandas to cudf ([#16645](https://github.com/NVIDIA/cudf/pull/16645)) [@Matt711](https://github.com/Matt711) +- Make isinstance check pass for proxy ndarrays ([#16601](https://github.com/NVIDIA/cudf/pull/16601)) [@Matt711](https://github.com/Matt711) +- [FEA] Add an environment variable to fail on fallback in `cudf.pandas` ([#16562](https://github.com/NVIDIA/cudf/pull/16562)) [@Matt711](https://github.com/Matt711) +- [FEA] Add support for `cudf.unique` ([#16554](https://github.com/NVIDIA/cudf/pull/16554)) [@Matt711](https://github.com/Matt711) +- [FEA] Support named aggregations in `df.groupby().agg()` ([#16528](https://github.com/NVIDIA/cudf/pull/16528)) [@Matt711](https://github.com/Matt711) +- Change IPv4 convert APIs to support UINT32 instead of INT64 ([#16489](https://github.com/NVIDIA/cudf/pull/16489)) [@davidwendt](https://github.com/davidwendt) +- enable list to be forced as string in JSON reader. ([#16472](https://github.com/NVIDIA/cudf/pull/16472)) [@karthikeyann](https://github.com/karthikeyann) +- Remove cuDF dependency from pylibcudf column from_device tests ([#16441](https://github.com/NVIDIA/cudf/pull/16441)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Enable cudf.pandas REPL and -c command support ([#16428](https://github.com/NVIDIA/cudf/pull/16428)) [@bdice](https://github.com/bdice) +- Setup pylibcudf package ([#16299](https://github.com/NVIDIA/cudf/pull/16299)) [@lithomas1](https://github.com/lithomas1) +- Add a libcudf/thrust-based TPC-H derived datagen ([#16294](https://github.com/NVIDIA/cudf/pull/16294)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Make proxy NumPy arrays pass isinstance check in `cudf.pandas` ([#16286](https://github.com/NVIDIA/cudf/pull/16286)) [@Matt711](https://github.com/Matt711) +- Add skiprows and nrows to parquet reader ([#16214](https://github.com/NVIDIA/cudf/pull/16214)) [@lithomas1](https://github.com/lithomas1) +- Upgrade to nvcomp 4.0.1 ([#16076](https://github.com/NVIDIA/cudf/pull/16076)) [@vuule](https://github.com/vuule) +- Migrate ORC reader to pylibcudf ([#16042](https://github.com/NVIDIA/cudf/pull/16042)) [@lithomas1](https://github.com/lithomas1) +- JSON reader validation of values ([#15968](https://github.com/NVIDIA/cudf/pull/15968)) [@karthikeyann](https://github.com/karthikeyann) +- Implement exposed null mask APIs in pylibcudf ([#15908](https://github.com/NVIDIA/cudf/pull/15908)) [@charlesbluca](https://github.com/charlesbluca) +- Word-based nvtext::minhash function ([#15368](https://github.com/NVIDIA/cudf/pull/15368)) [@davidwendt](https://github.com/davidwendt) ## 🛠️ Improvements -- Make tests deterministic ([#16910](https://github.com/rapidsai/cudf/pull/16910)) [@galipremsagar](https://github.com/galipremsagar) -- Update update-version.sh to use packaging lib ([#16891](https://github.com/rapidsai/cudf/pull/16891)) [@AyodeAwe](https://github.com/AyodeAwe) -- Pin polars for 24.10 and update polars test suite xfail list ([#16886](https://github.com/rapidsai/cudf/pull/16886)) [@wence-](https://github.com/wence-) -- Add in support for setting delim when parsing JSON through java ([#16867) (#16880](https://github.com/rapidsai/cudf/pull/16867) (#16880)) [@revans2](https://github.com/revans2) -- Remove unnecessary flag from build.sh ([#16879](https://github.com/rapidsai/cudf/pull/16879)) [@vyasr](https://github.com/vyasr) -- Ignore numba warning specific to ARM runners ([#16872](https://github.com/rapidsai/cudf/pull/16872)) [@galipremsagar](https://github.com/galipremsagar) -- Display deltas for `cudf.pandas` test summary ([#16864](https://github.com/rapidsai/cudf/pull/16864)) [@galipremsagar](https://github.com/galipremsagar) -- Switch to using native `traceback` ([#16851](https://github.com/rapidsai/cudf/pull/16851)) [@galipremsagar](https://github.com/galipremsagar) -- JSON tree algorithm code reorg ([#16836](https://github.com/rapidsai/cudf/pull/16836)) [@karthikeyann](https://github.com/karthikeyann) -- Add string.repeats API to pylibcudf ([#16834](https://github.com/rapidsai/cudf/pull/16834)) [@mroeschke](https://github.com/mroeschke) -- Use CI workflow branch 'branch-24.10' again ([#16832](https://github.com/rapidsai/cudf/pull/16832)) [@jameslamb](https://github.com/jameslamb) -- Rename the NDS-H benchmark binaries ([#16831](https://github.com/rapidsai/cudf/pull/16831)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Add string.findall APIs to pylibcudf ([#16825](https://github.com/rapidsai/cudf/pull/16825)) [@mroeschke](https://github.com/mroeschke) -- Add string.extract APIs to pylibcudf ([#16823](https://github.com/rapidsai/cudf/pull/16823)) [@mroeschke](https://github.com/mroeschke) -- use get-pr-info from nv-gha-runners ([#16819](https://github.com/rapidsai/cudf/pull/16819)) [@AyodeAwe](https://github.com/AyodeAwe) -- Add string.contains APIs to pylibcudf ([#16814](https://github.com/rapidsai/cudf/pull/16814)) [@mroeschke](https://github.com/mroeschke) -- Forward-merge branch-24.08 to branch-24.10 ([#16813](https://github.com/rapidsai/cudf/pull/16813)) [@bdice](https://github.com/bdice) -- Add io_type axis with default `PINNED_BUFFER` to nvbench PQ multithreaded reader ([#16809](https://github.com/rapidsai/cudf/pull/16809)) [@mhaseeb123](https://github.com/mhaseeb123) -- Update fmt (to 11.0.2) and spdlog (to 1.14.1). ([#16806](https://github.com/rapidsai/cudf/pull/16806)) [@jameslamb](https://github.com/jameslamb) -- Add ability to set parquet row group max #rows and #bytes in java ([#16805](https://github.com/rapidsai/cudf/pull/16805)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Add in option for Java JSON APIs to do column pruning in CUDF ([#16796](https://github.com/rapidsai/cudf/pull/16796)) [@revans2](https://github.com/revans2) -- Support drop_first in get_dummies ([#16795](https://github.com/rapidsai/cudf/pull/16795)) [@mroeschke](https://github.com/mroeschke) -- Exposed stream-ordering to join API ([#16793](https://github.com/rapidsai/cudf/pull/16793)) [@lamarrr](https://github.com/lamarrr) -- Add string.attributes APIs to pylibcudf ([#16785](https://github.com/rapidsai/cudf/pull/16785)) [@mroeschke](https://github.com/mroeschke) -- Java: Make ColumnVector.fromViewWithContiguousAllocation public ([#16784](https://github.com/rapidsai/cudf/pull/16784)) [@jlowe](https://github.com/jlowe) -- Add partitioning APIs to pylibcudf ([#16781](https://github.com/rapidsai/cudf/pull/16781)) [@mroeschke](https://github.com/mroeschke) -- Optimization of tdigest merge aggregation. ([#16780](https://github.com/rapidsai/cudf/pull/16780)) [@nvdbaranec](https://github.com/nvdbaranec) -- use libkvikio wheels in wheel builds ([#16778](https://github.com/rapidsai/cudf/pull/16778)) [@jameslamb](https://github.com/jameslamb) -- Exposed stream-ordering to datetime API ([#16774](https://github.com/rapidsai/cudf/pull/16774)) [@lamarrr](https://github.com/lamarrr) -- Add io/timezone APIs to pylibcudf ([#16771](https://github.com/rapidsai/cudf/pull/16771)) [@mroeschke](https://github.com/mroeschke) -- Remove `MultiIndex._poplevel` inplace implementation. ([#16767](https://github.com/rapidsai/cudf/pull/16767)) [@mroeschke](https://github.com/mroeschke) -- allow pandas patch version to float in cudf-pandas unit tests ([#16763](https://github.com/rapidsai/cudf/pull/16763)) [@jameslamb](https://github.com/jameslamb) -- Simplify the nvCOMP adapter ([#16762](https://github.com/rapidsai/cudf/pull/16762)) [@vuule](https://github.com/vuule) -- Add labeling APIs to pylibcudf ([#16761](https://github.com/rapidsai/cudf/pull/16761)) [@mroeschke](https://github.com/mroeschke) -- Add transform APIs to pylibcudf ([#16760](https://github.com/rapidsai/cudf/pull/16760)) [@mroeschke](https://github.com/mroeschke) -- Add a benchmark to study Parquet reader's performance for wide tables ([#16751](https://github.com/rapidsai/cudf/pull/16751)) [@mhaseeb123](https://github.com/mhaseeb123) -- Change the Parquet writer's `default_row_group_size_bytes` from 128MB to inf ([#16750](https://github.com/rapidsai/cudf/pull/16750)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add transpose API to pylibcudf ([#16749](https://github.com/rapidsai/cudf/pull/16749)) [@mroeschke](https://github.com/mroeschke) -- Add support for Python 3.12, update Kafka dependencies to 2.5.x ([#16745](https://github.com/rapidsai/cudf/pull/16745)) [@jameslamb](https://github.com/jameslamb) -- Generate GPU vs CPU usage metrics per pytest file in pandas testsuite for `cudf.pandas` ([#16739](https://github.com/rapidsai/cudf/pull/16739)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor cudf pandas integration tests CI ([#16728](https://github.com/rapidsai/cudf/pull/16728)) [@Matt711](https://github.com/Matt711) -- Remove ERROR_TEST gtest from libcudf ([#16722](https://github.com/rapidsai/cudf/pull/16722)) [@davidwendt](https://github.com/davidwendt) -- Use Series._from_column more consistently to avoid validation ([#16716](https://github.com/rapidsai/cudf/pull/16716)) [@mroeschke](https://github.com/mroeschke) -- remove some unnecessary libcudf nightly builds ([#16714](https://github.com/rapidsai/cudf/pull/16714)) [@jameslamb](https://github.com/jameslamb) -- Remove xfail from torch-cudf.pandas integration test ([#16705](https://github.com/rapidsai/cudf/pull/16705)) [@Matt711](https://github.com/Matt711) -- Add return type annotations to MultiIndex ([#16696](https://github.com/rapidsai/cudf/pull/16696)) [@mroeschke](https://github.com/mroeschke) -- Add type annotations to Index classes, utilize _from_column more ([#16695](https://github.com/rapidsai/cudf/pull/16695)) [@mroeschke](https://github.com/mroeschke) -- Have interval_range use IntervalIndex.from_breaks, remove column_empty_same_mask ([#16694](https://github.com/rapidsai/cudf/pull/16694)) [@mroeschke](https://github.com/mroeschke) -- Increase timeouts for couple of tests ([#16692](https://github.com/rapidsai/cudf/pull/16692)) [@galipremsagar](https://github.com/galipremsagar) -- Replace raw device_memory_resource pointer in pylibcudf Cython ([#16674](https://github.com/rapidsai/cudf/pull/16674)) [@harrism](https://github.com/harrism) -- switch from typing.Callable to collections.abc.Callable ([#16670](https://github.com/rapidsai/cudf/pull/16670)) [@jameslamb](https://github.com/jameslamb) -- Update rapidsai/pre-commit-hooks ([#16669](https://github.com/rapidsai/cudf/pull/16669)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Multi-file and Parquet-aware prefetching from remote storage ([#16657](https://github.com/rapidsai/cudf/pull/16657)) [@rjzamora](https://github.com/rjzamora) -- Access Frame attributes instead of ColumnAccessor attributes when available ([#16652](https://github.com/rapidsai/cudf/pull/16652)) [@mroeschke](https://github.com/mroeschke) -- Use non-mangled type names in nvbench output ([#16649](https://github.com/rapidsai/cudf/pull/16649)) [@davidwendt](https://github.com/davidwendt) -- Add pylibcudf build dir in build.sh for `clean` ([#16648](https://github.com/rapidsai/cudf/pull/16648)) [@galipremsagar](https://github.com/galipremsagar) -- Prune workflows based on changed files ([#16642](https://github.com/rapidsai/cudf/pull/16642)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Remove arrow dependency ([#16640](https://github.com/rapidsai/cudf/pull/16640)) [@vyasr](https://github.com/vyasr) -- Support reading multiple PQ sources with mismatching nullability for columns ([#16639](https://github.com/rapidsai/cudf/pull/16639)) [@mhaseeb123](https://github.com/mhaseeb123) -- Drop Python 3.9 support ([#16637](https://github.com/rapidsai/cudf/pull/16637)) [@jameslamb](https://github.com/jameslamb) -- Support DecimalDtype meta in dask_cudf ([#16634](https://github.com/rapidsai/cudf/pull/16634)) [@mroeschke](https://github.com/mroeschke) -- Add `num_multiprocessors` utility ([#16628](https://github.com/rapidsai/cudf/pull/16628)) [@PointKernel](https://github.com/PointKernel) -- Annotate `ColumnAccessor._data` labels as `Hashable` ([#16623](https://github.com/rapidsai/cudf/pull/16623)) [@mroeschke](https://github.com/mroeschke) -- Remove build_categorical_column in favor of CategoricalColumn constructor ([#16617](https://github.com/rapidsai/cudf/pull/16617)) [@mroeschke](https://github.com/mroeschke) -- Move apply_boolean_mask benchmark to nvbench ([#16616](https://github.com/rapidsai/cudf/pull/16616)) [@davidwendt](https://github.com/davidwendt) -- Revise `get_reader_filepath_or_buffer` to handle a list of data sources ([#16613](https://github.com/rapidsai/cudf/pull/16613)) [@rjzamora](https://github.com/rjzamora) -- do not install cudf in cudf_polars wheel tests ([#16612](https://github.com/rapidsai/cudf/pull/16612)) [@jameslamb](https://github.com/jameslamb) -- remove streamz git dependency, standardize build dependency names, consolidate some dependency lists ([#16611](https://github.com/rapidsai/cudf/pull/16611)) [@jameslamb](https://github.com/jameslamb) -- Fix C++ and Cython io types ([#16610](https://github.com/rapidsai/cudf/pull/16610)) [@vyasr](https://github.com/vyasr) -- Remove arrow_io_source ([#16607](https://github.com/rapidsai/cudf/pull/16607)) [@vyasr](https://github.com/vyasr) -- Remove thrust::optional from expression evaluator ([#16604](https://github.com/rapidsai/cudf/pull/16604)) [@bdice](https://github.com/bdice) -- Add stricter typing and validation to ColumnAccessor ([#16602](https://github.com/rapidsai/cudf/pull/16602)) [@mroeschke](https://github.com/mroeschke) -- make more use of YAML anchors in dependencies.yaml ([#16597](https://github.com/rapidsai/cudf/pull/16597)) [@jameslamb](https://github.com/jameslamb) -- Enable testing `cudf.pandas` unit tests for all minor versions of pandas ([#16595](https://github.com/rapidsai/cudf/pull/16595)) [@galipremsagar](https://github.com/galipremsagar) -- Extend the Parquet writer's dictionary encoding benchmark. ([#16591](https://github.com/rapidsai/cudf/pull/16591)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove legacy Arrow interop APIs ([#16590](https://github.com/rapidsai/cudf/pull/16590)) [@vyasr](https://github.com/vyasr) -- Remove NativeFile support from cudf Python ([#16589](https://github.com/rapidsai/cudf/pull/16589)) [@vyasr](https://github.com/vyasr) -- Add build job for pylibcudf ([#16587](https://github.com/rapidsai/cudf/pull/16587)) [@vyasr](https://github.com/vyasr) -- Add `public` qualifier for some member functions in Java class `Schema` ([#16583](https://github.com/rapidsai/cudf/pull/16583)) [@ttnghia](https://github.com/ttnghia) -- Enable gtests previously disabled for compute-sanitizer bug ([#16581](https://github.com/rapidsai/cudf/pull/16581)) [@davidwendt](https://github.com/davidwendt) -- [FEA] Add filesystem argument to `cudf.read_parquet` ([#16577](https://github.com/rapidsai/cudf/pull/16577)) [@rjzamora](https://github.com/rjzamora) -- Ensure size is always passed to NumericalColumn ([#16576](https://github.com/rapidsai/cudf/pull/16576)) [@mroeschke](https://github.com/mroeschke) -- standardize and consolidate wheel installations in testing scripts ([#16575](https://github.com/rapidsai/cudf/pull/16575)) [@jameslamb](https://github.com/jameslamb) -- Performance improvement for strings::slice for wide strings ([#16574](https://github.com/rapidsai/cudf/pull/16574)) [@davidwendt](https://github.com/davidwendt) -- Add `ToCudfBackend` expression to dask-cudf ([#16573](https://github.com/rapidsai/cudf/pull/16573)) [@rjzamora](https://github.com/rjzamora) -- CI: Test against old versions of key dependencies ([#16570](https://github.com/rapidsai/cudf/pull/16570)) [@seberg](https://github.com/seberg) -- Replace `NativeFile` dependency in dask-cudf Parquet reader ([#16569](https://github.com/rapidsai/cudf/pull/16569)) [@rjzamora](https://github.com/rjzamora) -- Align public utility function signatures with pandas 2.x ([#16565](https://github.com/rapidsai/cudf/pull/16565)) [@mroeschke](https://github.com/mroeschke) -- Move libcudf reduction google-benchmarks to nvbench ([#16564](https://github.com/rapidsai/cudf/pull/16564)) [@davidwendt](https://github.com/davidwendt) -- Rework strings::slice benchmark to use nvbench ([#16563](https://github.com/rapidsai/cudf/pull/16563)) [@davidwendt](https://github.com/davidwendt) -- Reenable arrow tests ([#16556](https://github.com/rapidsai/cudf/pull/16556)) [@vyasr](https://github.com/vyasr) -- Clean up reshaping ops ([#16553](https://github.com/rapidsai/cudf/pull/16553)) [@mroeschke](https://github.com/mroeschke) -- Disallow cudf.Index accepting column in favor of ._from_column ([#16549](https://github.com/rapidsai/cudf/pull/16549)) [@mroeschke](https://github.com/mroeschke) -- Rewrite remaining Python Arrow interop conversions using the C Data Interface ([#16548](https://github.com/rapidsai/cudf/pull/16548)) [@vyasr](https://github.com/vyasr) -- [REVIEW] JSON host tree algorithms ([#16545](https://github.com/rapidsai/cudf/pull/16545)) [@shrshi](https://github.com/shrshi) -- Refactor dictionary encoding in PQ writer to migrate to the new `cuco::static_map` ([#16541](https://github.com/rapidsai/cudf/pull/16541)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove hardcoded versions from workflows. ([#16540](https://github.com/rapidsai/cudf/pull/16540)) [@bdice](https://github.com/bdice) -- Ensure comparisons with pyints and integer series always succeed ([#16532](https://github.com/rapidsai/cudf/pull/16532)) [@seberg](https://github.com/seberg) -- Remove unneeded output size parameter from internal count_matches utility ([#16531](https://github.com/rapidsai/cudf/pull/16531)) [@davidwendt](https://github.com/davidwendt) -- Remove invalid column_view usage in string-scalar-to-column function ([#16530](https://github.com/rapidsai/cudf/pull/16530)) [@davidwendt](https://github.com/davidwendt) -- Raise NotImplementedError for Series.rename that's not a scalar ([#16525](https://github.com/rapidsai/cudf/pull/16525)) [@mroeschke](https://github.com/mroeschke) -- Remove deprecated public APIs from libcudf ([#16524](https://github.com/rapidsai/cudf/pull/16524)) [@davidwendt](https://github.com/davidwendt) -- Return Interval object in pandas compat mode for IntervalIndex reductions ([#16523](https://github.com/rapidsai/cudf/pull/16523)) [@mroeschke](https://github.com/mroeschke) -- Update json normalization to take device_buffer ([#16520](https://github.com/rapidsai/cudf/pull/16520)) [@karthikeyann](https://github.com/karthikeyann) -- Rework cudf::io::text::byte_range_info class member functions ([#16518](https://github.com/rapidsai/cudf/pull/16518)) [@davidwendt](https://github.com/davidwendt) -- Remove unneeded pair-iterator benchmark ([#16511](https://github.com/rapidsai/cudf/pull/16511)) [@davidwendt](https://github.com/davidwendt) -- Update pre-commit hooks ([#16510](https://github.com/rapidsai/cudf/pull/16510)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Improve update-version.sh ([#16506](https://github.com/rapidsai/cudf/pull/16506)) [@bdice](https://github.com/bdice) -- Use tool.scikit-build.cmake.version, set scikit-build-core minimum-version ([#16503](https://github.com/rapidsai/cudf/pull/16503)) [@jameslamb](https://github.com/jameslamb) -- Pass batch size to JSON reader using environment variable ([#16502](https://github.com/rapidsai/cudf/pull/16502)) [@shrshi](https://github.com/shrshi) -- Remove a deprecated multibyte_split API ([#16501](https://github.com/rapidsai/cudf/pull/16501)) [@davidwendt](https://github.com/davidwendt) -- Add interop example for `arrow::StringViewArray` to `cudf::column` ([#16498](https://github.com/rapidsai/cudf/pull/16498)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Add keep option to distinct nvbench ([#16497](https://github.com/rapidsai/cudf/pull/16497)) [@bdice](https://github.com/bdice) -- Use more idomatic cudf APIs in dask_cudf meta generation ([#16487](https://github.com/rapidsai/cudf/pull/16487)) [@mroeschke](https://github.com/mroeschke) -- Fix typo in dispatch_row_equal. ([#16473](https://github.com/rapidsai/cudf/pull/16473)) [@bdice](https://github.com/bdice) -- Use explicit construction of column subclass instead of `build_column` when type is known ([#16470](https://github.com/rapidsai/cudf/pull/16470)) [@mroeschke](https://github.com/mroeschke) -- Move exception handler into pylibcudf from cudf ([#16468](https://github.com/rapidsai/cudf/pull/16468)) [@lithomas1](https://github.com/lithomas1) -- Make StructColumn.__init__ strict ([#16467](https://github.com/rapidsai/cudf/pull/16467)) [@mroeschke](https://github.com/mroeschke) -- Make ListColumn.__init__ strict ([#16465](https://github.com/rapidsai/cudf/pull/16465)) [@mroeschke](https://github.com/mroeschke) -- Make Timedelta/DatetimeColumn.__init__ strict ([#16464](https://github.com/rapidsai/cudf/pull/16464)) [@mroeschke](https://github.com/mroeschke) -- Make NumericalColumn.__init__ strict ([#16457](https://github.com/rapidsai/cudf/pull/16457)) [@mroeschke](https://github.com/mroeschke) -- Make CategoricalColumn.__init__ strict ([#16456](https://github.com/rapidsai/cudf/pull/16456)) [@mroeschke](https://github.com/mroeschke) -- Disallow cudf.Series to accept column in favor of `._from_column` ([#16454](https://github.com/rapidsai/cudf/pull/16454)) [@mroeschke](https://github.com/mroeschke) -- Expose `stream` param in transform APIs ([#16452](https://github.com/rapidsai/cudf/pull/16452)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Add upper bound pin for polars ([#16442](https://github.com/rapidsai/cudf/pull/16442)) [@wence-](https://github.com/wence-) -- Make (Indexed)Frame.__init__ require data (and index) ([#16430](https://github.com/rapidsai/cudf/pull/16430)) [@mroeschke](https://github.com/mroeschke) -- Add Java APIs to copy column data to host asynchronously ([#16429](https://github.com/rapidsai/cudf/pull/16429)) [@jlowe](https://github.com/jlowe) -- Update docs of the TPC-H derived examples ([#16423](https://github.com/rapidsai/cudf/pull/16423)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Use RMM adaptor constructors instead of factories. ([#16414](https://github.com/rapidsai/cudf/pull/16414)) [@bdice](https://github.com/bdice) -- Align ewm APIs with pandas 2.x ([#16413](https://github.com/rapidsai/cudf/pull/16413)) [@mroeschke](https://github.com/mroeschke) -- Remove checking for specific tests in memcheck script ([#16412](https://github.com/rapidsai/cudf/pull/16412)) [@davidwendt](https://github.com/davidwendt) -- Add stream parameter to reshape APIs ([#16410](https://github.com/rapidsai/cudf/pull/16410)) [@davidwendt](https://github.com/davidwendt) -- Align groupby APIs with pandas 2.x ([#16403](https://github.com/rapidsai/cudf/pull/16403)) [@mroeschke](https://github.com/mroeschke) -- Align misc DataFrame and MultiIndex methods with pandas 2.x ([#16402](https://github.com/rapidsai/cudf/pull/16402)) [@mroeschke](https://github.com/mroeschke) -- update some branch references in GitHub Actions configs ([#16397](https://github.com/rapidsai/cudf/pull/16397)) [@jameslamb](https://github.com/jameslamb) -- Support reading matching projected and filter cols from Parquet files with otherwise mismatched schemas ([#16394](https://github.com/rapidsai/cudf/pull/16394)) [@mhaseeb123](https://github.com/mhaseeb123) -- Merge branch-24.08 into branch-24.10 ([#16393](https://github.com/rapidsai/cudf/pull/16393)) [@jameslamb](https://github.com/jameslamb) -- Add query 10 to the TPC-H suite ([#16392](https://github.com/rapidsai/cudf/pull/16392)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Use `make_host_vector` instead of `make_std_vector` to facilitate pinned memory optimizations ([#16386](https://github.com/rapidsai/cudf/pull/16386)) [@vuule](https://github.com/vuule) -- Fix some issues with deprecated / removed cccl facilities ([#16377](https://github.com/rapidsai/cudf/pull/16377)) [@miscco](https://github.com/miscco) -- Align IntervalIndex APIs with pandas 2.x ([#16371](https://github.com/rapidsai/cudf/pull/16371)) [@mroeschke](https://github.com/mroeschke) -- Align CategoricalIndex APIs with pandas 2.x ([#16369](https://github.com/rapidsai/cudf/pull/16369)) [@mroeschke](https://github.com/mroeschke) -- Align TimedeltaIndex APIs with pandas 2.x ([#16368](https://github.com/rapidsai/cudf/pull/16368)) [@mroeschke](https://github.com/mroeschke) -- Align DatetimeIndex APIs with pandas 2.x ([#16367](https://github.com/rapidsai/cudf/pull/16367)) [@mroeschke](https://github.com/mroeschke) -- fix [tool.setuptools] reference in custreamz config ([#16365](https://github.com/rapidsai/cudf/pull/16365)) [@jameslamb](https://github.com/jameslamb) -- Align Index APIs with pandas 2.x ([#16361](https://github.com/rapidsai/cudf/pull/16361)) [@mroeschke](https://github.com/mroeschke) -- Rebuild for & Support NumPy 2 ([#16300](https://github.com/rapidsai/cudf/pull/16300)) [@jakirkham](https://github.com/jakirkham) -- Add `stream` param to stream compaction APIs ([#16295](https://github.com/rapidsai/cudf/pull/16295)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Added batch memset to memset data and validity buffers in parquet reader ([#16281](https://github.com/rapidsai/cudf/pull/16281)) [@sdrp713](https://github.com/sdrp713) -- Deduplicate decimal32/decimal64 to decimal128 conversion function ([#16236](https://github.com/rapidsai/cudf/pull/16236)) [@mhaseeb123](https://github.com/mhaseeb123) -- Refactor mixed_semi_join using cuco::static_set ([#16230](https://github.com/rapidsai/cudf/pull/16230)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Improve performance of hash_character_ngrams using warp-per-string kernel ([#16212](https://github.com/rapidsai/cudf/pull/16212)) [@davidwendt](https://github.com/davidwendt) -- Add environment variable to log cudf.pandas fallback calls ([#16161](https://github.com/rapidsai/cudf/pull/16161)) [@mroeschke](https://github.com/mroeschke) -- Add libcudf example with large strings ([#15983](https://github.com/rapidsai/cudf/pull/15983)) [@davidwendt](https://github.com/davidwendt) -- JSON tree algorithms refactor I: CSR data structure for column tree ([#15979](https://github.com/rapidsai/cudf/pull/15979)) [@shrshi](https://github.com/shrshi) -- Support multiple new-line characters in regex APIs ([#15961](https://github.com/rapidsai/cudf/pull/15961)) [@davidwendt](https://github.com/davidwendt) -- adding wheel build for libcudf ([#15483](https://github.com/rapidsai/cudf/pull/15483)) [@msarahan](https://github.com/msarahan) -- Replace usages of `thrust::optional` with `std::optional` ([#15091](https://github.com/rapidsai/cudf/pull/15091)) [@miscco](https://github.com/miscco) +- Make tests deterministic ([#16910](https://github.com/NVIDIA/cudf/pull/16910)) [@galipremsagar](https://github.com/galipremsagar) +- Update update-version.sh to use packaging lib ([#16891](https://github.com/NVIDIA/cudf/pull/16891)) [@AyodeAwe](https://github.com/AyodeAwe) +- Pin polars for 24.10 and update polars test suite xfail list ([#16886](https://github.com/NVIDIA/cudf/pull/16886)) [@wence-](https://github.com/wence-) +- Add in support for setting delim when parsing JSON through java ([#16867) (#16880](https://github.com/NVIDIA/cudf/pull/16867) (#16880)) [@revans2](https://github.com/revans2) +- Remove unnecessary flag from build.sh ([#16879](https://github.com/NVIDIA/cudf/pull/16879)) [@vyasr](https://github.com/vyasr) +- Ignore numba warning specific to ARM runners ([#16872](https://github.com/NVIDIA/cudf/pull/16872)) [@galipremsagar](https://github.com/galipremsagar) +- Display deltas for `cudf.pandas` test summary ([#16864](https://github.com/NVIDIA/cudf/pull/16864)) [@galipremsagar](https://github.com/galipremsagar) +- Switch to using native `traceback` ([#16851](https://github.com/NVIDIA/cudf/pull/16851)) [@galipremsagar](https://github.com/galipremsagar) +- JSON tree algorithm code reorg ([#16836](https://github.com/NVIDIA/cudf/pull/16836)) [@karthikeyann](https://github.com/karthikeyann) +- Add string.repeats API to pylibcudf ([#16834](https://github.com/NVIDIA/cudf/pull/16834)) [@mroeschke](https://github.com/mroeschke) +- Use CI workflow branch 'branch-24.10' again ([#16832](https://github.com/NVIDIA/cudf/pull/16832)) [@jameslamb](https://github.com/jameslamb) +- Rename the NDS-H benchmark binaries ([#16831](https://github.com/NVIDIA/cudf/pull/16831)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Add string.findall APIs to pylibcudf ([#16825](https://github.com/NVIDIA/cudf/pull/16825)) [@mroeschke](https://github.com/mroeschke) +- Add string.extract APIs to pylibcudf ([#16823](https://github.com/NVIDIA/cudf/pull/16823)) [@mroeschke](https://github.com/mroeschke) +- use get-pr-info from nv-gha-runners ([#16819](https://github.com/NVIDIA/cudf/pull/16819)) [@AyodeAwe](https://github.com/AyodeAwe) +- Add string.contains APIs to pylibcudf ([#16814](https://github.com/NVIDIA/cudf/pull/16814)) [@mroeschke](https://github.com/mroeschke) +- Forward-merge branch-24.08 to branch-24.10 ([#16813](https://github.com/NVIDIA/cudf/pull/16813)) [@bdice](https://github.com/bdice) +- Add io_type axis with default `PINNED_BUFFER` to nvbench PQ multithreaded reader ([#16809](https://github.com/NVIDIA/cudf/pull/16809)) [@mhaseeb123](https://github.com/mhaseeb123) +- Update fmt (to 11.0.2) and spdlog (to 1.14.1). ([#16806](https://github.com/NVIDIA/cudf/pull/16806)) [@jameslamb](https://github.com/jameslamb) +- Add ability to set parquet row group max #rows and #bytes in java ([#16805](https://github.com/NVIDIA/cudf/pull/16805)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Add in option for Java JSON APIs to do column pruning in CUDF ([#16796](https://github.com/NVIDIA/cudf/pull/16796)) [@revans2](https://github.com/revans2) +- Support drop_first in get_dummies ([#16795](https://github.com/NVIDIA/cudf/pull/16795)) [@mroeschke](https://github.com/mroeschke) +- Exposed stream-ordering to join API ([#16793](https://github.com/NVIDIA/cudf/pull/16793)) [@lamarrr](https://github.com/lamarrr) +- Add string.attributes APIs to pylibcudf ([#16785](https://github.com/NVIDIA/cudf/pull/16785)) [@mroeschke](https://github.com/mroeschke) +- Java: Make ColumnVector.fromViewWithContiguousAllocation public ([#16784](https://github.com/NVIDIA/cudf/pull/16784)) [@jlowe](https://github.com/jlowe) +- Add partitioning APIs to pylibcudf ([#16781](https://github.com/NVIDIA/cudf/pull/16781)) [@mroeschke](https://github.com/mroeschke) +- Optimization of tdigest merge aggregation. ([#16780](https://github.com/NVIDIA/cudf/pull/16780)) [@nvdbaranec](https://github.com/nvdbaranec) +- use libkvikio wheels in wheel builds ([#16778](https://github.com/NVIDIA/cudf/pull/16778)) [@jameslamb](https://github.com/jameslamb) +- Exposed stream-ordering to datetime API ([#16774](https://github.com/NVIDIA/cudf/pull/16774)) [@lamarrr](https://github.com/lamarrr) +- Add io/timezone APIs to pylibcudf ([#16771](https://github.com/NVIDIA/cudf/pull/16771)) [@mroeschke](https://github.com/mroeschke) +- Remove `MultiIndex._poplevel` inplace implementation. ([#16767](https://github.com/NVIDIA/cudf/pull/16767)) [@mroeschke](https://github.com/mroeschke) +- allow pandas patch version to float in cudf-pandas unit tests ([#16763](https://github.com/NVIDIA/cudf/pull/16763)) [@jameslamb](https://github.com/jameslamb) +- Simplify the nvCOMP adapter ([#16762](https://github.com/NVIDIA/cudf/pull/16762)) [@vuule](https://github.com/vuule) +- Add labeling APIs to pylibcudf ([#16761](https://github.com/NVIDIA/cudf/pull/16761)) [@mroeschke](https://github.com/mroeschke) +- Add transform APIs to pylibcudf ([#16760](https://github.com/NVIDIA/cudf/pull/16760)) [@mroeschke](https://github.com/mroeschke) +- Add a benchmark to study Parquet reader's performance for wide tables ([#16751](https://github.com/NVIDIA/cudf/pull/16751)) [@mhaseeb123](https://github.com/mhaseeb123) +- Change the Parquet writer's `default_row_group_size_bytes` from 128MB to inf ([#16750](https://github.com/NVIDIA/cudf/pull/16750)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add transpose API to pylibcudf ([#16749](https://github.com/NVIDIA/cudf/pull/16749)) [@mroeschke](https://github.com/mroeschke) +- Add support for Python 3.12, update Kafka dependencies to 2.5.x ([#16745](https://github.com/NVIDIA/cudf/pull/16745)) [@jameslamb](https://github.com/jameslamb) +- Generate GPU vs CPU usage metrics per pytest file in pandas testsuite for `cudf.pandas` ([#16739](https://github.com/NVIDIA/cudf/pull/16739)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor cudf pandas integration tests CI ([#16728](https://github.com/NVIDIA/cudf/pull/16728)) [@Matt711](https://github.com/Matt711) +- Remove ERROR_TEST gtest from libcudf ([#16722](https://github.com/NVIDIA/cudf/pull/16722)) [@davidwendt](https://github.com/davidwendt) +- Use Series._from_column more consistently to avoid validation ([#16716](https://github.com/NVIDIA/cudf/pull/16716)) [@mroeschke](https://github.com/mroeschke) +- remove some unnecessary libcudf nightly builds ([#16714](https://github.com/NVIDIA/cudf/pull/16714)) [@jameslamb](https://github.com/jameslamb) +- Remove xfail from torch-cudf.pandas integration test ([#16705](https://github.com/NVIDIA/cudf/pull/16705)) [@Matt711](https://github.com/Matt711) +- Add return type annotations to MultiIndex ([#16696](https://github.com/NVIDIA/cudf/pull/16696)) [@mroeschke](https://github.com/mroeschke) +- Add type annotations to Index classes, utilize _from_column more ([#16695](https://github.com/NVIDIA/cudf/pull/16695)) [@mroeschke](https://github.com/mroeschke) +- Have interval_range use IntervalIndex.from_breaks, remove column_empty_same_mask ([#16694](https://github.com/NVIDIA/cudf/pull/16694)) [@mroeschke](https://github.com/mroeschke) +- Increase timeouts for couple of tests ([#16692](https://github.com/NVIDIA/cudf/pull/16692)) [@galipremsagar](https://github.com/galipremsagar) +- Replace raw device_memory_resource pointer in pylibcudf Cython ([#16674](https://github.com/NVIDIA/cudf/pull/16674)) [@harrism](https://github.com/harrism) +- switch from typing.Callable to collections.abc.Callable ([#16670](https://github.com/NVIDIA/cudf/pull/16670)) [@jameslamb](https://github.com/jameslamb) +- Update rapidsai/pre-commit-hooks ([#16669](https://github.com/NVIDIA/cudf/pull/16669)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Multi-file and Parquet-aware prefetching from remote storage ([#16657](https://github.com/NVIDIA/cudf/pull/16657)) [@rjzamora](https://github.com/rjzamora) +- Access Frame attributes instead of ColumnAccessor attributes when available ([#16652](https://github.com/NVIDIA/cudf/pull/16652)) [@mroeschke](https://github.com/mroeschke) +- Use non-mangled type names in nvbench output ([#16649](https://github.com/NVIDIA/cudf/pull/16649)) [@davidwendt](https://github.com/davidwendt) +- Add pylibcudf build dir in build.sh for `clean` ([#16648](https://github.com/NVIDIA/cudf/pull/16648)) [@galipremsagar](https://github.com/galipremsagar) +- Prune workflows based on changed files ([#16642](https://github.com/NVIDIA/cudf/pull/16642)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Remove arrow dependency ([#16640](https://github.com/NVIDIA/cudf/pull/16640)) [@vyasr](https://github.com/vyasr) +- Support reading multiple PQ sources with mismatching nullability for columns ([#16639](https://github.com/NVIDIA/cudf/pull/16639)) [@mhaseeb123](https://github.com/mhaseeb123) +- Drop Python 3.9 support ([#16637](https://github.com/NVIDIA/cudf/pull/16637)) [@jameslamb](https://github.com/jameslamb) +- Support DecimalDtype meta in dask_cudf ([#16634](https://github.com/NVIDIA/cudf/pull/16634)) [@mroeschke](https://github.com/mroeschke) +- Add `num_multiprocessors` utility ([#16628](https://github.com/NVIDIA/cudf/pull/16628)) [@PointKernel](https://github.com/PointKernel) +- Annotate `ColumnAccessor._data` labels as `Hashable` ([#16623](https://github.com/NVIDIA/cudf/pull/16623)) [@mroeschke](https://github.com/mroeschke) +- Remove build_categorical_column in favor of CategoricalColumn constructor ([#16617](https://github.com/NVIDIA/cudf/pull/16617)) [@mroeschke](https://github.com/mroeschke) +- Move apply_boolean_mask benchmark to nvbench ([#16616](https://github.com/NVIDIA/cudf/pull/16616)) [@davidwendt](https://github.com/davidwendt) +- Revise `get_reader_filepath_or_buffer` to handle a list of data sources ([#16613](https://github.com/NVIDIA/cudf/pull/16613)) [@rjzamora](https://github.com/rjzamora) +- do not install cudf in cudf_polars wheel tests ([#16612](https://github.com/NVIDIA/cudf/pull/16612)) [@jameslamb](https://github.com/jameslamb) +- remove streamz git dependency, standardize build dependency names, consolidate some dependency lists ([#16611](https://github.com/NVIDIA/cudf/pull/16611)) [@jameslamb](https://github.com/jameslamb) +- Fix C++ and Cython io types ([#16610](https://github.com/NVIDIA/cudf/pull/16610)) [@vyasr](https://github.com/vyasr) +- Remove arrow_io_source ([#16607](https://github.com/NVIDIA/cudf/pull/16607)) [@vyasr](https://github.com/vyasr) +- Remove thrust::optional from expression evaluator ([#16604](https://github.com/NVIDIA/cudf/pull/16604)) [@bdice](https://github.com/bdice) +- Add stricter typing and validation to ColumnAccessor ([#16602](https://github.com/NVIDIA/cudf/pull/16602)) [@mroeschke](https://github.com/mroeschke) +- make more use of YAML anchors in dependencies.yaml ([#16597](https://github.com/NVIDIA/cudf/pull/16597)) [@jameslamb](https://github.com/jameslamb) +- Enable testing `cudf.pandas` unit tests for all minor versions of pandas ([#16595](https://github.com/NVIDIA/cudf/pull/16595)) [@galipremsagar](https://github.com/galipremsagar) +- Extend the Parquet writer's dictionary encoding benchmark. ([#16591](https://github.com/NVIDIA/cudf/pull/16591)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove legacy Arrow interop APIs ([#16590](https://github.com/NVIDIA/cudf/pull/16590)) [@vyasr](https://github.com/vyasr) +- Remove NativeFile support from cudf Python ([#16589](https://github.com/NVIDIA/cudf/pull/16589)) [@vyasr](https://github.com/vyasr) +- Add build job for pylibcudf ([#16587](https://github.com/NVIDIA/cudf/pull/16587)) [@vyasr](https://github.com/vyasr) +- Add `public` qualifier for some member functions in Java class `Schema` ([#16583](https://github.com/NVIDIA/cudf/pull/16583)) [@ttnghia](https://github.com/ttnghia) +- Enable gtests previously disabled for compute-sanitizer bug ([#16581](https://github.com/NVIDIA/cudf/pull/16581)) [@davidwendt](https://github.com/davidwendt) +- [FEA] Add filesystem argument to `cudf.read_parquet` ([#16577](https://github.com/NVIDIA/cudf/pull/16577)) [@rjzamora](https://github.com/rjzamora) +- Ensure size is always passed to NumericalColumn ([#16576](https://github.com/NVIDIA/cudf/pull/16576)) [@mroeschke](https://github.com/mroeschke) +- standardize and consolidate wheel installations in testing scripts ([#16575](https://github.com/NVIDIA/cudf/pull/16575)) [@jameslamb](https://github.com/jameslamb) +- Performance improvement for strings::slice for wide strings ([#16574](https://github.com/NVIDIA/cudf/pull/16574)) [@davidwendt](https://github.com/davidwendt) +- Add `ToCudfBackend` expression to dask-cudf ([#16573](https://github.com/NVIDIA/cudf/pull/16573)) [@rjzamora](https://github.com/rjzamora) +- CI: Test against old versions of key dependencies ([#16570](https://github.com/NVIDIA/cudf/pull/16570)) [@seberg](https://github.com/seberg) +- Replace `NativeFile` dependency in dask-cudf Parquet reader ([#16569](https://github.com/NVIDIA/cudf/pull/16569)) [@rjzamora](https://github.com/rjzamora) +- Align public utility function signatures with pandas 2.x ([#16565](https://github.com/NVIDIA/cudf/pull/16565)) [@mroeschke](https://github.com/mroeschke) +- Move libcudf reduction google-benchmarks to nvbench ([#16564](https://github.com/NVIDIA/cudf/pull/16564)) [@davidwendt](https://github.com/davidwendt) +- Rework strings::slice benchmark to use nvbench ([#16563](https://github.com/NVIDIA/cudf/pull/16563)) [@davidwendt](https://github.com/davidwendt) +- Reenable arrow tests ([#16556](https://github.com/NVIDIA/cudf/pull/16556)) [@vyasr](https://github.com/vyasr) +- Clean up reshaping ops ([#16553](https://github.com/NVIDIA/cudf/pull/16553)) [@mroeschke](https://github.com/mroeschke) +- Disallow cudf.Index accepting column in favor of ._from_column ([#16549](https://github.com/NVIDIA/cudf/pull/16549)) [@mroeschke](https://github.com/mroeschke) +- Rewrite remaining Python Arrow interop conversions using the C Data Interface ([#16548](https://github.com/NVIDIA/cudf/pull/16548)) [@vyasr](https://github.com/vyasr) +- [REVIEW] JSON host tree algorithms ([#16545](https://github.com/NVIDIA/cudf/pull/16545)) [@shrshi](https://github.com/shrshi) +- Refactor dictionary encoding in PQ writer to migrate to the new `cuco::static_map` ([#16541](https://github.com/NVIDIA/cudf/pull/16541)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove hardcoded versions from workflows. ([#16540](https://github.com/NVIDIA/cudf/pull/16540)) [@bdice](https://github.com/bdice) +- Ensure comparisons with pyints and integer series always succeed ([#16532](https://github.com/NVIDIA/cudf/pull/16532)) [@seberg](https://github.com/seberg) +- Remove unneeded output size parameter from internal count_matches utility ([#16531](https://github.com/NVIDIA/cudf/pull/16531)) [@davidwendt](https://github.com/davidwendt) +- Remove invalid column_view usage in string-scalar-to-column function ([#16530](https://github.com/NVIDIA/cudf/pull/16530)) [@davidwendt](https://github.com/davidwendt) +- Raise NotImplementedError for Series.rename that's not a scalar ([#16525](https://github.com/NVIDIA/cudf/pull/16525)) [@mroeschke](https://github.com/mroeschke) +- Remove deprecated public APIs from libcudf ([#16524](https://github.com/NVIDIA/cudf/pull/16524)) [@davidwendt](https://github.com/davidwendt) +- Return Interval object in pandas compat mode for IntervalIndex reductions ([#16523](https://github.com/NVIDIA/cudf/pull/16523)) [@mroeschke](https://github.com/mroeschke) +- Update json normalization to take device_buffer ([#16520](https://github.com/NVIDIA/cudf/pull/16520)) [@karthikeyann](https://github.com/karthikeyann) +- Rework cudf::io::text::byte_range_info class member functions ([#16518](https://github.com/NVIDIA/cudf/pull/16518)) [@davidwendt](https://github.com/davidwendt) +- Remove unneeded pair-iterator benchmark ([#16511](https://github.com/NVIDIA/cudf/pull/16511)) [@davidwendt](https://github.com/davidwendt) +- Update pre-commit hooks ([#16510](https://github.com/NVIDIA/cudf/pull/16510)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Improve update-version.sh ([#16506](https://github.com/NVIDIA/cudf/pull/16506)) [@bdice](https://github.com/bdice) +- Use tool.scikit-build.cmake.version, set scikit-build-core minimum-version ([#16503](https://github.com/NVIDIA/cudf/pull/16503)) [@jameslamb](https://github.com/jameslamb) +- Pass batch size to JSON reader using environment variable ([#16502](https://github.com/NVIDIA/cudf/pull/16502)) [@shrshi](https://github.com/shrshi) +- Remove a deprecated multibyte_split API ([#16501](https://github.com/NVIDIA/cudf/pull/16501)) [@davidwendt](https://github.com/davidwendt) +- Add interop example for `arrow::StringViewArray` to `cudf::column` ([#16498](https://github.com/NVIDIA/cudf/pull/16498)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Add keep option to distinct nvbench ([#16497](https://github.com/NVIDIA/cudf/pull/16497)) [@bdice](https://github.com/bdice) +- Use more idomatic cudf APIs in dask_cudf meta generation ([#16487](https://github.com/NVIDIA/cudf/pull/16487)) [@mroeschke](https://github.com/mroeschke) +- Fix typo in dispatch_row_equal. ([#16473](https://github.com/NVIDIA/cudf/pull/16473)) [@bdice](https://github.com/bdice) +- Use explicit construction of column subclass instead of `build_column` when type is known ([#16470](https://github.com/NVIDIA/cudf/pull/16470)) [@mroeschke](https://github.com/mroeschke) +- Move exception handler into pylibcudf from cudf ([#16468](https://github.com/NVIDIA/cudf/pull/16468)) [@lithomas1](https://github.com/lithomas1) +- Make StructColumn.__init__ strict ([#16467](https://github.com/NVIDIA/cudf/pull/16467)) [@mroeschke](https://github.com/mroeschke) +- Make ListColumn.__init__ strict ([#16465](https://github.com/NVIDIA/cudf/pull/16465)) [@mroeschke](https://github.com/mroeschke) +- Make Timedelta/DatetimeColumn.__init__ strict ([#16464](https://github.com/NVIDIA/cudf/pull/16464)) [@mroeschke](https://github.com/mroeschke) +- Make NumericalColumn.__init__ strict ([#16457](https://github.com/NVIDIA/cudf/pull/16457)) [@mroeschke](https://github.com/mroeschke) +- Make CategoricalColumn.__init__ strict ([#16456](https://github.com/NVIDIA/cudf/pull/16456)) [@mroeschke](https://github.com/mroeschke) +- Disallow cudf.Series to accept column in favor of `._from_column` ([#16454](https://github.com/NVIDIA/cudf/pull/16454)) [@mroeschke](https://github.com/mroeschke) +- Expose `stream` param in transform APIs ([#16452](https://github.com/NVIDIA/cudf/pull/16452)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Add upper bound pin for polars ([#16442](https://github.com/NVIDIA/cudf/pull/16442)) [@wence-](https://github.com/wence-) +- Make (Indexed)Frame.__init__ require data (and index) ([#16430](https://github.com/NVIDIA/cudf/pull/16430)) [@mroeschke](https://github.com/mroeschke) +- Add Java APIs to copy column data to host asynchronously ([#16429](https://github.com/NVIDIA/cudf/pull/16429)) [@jlowe](https://github.com/jlowe) +- Update docs of the TPC-H derived examples ([#16423](https://github.com/NVIDIA/cudf/pull/16423)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Use RMM adaptor constructors instead of factories. ([#16414](https://github.com/NVIDIA/cudf/pull/16414)) [@bdice](https://github.com/bdice) +- Align ewm APIs with pandas 2.x ([#16413](https://github.com/NVIDIA/cudf/pull/16413)) [@mroeschke](https://github.com/mroeschke) +- Remove checking for specific tests in memcheck script ([#16412](https://github.com/NVIDIA/cudf/pull/16412)) [@davidwendt](https://github.com/davidwendt) +- Add stream parameter to reshape APIs ([#16410](https://github.com/NVIDIA/cudf/pull/16410)) [@davidwendt](https://github.com/davidwendt) +- Align groupby APIs with pandas 2.x ([#16403](https://github.com/NVIDIA/cudf/pull/16403)) [@mroeschke](https://github.com/mroeschke) +- Align misc DataFrame and MultiIndex methods with pandas 2.x ([#16402](https://github.com/NVIDIA/cudf/pull/16402)) [@mroeschke](https://github.com/mroeschke) +- update some branch references in GitHub Actions configs ([#16397](https://github.com/NVIDIA/cudf/pull/16397)) [@jameslamb](https://github.com/jameslamb) +- Support reading matching projected and filter cols from Parquet files with otherwise mismatched schemas ([#16394](https://github.com/NVIDIA/cudf/pull/16394)) [@mhaseeb123](https://github.com/mhaseeb123) +- Merge branch-24.08 into branch-24.10 ([#16393](https://github.com/NVIDIA/cudf/pull/16393)) [@jameslamb](https://github.com/jameslamb) +- Add query 10 to the TPC-H suite ([#16392](https://github.com/NVIDIA/cudf/pull/16392)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Use `make_host_vector` instead of `make_std_vector` to facilitate pinned memory optimizations ([#16386](https://github.com/NVIDIA/cudf/pull/16386)) [@vuule](https://github.com/vuule) +- Fix some issues with deprecated / removed cccl facilities ([#16377](https://github.com/NVIDIA/cudf/pull/16377)) [@miscco](https://github.com/miscco) +- Align IntervalIndex APIs with pandas 2.x ([#16371](https://github.com/NVIDIA/cudf/pull/16371)) [@mroeschke](https://github.com/mroeschke) +- Align CategoricalIndex APIs with pandas 2.x ([#16369](https://github.com/NVIDIA/cudf/pull/16369)) [@mroeschke](https://github.com/mroeschke) +- Align TimedeltaIndex APIs with pandas 2.x ([#16368](https://github.com/NVIDIA/cudf/pull/16368)) [@mroeschke](https://github.com/mroeschke) +- Align DatetimeIndex APIs with pandas 2.x ([#16367](https://github.com/NVIDIA/cudf/pull/16367)) [@mroeschke](https://github.com/mroeschke) +- fix [tool.setuptools] reference in custreamz config ([#16365](https://github.com/NVIDIA/cudf/pull/16365)) [@jameslamb](https://github.com/jameslamb) +- Align Index APIs with pandas 2.x ([#16361](https://github.com/NVIDIA/cudf/pull/16361)) [@mroeschke](https://github.com/mroeschke) +- Rebuild for & Support NumPy 2 ([#16300](https://github.com/NVIDIA/cudf/pull/16300)) [@jakirkham](https://github.com/jakirkham) +- Add `stream` param to stream compaction APIs ([#16295](https://github.com/NVIDIA/cudf/pull/16295)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Added batch memset to memset data and validity buffers in parquet reader ([#16281](https://github.com/NVIDIA/cudf/pull/16281)) [@sdrp713](https://github.com/sdrp713) +- Deduplicate decimal32/decimal64 to decimal128 conversion function ([#16236](https://github.com/NVIDIA/cudf/pull/16236)) [@mhaseeb123](https://github.com/mhaseeb123) +- Refactor mixed_semi_join using cuco::static_set ([#16230](https://github.com/NVIDIA/cudf/pull/16230)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Improve performance of hash_character_ngrams using warp-per-string kernel ([#16212](https://github.com/NVIDIA/cudf/pull/16212)) [@davidwendt](https://github.com/davidwendt) +- Add environment variable to log cudf.pandas fallback calls ([#16161](https://github.com/NVIDIA/cudf/pull/16161)) [@mroeschke](https://github.com/mroeschke) +- Add libcudf example with large strings ([#15983](https://github.com/NVIDIA/cudf/pull/15983)) [@davidwendt](https://github.com/davidwendt) +- JSON tree algorithms refactor I: CSR data structure for column tree ([#15979](https://github.com/NVIDIA/cudf/pull/15979)) [@shrshi](https://github.com/shrshi) +- Support multiple new-line characters in regex APIs ([#15961](https://github.com/NVIDIA/cudf/pull/15961)) [@davidwendt](https://github.com/davidwendt) +- adding wheel build for libcudf ([#15483](https://github.com/NVIDIA/cudf/pull/15483)) [@msarahan](https://github.com/msarahan) +- Replace usages of `thrust::optional` with `std::optional` ([#15091](https://github.com/NVIDIA/cudf/pull/15091)) [@miscco](https://github.com/miscco) # cudf 24.08.00 (7 Aug 2024) ## 🚨 Breaking Changes -- Align Index __init__ APIs with pandas 2.x ([#16362](https://github.com/rapidsai/cudf/pull/16362)) [@mroeschke](https://github.com/mroeschke) -- Align Series APIs with pandas 2.x ([#16333](https://github.com/rapidsai/cudf/pull/16333)) [@mroeschke](https://github.com/mroeschke) -- Add missing `stream` param to dictionary factory APIs ([#16319](https://github.com/rapidsai/cudf/pull/16319)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Deprecate dtype= parameter in reduction methods ([#16313](https://github.com/rapidsai/cudf/pull/16313)) [@mroeschke](https://github.com/mroeschke) -- Remove squeeze argument from groupby ([#16312](https://github.com/rapidsai/cudf/pull/16312)) [@mroeschke](https://github.com/mroeschke) -- Align more DataFrame APIs with pandas ([#16310](https://github.com/rapidsai/cudf/pull/16310)) [@mroeschke](https://github.com/mroeschke) -- Remove `mr` param from `write_csv` and `write_json` ([#16231](https://github.com/rapidsai/cudf/pull/16231)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Report number of rows per file read by PQ reader when no row selection and fix segfault in chunked PQ reader when skip_rows > 0 ([#16195](https://github.com/rapidsai/cudf/pull/16195)) [@mhaseeb123](https://github.com/mhaseeb123) -- Refactor from_arrow_device/host to use resource_ref ([#16160](https://github.com/rapidsai/cudf/pull/16160)) [@harrism](https://github.com/harrism) -- Deprecate Arrow support in I/O ([#16132](https://github.com/rapidsai/cudf/pull/16132)) [@lithomas1](https://github.com/lithomas1) -- Return `FrozenList` for `Index.names` ([#16047](https://github.com/rapidsai/cudf/pull/16047)) [@galipremsagar](https://github.com/galipremsagar) -- Add compile option to enable large strings support ([#16037](https://github.com/rapidsai/cudf/pull/16037)) [@davidwendt](https://github.com/davidwendt) -- Hide visibility of non public symbols ([#15982](https://github.com/rapidsai/cudf/pull/15982)) [@robertmaynard](https://github.com/robertmaynard) -- Rename strings multiple target replace API ([#15898](https://github.com/rapidsai/cudf/pull/15898)) [@davidwendt](https://github.com/davidwendt) -- Pinned vector factory that uses the global pool ([#15895](https://github.com/rapidsai/cudf/pull/15895)) [@vuule](https://github.com/vuule) -- Apply clang-tidy autofixes ([#15894](https://github.com/rapidsai/cudf/pull/15894)) [@vyasr](https://github.com/vyasr) -- Support `arrow:schema` in Parquet writer to faithfully roundtrip `duration` types with Arrow ([#15875](https://github.com/rapidsai/cudf/pull/15875)) [@mhaseeb123](https://github.com/mhaseeb123) -- Expose stream parameter to public rolling APIs ([#15865](https://github.com/rapidsai/cudf/pull/15865)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Fix large strings handling in nvtext::character_tokenize ([#15829](https://github.com/rapidsai/cudf/pull/15829)) [@davidwendt](https://github.com/davidwendt) -- Remove legacy JSON reader and concurrent_unordered_map.cuh. ([#15813](https://github.com/rapidsai/cudf/pull/15813)) [@bdice](https://github.com/bdice) +- Align Index __init__ APIs with pandas 2.x ([#16362](https://github.com/NVIDIA/cudf/pull/16362)) [@mroeschke](https://github.com/mroeschke) +- Align Series APIs with pandas 2.x ([#16333](https://github.com/NVIDIA/cudf/pull/16333)) [@mroeschke](https://github.com/mroeschke) +- Add missing `stream` param to dictionary factory APIs ([#16319](https://github.com/NVIDIA/cudf/pull/16319)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Deprecate dtype= parameter in reduction methods ([#16313](https://github.com/NVIDIA/cudf/pull/16313)) [@mroeschke](https://github.com/mroeschke) +- Remove squeeze argument from groupby ([#16312](https://github.com/NVIDIA/cudf/pull/16312)) [@mroeschke](https://github.com/mroeschke) +- Align more DataFrame APIs with pandas ([#16310](https://github.com/NVIDIA/cudf/pull/16310)) [@mroeschke](https://github.com/mroeschke) +- Remove `mr` param from `write_csv` and `write_json` ([#16231](https://github.com/NVIDIA/cudf/pull/16231)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Report number of rows per file read by PQ reader when no row selection and fix segfault in chunked PQ reader when skip_rows > 0 ([#16195](https://github.com/NVIDIA/cudf/pull/16195)) [@mhaseeb123](https://github.com/mhaseeb123) +- Refactor from_arrow_device/host to use resource_ref ([#16160](https://github.com/NVIDIA/cudf/pull/16160)) [@harrism](https://github.com/harrism) +- Deprecate Arrow support in I/O ([#16132](https://github.com/NVIDIA/cudf/pull/16132)) [@lithomas1](https://github.com/lithomas1) +- Return `FrozenList` for `Index.names` ([#16047](https://github.com/NVIDIA/cudf/pull/16047)) [@galipremsagar](https://github.com/galipremsagar) +- Add compile option to enable large strings support ([#16037](https://github.com/NVIDIA/cudf/pull/16037)) [@davidwendt](https://github.com/davidwendt) +- Hide visibility of non public symbols ([#15982](https://github.com/NVIDIA/cudf/pull/15982)) [@robertmaynard](https://github.com/robertmaynard) +- Rename strings multiple target replace API ([#15898](https://github.com/NVIDIA/cudf/pull/15898)) [@davidwendt](https://github.com/davidwendt) +- Pinned vector factory that uses the global pool ([#15895](https://github.com/NVIDIA/cudf/pull/15895)) [@vuule](https://github.com/vuule) +- Apply clang-tidy autofixes ([#15894](https://github.com/NVIDIA/cudf/pull/15894)) [@vyasr](https://github.com/vyasr) +- Support `arrow:schema` in Parquet writer to faithfully roundtrip `duration` types with Arrow ([#15875](https://github.com/NVIDIA/cudf/pull/15875)) [@mhaseeb123](https://github.com/mhaseeb123) +- Expose stream parameter to public rolling APIs ([#15865](https://github.com/NVIDIA/cudf/pull/15865)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Fix large strings handling in nvtext::character_tokenize ([#15829](https://github.com/NVIDIA/cudf/pull/15829)) [@davidwendt](https://github.com/davidwendt) +- Remove legacy JSON reader and concurrent_unordered_map.cuh. ([#15813](https://github.com/NVIDIA/cudf/pull/15813)) [@bdice](https://github.com/bdice) ## 🐛 Bug Fixes -- Add `flatbuffers` to `libcudf` build ([#16446](https://github.com/rapidsai/cudf/pull/16446)) [@galipremsagar](https://github.com/galipremsagar) -- Fix parquet_field_list read_func lambda capture invalid this pointer ([#16440](https://github.com/rapidsai/cudf/pull/16440)) [@davidwendt](https://github.com/davidwendt) -- Enable prefetching in cudf.pandas.install() ([#16439](https://github.com/rapidsai/cudf/pull/16439)) [@bdice](https://github.com/bdice) -- Enable prefetching before `runpy` ([#16427](https://github.com/rapidsai/cudf/pull/16427)) [@galipremsagar](https://github.com/galipremsagar) -- Support thread-safe for `prefetch_config::get` and `prefetch_config::set` ([#16425](https://github.com/rapidsai/cudf/pull/16425)) [@ttnghia](https://github.com/ttnghia) -- Fix a `pandas-2.0` missing attribute error ([#16416](https://github.com/rapidsai/cudf/pull/16416)) [@galipremsagar](https://github.com/galipremsagar) -- [Bug] Remove loud `NativeFile` deprecation noise for `read_parquet` from S3 ([#16415](https://github.com/rapidsai/cudf/pull/16415)) [@rjzamora](https://github.com/rjzamora) -- Fix nightly memcheck error for empty STREAM_INTEROP_TEST ([#16406](https://github.com/rapidsai/cudf/pull/16406)) [@davidwendt](https://github.com/davidwendt) -- Gate ArrowStringArrayNumpySemantics cudf.pandas proxy behind version check ([#16401](https://github.com/rapidsai/cudf/pull/16401)) [@mroeschke](https://github.com/mroeschke) -- Don't export bs_thread_pool ([#16398](https://github.com/rapidsai/cudf/pull/16398)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Require fixed width types for casting in `cudf-polars` ([#16381](https://github.com/rapidsai/cudf/pull/16381)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix docstring of `DataFrame.apply` ([#16351](https://github.com/rapidsai/cudf/pull/16351)) [@galipremsagar](https://github.com/galipremsagar) -- Make __bool__ raise for more cudf objects ([#16311](https://github.com/rapidsai/cudf/pull/16311)) [@mroeschke](https://github.com/mroeschke) -- Rename `.devcontainer`s for CUDA 12.5 ([#16293](https://github.com/rapidsai/cudf/pull/16293)) [@jakirkham](https://github.com/jakirkham) -- Fix split_record for all empty strings column ([#16291](https://github.com/rapidsai/cudf/pull/16291)) [@davidwendt](https://github.com/davidwendt) -- Fix logic in to_arrow for empty list column ([#16279](https://github.com/rapidsai/cudf/pull/16279)) [@wence-](https://github.com/wence-) -- [BUG] Make name attr of Index fast slow attrs ([#16270](https://github.com/rapidsai/cudf/pull/16270)) [@Matt711](https://github.com/Matt711) -- Add custom name setter and getter for proxy objects in `cudf.pandas` ([#16234](https://github.com/rapidsai/cudf/pull/16234)) [@Matt711](https://github.com/Matt711) -- Fall back when casting a timestamp to numeric in cudf-polars ([#16232](https://github.com/rapidsai/cudf/pull/16232)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Disable large string support for Java build ([#16216](https://github.com/rapidsai/cudf/pull/16216)) [@jlowe](https://github.com/jlowe) -- Remove CCCL patch for PR 211. ([#16207](https://github.com/rapidsai/cudf/pull/16207)) [@bdice](https://github.com/bdice) -- Add single offset to an empty ListArray in cudf::to_arrow ([#16201](https://github.com/rapidsai/cudf/pull/16201)) [@davidwendt](https://github.com/davidwendt) -- Fix `memory_usage` when calculating nested list column ([#16193](https://github.com/rapidsai/cudf/pull/16193)) [@mroeschke](https://github.com/mroeschke) -- Support at/iat indexers in cudf.pandas ([#16177](https://github.com/rapidsai/cudf/pull/16177)) [@mroeschke](https://github.com/mroeschke) -- Fix unused-return-value debug build error in from_arrow_stream_test.cpp ([#16168](https://github.com/rapidsai/cudf/pull/16168)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf::strings::replace_multiple hang on empty target ([#16167](https://github.com/rapidsai/cudf/pull/16167)) [@davidwendt](https://github.com/davidwendt) -- Refactor from_arrow_device/host to use resource_ref ([#16160](https://github.com/rapidsai/cudf/pull/16160)) [@harrism](https://github.com/harrism) -- interpolate returns new column if no values are interpolated ([#16158](https://github.com/rapidsai/cudf/pull/16158)) [@mroeschke](https://github.com/mroeschke) -- Use provided memory resource for allocating mixed join results. ([#16153](https://github.com/rapidsai/cudf/pull/16153)) [@bdice](https://github.com/bdice) -- Run DFG after verify-alpha-spec ([#16151](https://github.com/rapidsai/cudf/pull/16151)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Use size_t to allow large conditional joins ([#16127](https://github.com/rapidsai/cudf/pull/16127)) [@bdice](https://github.com/bdice) -- Allow only scale=0 fixed-point values in fixed_width_column_wrapper ([#16120](https://github.com/rapidsai/cudf/pull/16120)) [@davidwendt](https://github.com/davidwendt) -- Fix pylibcudf Table.num_rows for 0 columns case and add interop to docs ([#16108](https://github.com/rapidsai/cudf/pull/16108)) [@lithomas1](https://github.com/lithomas1) -- Add support for proxy `np.flatiter` objects ([#16107](https://github.com/rapidsai/cudf/pull/16107)) [@Matt711](https://github.com/Matt711) -- Ensure cudf objects can astype to any type when empty ([#16106](https://github.com/rapidsai/cudf/pull/16106)) [@mroeschke](https://github.com/mroeschke) -- Support `pd.read_pickle` and `pd.to_pickle` in `cudf.pandas` ([#16105](https://github.com/rapidsai/cudf/pull/16105)) [@Matt711](https://github.com/Matt711) -- Fix unnecessarily strict check in parquet chunked reader for choosing split locations. ([#16099](https://github.com/rapidsai/cudf/pull/16099)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix `is_monotonic_*` APIs to include `nan's` ([#16085](https://github.com/rapidsai/cudf/pull/16085)) [@galipremsagar](https://github.com/galipremsagar) -- More safely parse CUDA versions when subprocess output is contaminated ([#16067](https://github.com/rapidsai/cudf/pull/16067)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- fast_slow_proxy: Don't import assert_eq at top-level ([#16063](https://github.com/rapidsai/cudf/pull/16063)) [@wence-](https://github.com/wence-) -- Prevent bad ColumnAccessor state after .sort_index(axis=1, ignore_index=True) ([#16061](https://github.com/rapidsai/cudf/pull/16061)) [@mroeschke](https://github.com/mroeschke) -- Fix ArrowDeviceArray interface to pass address of event ([#16058](https://github.com/rapidsai/cudf/pull/16058)) [@zeroshade](https://github.com/zeroshade) -- Fix a size overflow bug in hash groupby ([#16053](https://github.com/rapidsai/cudf/pull/16053)) [@PointKernel](https://github.com/PointKernel) -- Fix `atomic_ref` scope when multiple blocks are updating the same output ([#16051](https://github.com/rapidsai/cudf/pull/16051)) [@vuule](https://github.com/vuule) -- Fix initialization error in to_arrow for empty string views ([#16033](https://github.com/rapidsai/cudf/pull/16033)) [@wence-](https://github.com/wence-) -- Fix the int32 overflow when computing page fragment sizes for large string columns ([#16028](https://github.com/rapidsai/cudf/pull/16028)) [@mhaseeb123](https://github.com/mhaseeb123) -- Fix the pool size alignment issue ([#16024](https://github.com/rapidsai/cudf/pull/16024)) [@PointKernel](https://github.com/PointKernel) -- Improve multibyte-split byte-range performance ([#16019](https://github.com/rapidsai/cudf/pull/16019)) [@davidwendt](https://github.com/davidwendt) -- Fix target counting in strings char-parallel replace ([#16017](https://github.com/rapidsai/cudf/pull/16017)) [@davidwendt](https://github.com/davidwendt) -- Support IntervalDtype in cudf.from_pandas ([#16014](https://github.com/rapidsai/cudf/pull/16014)) [@mroeschke](https://github.com/mroeschke) -- Fix memory size in create_byte_range_infos_consecutive ([#16012](https://github.com/rapidsai/cudf/pull/16012)) [@davidwendt](https://github.com/davidwendt) -- Hide visibility of non public symbols ([#15982](https://github.com/rapidsai/cudf/pull/15982)) [@robertmaynard](https://github.com/robertmaynard) -- Fix Cython typo preventing proper inheritance ([#15978](https://github.com/rapidsai/cudf/pull/15978)) [@vyasr](https://github.com/vyasr) -- Fix convert_dtypes with convert_integer=False/convert_floating=True ([#15964](https://github.com/rapidsai/cudf/pull/15964)) [@mroeschke](https://github.com/mroeschke) -- Fix nunique for `MultiIndex`, `DataFrame`, and all NA case with `dropna=False` ([#15962](https://github.com/rapidsai/cudf/pull/15962)) [@mroeschke](https://github.com/mroeschke) -- Explicitly build for all GPU architectures ([#15959](https://github.com/rapidsai/cudf/pull/15959)) [@vyasr](https://github.com/vyasr) -- Preserve column type and class information in more DataFrame operations ([#15949](https://github.com/rapidsai/cudf/pull/15949)) [@mroeschke](https://github.com/mroeschke) -- Add __array_interface__ to cudf.pandas numpy.ndarray proxy ([#15936](https://github.com/rapidsai/cudf/pull/15936)) [@mroeschke](https://github.com/mroeschke) -- Allow tests to be built when stream util is disabled ([#15933](https://github.com/rapidsai/cudf/pull/15933)) [@robertmaynard](https://github.com/robertmaynard) -- Fix JSON multi-source reading when total source size exceeds `INT_MAX` bytes ([#15930](https://github.com/rapidsai/cudf/pull/15930)) [@shrshi](https://github.com/shrshi) -- Fix `dask_cudf.read_parquet` regression for legacy timestamp data ([#15929](https://github.com/rapidsai/cudf/pull/15929)) [@rjzamora](https://github.com/rjzamora) -- Fix offsetalator when accessing over 268 million rows ([#15921](https://github.com/rapidsai/cudf/pull/15921)) [@davidwendt](https://github.com/davidwendt) -- Fix debug assert in rowgroup_char_counts_kernel ([#15902](https://github.com/rapidsai/cudf/pull/15902)) [@davidwendt](https://github.com/davidwendt) -- Fix categorical conversion from chunked arrow arrays ([#15886](https://github.com/rapidsai/cudf/pull/15886)) [@vyasr](https://github.com/vyasr) -- Handling for `NaN` and `inf` when converting floating point to fixed point types ([#15885](https://github.com/rapidsai/cudf/pull/15885)) [@ttnghia](https://github.com/ttnghia) -- Manual merge of Branch 24.08 from 24.06 ([#15869](https://github.com/rapidsai/cudf/pull/15869)) [@galipremsagar](https://github.com/galipremsagar) -- Avoid unnecessary `Index` cast in `IndexedFrame.index` setter ([#15843](https://github.com/rapidsai/cudf/pull/15843)) [@charlesbluca](https://github.com/charlesbluca) -- Fix large strings handling in nvtext::character_tokenize ([#15829](https://github.com/rapidsai/cudf/pull/15829)) [@davidwendt](https://github.com/davidwendt) -- Fix multi-replace target count logic for large strings ([#15807](https://github.com/rapidsai/cudf/pull/15807)) [@davidwendt](https://github.com/davidwendt) -- Fix JSON parsing memory corruption - Fix Mixed types nested children removal ([#15798](https://github.com/rapidsai/cudf/pull/15798)) [@karthikeyann](https://github.com/karthikeyann) -- Allow anonymous user in devcontainer name. ([#15784](https://github.com/rapidsai/cudf/pull/15784)) [@bdice](https://github.com/bdice) -- Add support for additional metaclasses of proxies and use for ExcelWriter ([#15399](https://github.com/rapidsai/cudf/pull/15399)) [@vyasr](https://github.com/vyasr) +- Add `flatbuffers` to `libcudf` build ([#16446](https://github.com/NVIDIA/cudf/pull/16446)) [@galipremsagar](https://github.com/galipremsagar) +- Fix parquet_field_list read_func lambda capture invalid this pointer ([#16440](https://github.com/NVIDIA/cudf/pull/16440)) [@davidwendt](https://github.com/davidwendt) +- Enable prefetching in cudf.pandas.install() ([#16439](https://github.com/NVIDIA/cudf/pull/16439)) [@bdice](https://github.com/bdice) +- Enable prefetching before `runpy` ([#16427](https://github.com/NVIDIA/cudf/pull/16427)) [@galipremsagar](https://github.com/galipremsagar) +- Support thread-safe for `prefetch_config::get` and `prefetch_config::set` ([#16425](https://github.com/NVIDIA/cudf/pull/16425)) [@ttnghia](https://github.com/ttnghia) +- Fix a `pandas-2.0` missing attribute error ([#16416](https://github.com/NVIDIA/cudf/pull/16416)) [@galipremsagar](https://github.com/galipremsagar) +- [Bug] Remove loud `NativeFile` deprecation noise for `read_parquet` from S3 ([#16415](https://github.com/NVIDIA/cudf/pull/16415)) [@rjzamora](https://github.com/rjzamora) +- Fix nightly memcheck error for empty STREAM_INTEROP_TEST ([#16406](https://github.com/NVIDIA/cudf/pull/16406)) [@davidwendt](https://github.com/davidwendt) +- Gate ArrowStringArrayNumpySemantics cudf.pandas proxy behind version check ([#16401](https://github.com/NVIDIA/cudf/pull/16401)) [@mroeschke](https://github.com/mroeschke) +- Don't export bs_thread_pool ([#16398](https://github.com/NVIDIA/cudf/pull/16398)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Require fixed width types for casting in `cudf-polars` ([#16381](https://github.com/NVIDIA/cudf/pull/16381)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix docstring of `DataFrame.apply` ([#16351](https://github.com/NVIDIA/cudf/pull/16351)) [@galipremsagar](https://github.com/galipremsagar) +- Make __bool__ raise for more cudf objects ([#16311](https://github.com/NVIDIA/cudf/pull/16311)) [@mroeschke](https://github.com/mroeschke) +- Rename `.devcontainer`s for CUDA 12.5 ([#16293](https://github.com/NVIDIA/cudf/pull/16293)) [@jakirkham](https://github.com/jakirkham) +- Fix split_record for all empty strings column ([#16291](https://github.com/NVIDIA/cudf/pull/16291)) [@davidwendt](https://github.com/davidwendt) +- Fix logic in to_arrow for empty list column ([#16279](https://github.com/NVIDIA/cudf/pull/16279)) [@wence-](https://github.com/wence-) +- [BUG] Make name attr of Index fast slow attrs ([#16270](https://github.com/NVIDIA/cudf/pull/16270)) [@Matt711](https://github.com/Matt711) +- Add custom name setter and getter for proxy objects in `cudf.pandas` ([#16234](https://github.com/NVIDIA/cudf/pull/16234)) [@Matt711](https://github.com/Matt711) +- Fall back when casting a timestamp to numeric in cudf-polars ([#16232](https://github.com/NVIDIA/cudf/pull/16232)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Disable large string support for Java build ([#16216](https://github.com/NVIDIA/cudf/pull/16216)) [@jlowe](https://github.com/jlowe) +- Remove CCCL patch for PR 211. ([#16207](https://github.com/NVIDIA/cudf/pull/16207)) [@bdice](https://github.com/bdice) +- Add single offset to an empty ListArray in cudf::to_arrow ([#16201](https://github.com/NVIDIA/cudf/pull/16201)) [@davidwendt](https://github.com/davidwendt) +- Fix `memory_usage` when calculating nested list column ([#16193](https://github.com/NVIDIA/cudf/pull/16193)) [@mroeschke](https://github.com/mroeschke) +- Support at/iat indexers in cudf.pandas ([#16177](https://github.com/NVIDIA/cudf/pull/16177)) [@mroeschke](https://github.com/mroeschke) +- Fix unused-return-value debug build error in from_arrow_stream_test.cpp ([#16168](https://github.com/NVIDIA/cudf/pull/16168)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf::strings::replace_multiple hang on empty target ([#16167](https://github.com/NVIDIA/cudf/pull/16167)) [@davidwendt](https://github.com/davidwendt) +- Refactor from_arrow_device/host to use resource_ref ([#16160](https://github.com/NVIDIA/cudf/pull/16160)) [@harrism](https://github.com/harrism) +- interpolate returns new column if no values are interpolated ([#16158](https://github.com/NVIDIA/cudf/pull/16158)) [@mroeschke](https://github.com/mroeschke) +- Use provided memory resource for allocating mixed join results. ([#16153](https://github.com/NVIDIA/cudf/pull/16153)) [@bdice](https://github.com/bdice) +- Run DFG after verify-alpha-spec ([#16151](https://github.com/NVIDIA/cudf/pull/16151)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Use size_t to allow large conditional joins ([#16127](https://github.com/NVIDIA/cudf/pull/16127)) [@bdice](https://github.com/bdice) +- Allow only scale=0 fixed-point values in fixed_width_column_wrapper ([#16120](https://github.com/NVIDIA/cudf/pull/16120)) [@davidwendt](https://github.com/davidwendt) +- Fix pylibcudf Table.num_rows for 0 columns case and add interop to docs ([#16108](https://github.com/NVIDIA/cudf/pull/16108)) [@lithomas1](https://github.com/lithomas1) +- Add support for proxy `np.flatiter` objects ([#16107](https://github.com/NVIDIA/cudf/pull/16107)) [@Matt711](https://github.com/Matt711) +- Ensure cudf objects can astype to any type when empty ([#16106](https://github.com/NVIDIA/cudf/pull/16106)) [@mroeschke](https://github.com/mroeschke) +- Support `pd.read_pickle` and `pd.to_pickle` in `cudf.pandas` ([#16105](https://github.com/NVIDIA/cudf/pull/16105)) [@Matt711](https://github.com/Matt711) +- Fix unnecessarily strict check in parquet chunked reader for choosing split locations. ([#16099](https://github.com/NVIDIA/cudf/pull/16099)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix `is_monotonic_*` APIs to include `nan's` ([#16085](https://github.com/NVIDIA/cudf/pull/16085)) [@galipremsagar](https://github.com/galipremsagar) +- More safely parse CUDA versions when subprocess output is contaminated ([#16067](https://github.com/NVIDIA/cudf/pull/16067)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- fast_slow_proxy: Don't import assert_eq at top-level ([#16063](https://github.com/NVIDIA/cudf/pull/16063)) [@wence-](https://github.com/wence-) +- Prevent bad ColumnAccessor state after .sort_index(axis=1, ignore_index=True) ([#16061](https://github.com/NVIDIA/cudf/pull/16061)) [@mroeschke](https://github.com/mroeschke) +- Fix ArrowDeviceArray interface to pass address of event ([#16058](https://github.com/NVIDIA/cudf/pull/16058)) [@zeroshade](https://github.com/zeroshade) +- Fix a size overflow bug in hash groupby ([#16053](https://github.com/NVIDIA/cudf/pull/16053)) [@PointKernel](https://github.com/PointKernel) +- Fix `atomic_ref` scope when multiple blocks are updating the same output ([#16051](https://github.com/NVIDIA/cudf/pull/16051)) [@vuule](https://github.com/vuule) +- Fix initialization error in to_arrow for empty string views ([#16033](https://github.com/NVIDIA/cudf/pull/16033)) [@wence-](https://github.com/wence-) +- Fix the int32 overflow when computing page fragment sizes for large string columns ([#16028](https://github.com/NVIDIA/cudf/pull/16028)) [@mhaseeb123](https://github.com/mhaseeb123) +- Fix the pool size alignment issue ([#16024](https://github.com/NVIDIA/cudf/pull/16024)) [@PointKernel](https://github.com/PointKernel) +- Improve multibyte-split byte-range performance ([#16019](https://github.com/NVIDIA/cudf/pull/16019)) [@davidwendt](https://github.com/davidwendt) +- Fix target counting in strings char-parallel replace ([#16017](https://github.com/NVIDIA/cudf/pull/16017)) [@davidwendt](https://github.com/davidwendt) +- Support IntervalDtype in cudf.from_pandas ([#16014](https://github.com/NVIDIA/cudf/pull/16014)) [@mroeschke](https://github.com/mroeschke) +- Fix memory size in create_byte_range_infos_consecutive ([#16012](https://github.com/NVIDIA/cudf/pull/16012)) [@davidwendt](https://github.com/davidwendt) +- Hide visibility of non public symbols ([#15982](https://github.com/NVIDIA/cudf/pull/15982)) [@robertmaynard](https://github.com/robertmaynard) +- Fix Cython typo preventing proper inheritance ([#15978](https://github.com/NVIDIA/cudf/pull/15978)) [@vyasr](https://github.com/vyasr) +- Fix convert_dtypes with convert_integer=False/convert_floating=True ([#15964](https://github.com/NVIDIA/cudf/pull/15964)) [@mroeschke](https://github.com/mroeschke) +- Fix nunique for `MultiIndex`, `DataFrame`, and all NA case with `dropna=False` ([#15962](https://github.com/NVIDIA/cudf/pull/15962)) [@mroeschke](https://github.com/mroeschke) +- Explicitly build for all GPU architectures ([#15959](https://github.com/NVIDIA/cudf/pull/15959)) [@vyasr](https://github.com/vyasr) +- Preserve column type and class information in more DataFrame operations ([#15949](https://github.com/NVIDIA/cudf/pull/15949)) [@mroeschke](https://github.com/mroeschke) +- Add __array_interface__ to cudf.pandas numpy.ndarray proxy ([#15936](https://github.com/NVIDIA/cudf/pull/15936)) [@mroeschke](https://github.com/mroeschke) +- Allow tests to be built when stream util is disabled ([#15933](https://github.com/NVIDIA/cudf/pull/15933)) [@robertmaynard](https://github.com/robertmaynard) +- Fix JSON multi-source reading when total source size exceeds `INT_MAX` bytes ([#15930](https://github.com/NVIDIA/cudf/pull/15930)) [@shrshi](https://github.com/shrshi) +- Fix `dask_cudf.read_parquet` regression for legacy timestamp data ([#15929](https://github.com/NVIDIA/cudf/pull/15929)) [@rjzamora](https://github.com/rjzamora) +- Fix offsetalator when accessing over 268 million rows ([#15921](https://github.com/NVIDIA/cudf/pull/15921)) [@davidwendt](https://github.com/davidwendt) +- Fix debug assert in rowgroup_char_counts_kernel ([#15902](https://github.com/NVIDIA/cudf/pull/15902)) [@davidwendt](https://github.com/davidwendt) +- Fix categorical conversion from chunked arrow arrays ([#15886](https://github.com/NVIDIA/cudf/pull/15886)) [@vyasr](https://github.com/vyasr) +- Handling for `NaN` and `inf` when converting floating point to fixed point types ([#15885](https://github.com/NVIDIA/cudf/pull/15885)) [@ttnghia](https://github.com/ttnghia) +- Manual merge of Branch 24.08 from 24.06 ([#15869](https://github.com/NVIDIA/cudf/pull/15869)) [@galipremsagar](https://github.com/galipremsagar) +- Avoid unnecessary `Index` cast in `IndexedFrame.index` setter ([#15843](https://github.com/NVIDIA/cudf/pull/15843)) [@charlesbluca](https://github.com/charlesbluca) +- Fix large strings handling in nvtext::character_tokenize ([#15829](https://github.com/NVIDIA/cudf/pull/15829)) [@davidwendt](https://github.com/davidwendt) +- Fix multi-replace target count logic for large strings ([#15807](https://github.com/NVIDIA/cudf/pull/15807)) [@davidwendt](https://github.com/davidwendt) +- Fix JSON parsing memory corruption - Fix Mixed types nested children removal ([#15798](https://github.com/NVIDIA/cudf/pull/15798)) [@karthikeyann](https://github.com/karthikeyann) +- Allow anonymous user in devcontainer name. ([#15784](https://github.com/NVIDIA/cudf/pull/15784)) [@bdice](https://github.com/bdice) +- Add support for additional metaclasses of proxies and use for ExcelWriter ([#15399](https://github.com/NVIDIA/cudf/pull/15399)) [@vyasr](https://github.com/vyasr) ## 📖 Documentation -- Add docstring for from_dataframe ([#16260](https://github.com/rapidsai/cudf/pull/16260)) [@mroeschke](https://github.com/mroeschke) -- Update libcudf compiler requirements in contributing doc ([#16103](https://github.com/rapidsai/cudf/pull/16103)) [@davidwendt](https://github.com/davidwendt) -- Add libcudf public/detail API pattern to developer guide ([#16086](https://github.com/rapidsai/cudf/pull/16086)) [@davidwendt](https://github.com/davidwendt) -- Explain line profiler and how to know which functions are GPU-accelerated. ([#16079](https://github.com/rapidsai/cudf/pull/16079)) [@bdice](https://github.com/bdice) -- cudf.pandas documentation improvement ([#15948](https://github.com/rapidsai/cudf/pull/15948)) [@Matt711](https://github.com/Matt711) -- Reland "Fix docs for IO readers and strings_convert" ([#15872)" (#15941](https://github.com/rapidsai/cudf/pull/15872)" (#15941)) [@lithomas1](https://github.com/lithomas1) -- Document how to use cudf.pandas in tandem with multiprocessing ([#15940](https://github.com/rapidsai/cudf/pull/15940)) [@wence-](https://github.com/wence-) -- DOC: Add documentation for cudf.pandas in the Developer Guide ([#15889](https://github.com/rapidsai/cudf/pull/15889)) [@Matt711](https://github.com/Matt711) -- Improve options docs ([#15888](https://github.com/rapidsai/cudf/pull/15888)) [@bdice](https://github.com/bdice) -- DOC: add linkcode to docs ([#15860](https://github.com/rapidsai/cudf/pull/15860)) [@raybellwaves](https://github.com/raybellwaves) -- DOC: use intersphinx mapping in pandas-compat ext ([#15846](https://github.com/rapidsai/cudf/pull/15846)) [@raybellwaves](https://github.com/raybellwaves) -- Fix inconsistent usage of 'results' and 'records' in read-json.md ([#15766](https://github.com/rapidsai/cudf/pull/15766)) [@dagardner-nv](https://github.com/dagardner-nv) -- Update PandasCompat.py to resolve references ([#15704](https://github.com/rapidsai/cudf/pull/15704)) [@raybellwaves](https://github.com/raybellwaves) +- Add docstring for from_dataframe ([#16260](https://github.com/NVIDIA/cudf/pull/16260)) [@mroeschke](https://github.com/mroeschke) +- Update libcudf compiler requirements in contributing doc ([#16103](https://github.com/NVIDIA/cudf/pull/16103)) [@davidwendt](https://github.com/davidwendt) +- Add libcudf public/detail API pattern to developer guide ([#16086](https://github.com/NVIDIA/cudf/pull/16086)) [@davidwendt](https://github.com/davidwendt) +- Explain line profiler and how to know which functions are GPU-accelerated. ([#16079](https://github.com/NVIDIA/cudf/pull/16079)) [@bdice](https://github.com/bdice) +- cudf.pandas documentation improvement ([#15948](https://github.com/NVIDIA/cudf/pull/15948)) [@Matt711](https://github.com/Matt711) +- Reland "Fix docs for IO readers and strings_convert" ([#15872)" (#15941](https://github.com/NVIDIA/cudf/pull/15872)" (#15941)) [@lithomas1](https://github.com/lithomas1) +- Document how to use cudf.pandas in tandem with multiprocessing ([#15940](https://github.com/NVIDIA/cudf/pull/15940)) [@wence-](https://github.com/wence-) +- DOC: Add documentation for cudf.pandas in the Developer Guide ([#15889](https://github.com/NVIDIA/cudf/pull/15889)) [@Matt711](https://github.com/Matt711) +- Improve options docs ([#15888](https://github.com/NVIDIA/cudf/pull/15888)) [@bdice](https://github.com/bdice) +- DOC: add linkcode to docs ([#15860](https://github.com/NVIDIA/cudf/pull/15860)) [@raybellwaves](https://github.com/raybellwaves) +- DOC: use intersphinx mapping in pandas-compat ext ([#15846](https://github.com/NVIDIA/cudf/pull/15846)) [@raybellwaves](https://github.com/raybellwaves) +- Fix inconsistent usage of 'results' and 'records' in read-json.md ([#15766](https://github.com/NVIDIA/cudf/pull/15766)) [@dagardner-nv](https://github.com/dagardner-nv) +- Update PandasCompat.py to resolve references ([#15704](https://github.com/NVIDIA/cudf/pull/15704)) [@raybellwaves](https://github.com/raybellwaves) ## 🚀 New Features -- Warn on cuDF failure when `POLARS_VERBOSE` is true ([#16308](https://github.com/rapidsai/cudf/pull/16308)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add `drop_nulls` in `cudf-polars` ([#16290](https://github.com/rapidsai/cudf/pull/16290)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- [JNI] Add setKernelPinnedCopyThreshold and setPinnedAllocationThreshold ([#16288](https://github.com/rapidsai/cudf/pull/16288)) [@abellina](https://github.com/abellina) -- Implement support for scan_ndjson in cudf-polars ([#16263](https://github.com/rapidsai/cudf/pull/16263)) [@lithomas1](https://github.com/lithomas1) -- Publish cudf-polars nightlies ([#16213](https://github.com/rapidsai/cudf/pull/16213)) [@lithomas1](https://github.com/lithomas1) -- Modify `make_host_vector` and `make_device_uvector` factories to optionally use pinned memory and kernel copy ([#16206](https://github.com/rapidsai/cudf/pull/16206)) [@vuule](https://github.com/vuule) -- Migrate lists/set_operations to pylibcudf ([#16190](https://github.com/rapidsai/cudf/pull/16190)) [@Matt711](https://github.com/Matt711) -- Migrate lists/filling to pylibcudf ([#16189](https://github.com/rapidsai/cudf/pull/16189)) [@Matt711](https://github.com/Matt711) -- Fall back to CPU for unsupported libcudf binaryops in cudf-polars ([#16188](https://github.com/rapidsai/cudf/pull/16188)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Use resource_ref for upstream in stream_checking_resource_adaptor ([#16187](https://github.com/rapidsai/cudf/pull/16187)) [@harrism](https://github.com/harrism) -- Migrate lists/modifying to pylibcudf ([#16185](https://github.com/rapidsai/cudf/pull/16185)) [@Matt711](https://github.com/Matt711) -- Migrate lists/filtering to pylibcudf ([#16184](https://github.com/rapidsai/cudf/pull/16184)) [@Matt711](https://github.com/Matt711) -- Migrate lists/sorting to pylibcudf ([#16179](https://github.com/rapidsai/cudf/pull/16179)) [@Matt711](https://github.com/Matt711) -- Add missing methods to lists/list_column_view.pxd in pylibcudf ([#16175](https://github.com/rapidsai/cudf/pull/16175)) [@Matt711](https://github.com/Matt711) -- Migrate pylibcudf lists gathering ([#16170](https://github.com/rapidsai/cudf/pull/16170)) [@Matt711](https://github.com/Matt711) -- Move kernel vis over to CUDF_HIDDEN ([#16165](https://github.com/rapidsai/cudf/pull/16165)) [@robertmaynard](https://github.com/robertmaynard) -- Add groupby_max multi-threaded benchmark ([#16154](https://github.com/rapidsai/cudf/pull/16154)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Promote has_nested_columns to cudf public API ([#16131](https://github.com/rapidsai/cudf/pull/16131)) [@robertmaynard](https://github.com/robertmaynard) -- Promote IO support queries to cudf API ([#16125](https://github.com/rapidsai/cudf/pull/16125)) [@robertmaynard](https://github.com/robertmaynard) -- cudf::merge public API now support passing a user stream ([#16124](https://github.com/rapidsai/cudf/pull/16124)) [@robertmaynard](https://github.com/robertmaynard) -- Add TPC-H inspired examples for Libcudf ([#16088](https://github.com/rapidsai/cudf/pull/16088)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Installed cudf header use cudf::allocate_like ([#16087](https://github.com/rapidsai/cudf/pull/16087)) [@robertmaynard](https://github.com/robertmaynard) -- `cudf-polars` string slicing ([#16082](https://github.com/rapidsai/cudf/pull/16082)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate Parquet reader to pylibcudf ([#16078](https://github.com/rapidsai/cudf/pull/16078)) [@lithomas1](https://github.com/lithomas1) -- Migrate lists/count_elements to pylibcudf ([#16072](https://github.com/rapidsai/cudf/pull/16072)) [@Matt711](https://github.com/Matt711) -- Migrate lists/extract to pylibcudf ([#16071](https://github.com/rapidsai/cudf/pull/16071)) [@Matt711](https://github.com/Matt711) -- Move common string utilities to public api ([#16070](https://github.com/rapidsai/cudf/pull/16070)) [@robertmaynard](https://github.com/robertmaynard) -- stable_distinct public api now has a stream parameter ([#16068](https://github.com/rapidsai/cudf/pull/16068)) [@robertmaynard](https://github.com/robertmaynard) -- Migrate expressions to pylibcudf ([#16056](https://github.com/rapidsai/cudf/pull/16056)) [@lithomas1](https://github.com/lithomas1) -- Add support to ArrowDataSource in SourceInfo ([#16050](https://github.com/rapidsai/cudf/pull/16050)) [@lithomas1](https://github.com/lithomas1) -- Experimental support for configurable prefetching ([#16020](https://github.com/rapidsai/cudf/pull/16020)) [@vyasr](https://github.com/vyasr) -- Migrate CSV reader to pylibcudf ([#16011](https://github.com/rapidsai/cudf/pull/16011)) [@lithomas1](https://github.com/lithomas1) -- Migrate string `slice` APIs to `pylibcudf` ([#15988](https://github.com/rapidsai/cudf/pull/15988)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate lists/contains to pylibcudf ([#15981](https://github.com/rapidsai/cudf/pull/15981)) [@Matt711](https://github.com/Matt711) -- Remove CCCL 2.2 patches as we now always use 2.5+ ([#15969](https://github.com/rapidsai/cudf/pull/15969)) [@robertmaynard](https://github.com/robertmaynard) -- Migrate JSON reader to pylibcudf ([#15966](https://github.com/rapidsai/cudf/pull/15966)) [@lithomas1](https://github.com/lithomas1) -- Add a developer check for proxy objects ([#15956](https://github.com/rapidsai/cudf/pull/15956)) [@Matt711](https://github.com/Matt711) -- Start migrating I/O writers to pylibcudf (starting with JSON) ([#15952](https://github.com/rapidsai/cudf/pull/15952)) [@lithomas1](https://github.com/lithomas1) -- Kernel copy for pinned memory ([#15934](https://github.com/rapidsai/cudf/pull/15934)) [@vuule](https://github.com/vuule) -- Migrate left join and conditional join benchmarks to use nvbench ([#15931](https://github.com/rapidsai/cudf/pull/15931)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Migrate lists/combine to pylibcudf ([#15928](https://github.com/rapidsai/cudf/pull/15928)) [@Matt711](https://github.com/Matt711) -- Plumb pylibcudf strings `contains_re` through cudf_polars ([#15918](https://github.com/rapidsai/cudf/pull/15918)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Start migrating I/O to pylibcudf ([#15899](https://github.com/rapidsai/cudf/pull/15899)) [@lithomas1](https://github.com/lithomas1) -- Pinned vector factory that uses the global pool ([#15895](https://github.com/rapidsai/cudf/pull/15895)) [@vuule](https://github.com/vuule) -- Migrate strings `contains` operations to `pylibcudf` ([#15880](https://github.com/rapidsai/cudf/pull/15880)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate quantile.pxd to pylibcudf ([#15874](https://github.com/rapidsai/cudf/pull/15874)) [@lithomas1](https://github.com/lithomas1) -- Migrate round to pylibcudf ([#15863](https://github.com/rapidsai/cudf/pull/15863)) [@lithomas1](https://github.com/lithomas1) -- Migrate string replace.pxd to pylibcudf ([#15839](https://github.com/rapidsai/cudf/pull/15839)) [@lithomas1](https://github.com/lithomas1) -- Add an Environment Variable for debugging the fast path in cudf.pandas ([#15837](https://github.com/rapidsai/cudf/pull/15837)) [@Matt711](https://github.com/Matt711) -- Add an option to run cuIO benchmarks with pinned buffers as input ([#15830](https://github.com/rapidsai/cudf/pull/15830)) [@vuule](https://github.com/vuule) -- Update `pylibcudf` testing utilities ([#15772](https://github.com/rapidsai/cudf/pull/15772)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate string `capitalize` APIs to `pylibcudf` ([#15503](https://github.com/rapidsai/cudf/pull/15503)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add tests for `pylibcudf` binaryops ([#15470](https://github.com/rapidsai/cudf/pull/15470)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrate column factories to pylibcudf ([#15257](https://github.com/rapidsai/cudf/pull/15257)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- cuDF/libcudf exponentially weighted moving averages ([#9027](https://github.com/rapidsai/cudf/pull/9027)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Warn on cuDF failure when `POLARS_VERBOSE` is true ([#16308](https://github.com/NVIDIA/cudf/pull/16308)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add `drop_nulls` in `cudf-polars` ([#16290](https://github.com/NVIDIA/cudf/pull/16290)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- [JNI] Add setKernelPinnedCopyThreshold and setPinnedAllocationThreshold ([#16288](https://github.com/NVIDIA/cudf/pull/16288)) [@abellina](https://github.com/abellina) +- Implement support for scan_ndjson in cudf-polars ([#16263](https://github.com/NVIDIA/cudf/pull/16263)) [@lithomas1](https://github.com/lithomas1) +- Publish cudf-polars nightlies ([#16213](https://github.com/NVIDIA/cudf/pull/16213)) [@lithomas1](https://github.com/lithomas1) +- Modify `make_host_vector` and `make_device_uvector` factories to optionally use pinned memory and kernel copy ([#16206](https://github.com/NVIDIA/cudf/pull/16206)) [@vuule](https://github.com/vuule) +- Migrate lists/set_operations to pylibcudf ([#16190](https://github.com/NVIDIA/cudf/pull/16190)) [@Matt711](https://github.com/Matt711) +- Migrate lists/filling to pylibcudf ([#16189](https://github.com/NVIDIA/cudf/pull/16189)) [@Matt711](https://github.com/Matt711) +- Fall back to CPU for unsupported libcudf binaryops in cudf-polars ([#16188](https://github.com/NVIDIA/cudf/pull/16188)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Use resource_ref for upstream in stream_checking_resource_adaptor ([#16187](https://github.com/NVIDIA/cudf/pull/16187)) [@harrism](https://github.com/harrism) +- Migrate lists/modifying to pylibcudf ([#16185](https://github.com/NVIDIA/cudf/pull/16185)) [@Matt711](https://github.com/Matt711) +- Migrate lists/filtering to pylibcudf ([#16184](https://github.com/NVIDIA/cudf/pull/16184)) [@Matt711](https://github.com/Matt711) +- Migrate lists/sorting to pylibcudf ([#16179](https://github.com/NVIDIA/cudf/pull/16179)) [@Matt711](https://github.com/Matt711) +- Add missing methods to lists/list_column_view.pxd in pylibcudf ([#16175](https://github.com/NVIDIA/cudf/pull/16175)) [@Matt711](https://github.com/Matt711) +- Migrate pylibcudf lists gathering ([#16170](https://github.com/NVIDIA/cudf/pull/16170)) [@Matt711](https://github.com/Matt711) +- Move kernel vis over to CUDF_HIDDEN ([#16165](https://github.com/NVIDIA/cudf/pull/16165)) [@robertmaynard](https://github.com/robertmaynard) +- Add groupby_max multi-threaded benchmark ([#16154](https://github.com/NVIDIA/cudf/pull/16154)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Promote has_nested_columns to cudf public API ([#16131](https://github.com/NVIDIA/cudf/pull/16131)) [@robertmaynard](https://github.com/robertmaynard) +- Promote IO support queries to cudf API ([#16125](https://github.com/NVIDIA/cudf/pull/16125)) [@robertmaynard](https://github.com/robertmaynard) +- cudf::merge public API now support passing a user stream ([#16124](https://github.com/NVIDIA/cudf/pull/16124)) [@robertmaynard](https://github.com/robertmaynard) +- Add TPC-H inspired examples for Libcudf ([#16088](https://github.com/NVIDIA/cudf/pull/16088)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Installed cudf header use cudf::allocate_like ([#16087](https://github.com/NVIDIA/cudf/pull/16087)) [@robertmaynard](https://github.com/robertmaynard) +- `cudf-polars` string slicing ([#16082](https://github.com/NVIDIA/cudf/pull/16082)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate Parquet reader to pylibcudf ([#16078](https://github.com/NVIDIA/cudf/pull/16078)) [@lithomas1](https://github.com/lithomas1) +- Migrate lists/count_elements to pylibcudf ([#16072](https://github.com/NVIDIA/cudf/pull/16072)) [@Matt711](https://github.com/Matt711) +- Migrate lists/extract to pylibcudf ([#16071](https://github.com/NVIDIA/cudf/pull/16071)) [@Matt711](https://github.com/Matt711) +- Move common string utilities to public api ([#16070](https://github.com/NVIDIA/cudf/pull/16070)) [@robertmaynard](https://github.com/robertmaynard) +- stable_distinct public api now has a stream parameter ([#16068](https://github.com/NVIDIA/cudf/pull/16068)) [@robertmaynard](https://github.com/robertmaynard) +- Migrate expressions to pylibcudf ([#16056](https://github.com/NVIDIA/cudf/pull/16056)) [@lithomas1](https://github.com/lithomas1) +- Add support to ArrowDataSource in SourceInfo ([#16050](https://github.com/NVIDIA/cudf/pull/16050)) [@lithomas1](https://github.com/lithomas1) +- Experimental support for configurable prefetching ([#16020](https://github.com/NVIDIA/cudf/pull/16020)) [@vyasr](https://github.com/vyasr) +- Migrate CSV reader to pylibcudf ([#16011](https://github.com/NVIDIA/cudf/pull/16011)) [@lithomas1](https://github.com/lithomas1) +- Migrate string `slice` APIs to `pylibcudf` ([#15988](https://github.com/NVIDIA/cudf/pull/15988)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate lists/contains to pylibcudf ([#15981](https://github.com/NVIDIA/cudf/pull/15981)) [@Matt711](https://github.com/Matt711) +- Remove CCCL 2.2 patches as we now always use 2.5+ ([#15969](https://github.com/NVIDIA/cudf/pull/15969)) [@robertmaynard](https://github.com/robertmaynard) +- Migrate JSON reader to pylibcudf ([#15966](https://github.com/NVIDIA/cudf/pull/15966)) [@lithomas1](https://github.com/lithomas1) +- Add a developer check for proxy objects ([#15956](https://github.com/NVIDIA/cudf/pull/15956)) [@Matt711](https://github.com/Matt711) +- Start migrating I/O writers to pylibcudf (starting with JSON) ([#15952](https://github.com/NVIDIA/cudf/pull/15952)) [@lithomas1](https://github.com/lithomas1) +- Kernel copy for pinned memory ([#15934](https://github.com/NVIDIA/cudf/pull/15934)) [@vuule](https://github.com/vuule) +- Migrate left join and conditional join benchmarks to use nvbench ([#15931](https://github.com/NVIDIA/cudf/pull/15931)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Migrate lists/combine to pylibcudf ([#15928](https://github.com/NVIDIA/cudf/pull/15928)) [@Matt711](https://github.com/Matt711) +- Plumb pylibcudf strings `contains_re` through cudf_polars ([#15918](https://github.com/NVIDIA/cudf/pull/15918)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Start migrating I/O to pylibcudf ([#15899](https://github.com/NVIDIA/cudf/pull/15899)) [@lithomas1](https://github.com/lithomas1) +- Pinned vector factory that uses the global pool ([#15895](https://github.com/NVIDIA/cudf/pull/15895)) [@vuule](https://github.com/vuule) +- Migrate strings `contains` operations to `pylibcudf` ([#15880](https://github.com/NVIDIA/cudf/pull/15880)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate quantile.pxd to pylibcudf ([#15874](https://github.com/NVIDIA/cudf/pull/15874)) [@lithomas1](https://github.com/lithomas1) +- Migrate round to pylibcudf ([#15863](https://github.com/NVIDIA/cudf/pull/15863)) [@lithomas1](https://github.com/lithomas1) +- Migrate string replace.pxd to pylibcudf ([#15839](https://github.com/NVIDIA/cudf/pull/15839)) [@lithomas1](https://github.com/lithomas1) +- Add an Environment Variable for debugging the fast path in cudf.pandas ([#15837](https://github.com/NVIDIA/cudf/pull/15837)) [@Matt711](https://github.com/Matt711) +- Add an option to run cuIO benchmarks with pinned buffers as input ([#15830](https://github.com/NVIDIA/cudf/pull/15830)) [@vuule](https://github.com/vuule) +- Update `pylibcudf` testing utilities ([#15772](https://github.com/NVIDIA/cudf/pull/15772)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate string `capitalize` APIs to `pylibcudf` ([#15503](https://github.com/NVIDIA/cudf/pull/15503)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add tests for `pylibcudf` binaryops ([#15470](https://github.com/NVIDIA/cudf/pull/15470)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrate column factories to pylibcudf ([#15257](https://github.com/NVIDIA/cudf/pull/15257)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- cuDF/libcudf exponentially weighted moving averages ([#9027](https://github.com/NVIDIA/cudf/pull/9027)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🛠️ Improvements -- Ensure objects with __interface__ are converted to cupy/numpy arrays ([#16436](https://github.com/rapidsai/cudf/pull/16436)) [@mroeschke](https://github.com/mroeschke) -- Add about rmm modes in `cudf.pandas` docs ([#16404](https://github.com/rapidsai/cudf/pull/16404)) [@galipremsagar](https://github.com/galipremsagar) -- Gracefully CUDF_FAIL when `skip_rows > 0` in Chunked Parquet reader ([#16385](https://github.com/rapidsai/cudf/pull/16385)) [@mhaseeb123](https://github.com/mhaseeb123) -- Make C++ compilation warning free after #16297 ([#16379](https://github.com/rapidsai/cudf/pull/16379)) [@wence-](https://github.com/wence-) -- Align Index __init__ APIs with pandas 2.x ([#16362](https://github.com/rapidsai/cudf/pull/16362)) [@mroeschke](https://github.com/mroeschke) -- Use rapids_cpm_bs_thread_pool() ([#16360](https://github.com/rapidsai/cudf/pull/16360)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Rename PrefetchConfig to prefetch_config. ([#16358](https://github.com/rapidsai/cudf/pull/16358)) [@bdice](https://github.com/bdice) -- Implement parquet reading using pylibcudf in cudf-polars ([#16346](https://github.com/rapidsai/cudf/pull/16346)) [@lithomas1](https://github.com/lithomas1) -- Fix compile warnings with `jni_utils.hpp` ([#16336](https://github.com/rapidsai/cudf/pull/16336)) [@ttnghia](https://github.com/ttnghia) -- Align Series APIs with pandas 2.x ([#16333](https://github.com/rapidsai/cudf/pull/16333)) [@mroeschke](https://github.com/mroeschke) -- Add missing `stream` param to dictionary factory APIs ([#16319](https://github.com/rapidsai/cudf/pull/16319)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Mark cudf._typing as a typing module in ruff ([#16318](https://github.com/rapidsai/cudf/pull/16318)) [@mroeschke](https://github.com/mroeschke) -- Add `stream` param to list explode APIs ([#16317](https://github.com/rapidsai/cudf/pull/16317)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Fix polars for 1.2.1 ([#16316](https://github.com/rapidsai/cudf/pull/16316)) [@lithomas1](https://github.com/lithomas1) -- Use workflow branch 24.08 again ([#16314](https://github.com/rapidsai/cudf/pull/16314)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Deprecate dtype= parameter in reduction methods ([#16313](https://github.com/rapidsai/cudf/pull/16313)) [@mroeschke](https://github.com/mroeschke) -- Remove squeeze argument from groupby ([#16312](https://github.com/rapidsai/cudf/pull/16312)) [@mroeschke](https://github.com/mroeschke) -- Align more DataFrame APIs with pandas ([#16310](https://github.com/rapidsai/cudf/pull/16310)) [@mroeschke](https://github.com/mroeschke) -- Clean unneeded/redudant dtype utils ([#16309](https://github.com/rapidsai/cudf/pull/16309)) [@mroeschke](https://github.com/mroeschke) -- Implement read_csv in cudf-polars using pylibcudf ([#16307](https://github.com/rapidsai/cudf/pull/16307)) [@lithomas1](https://github.com/lithomas1) -- Use Column.can_cast_safely instead of some ad-hoc dtype functions in .where ([#16303](https://github.com/rapidsai/cudf/pull/16303)) [@mroeschke](https://github.com/mroeschke) -- Drop `{{ pin_compatible('numpy', max_pin='x') }}` ([#16301](https://github.com/rapidsai/cudf/pull/16301)) [@jakirkham](https://github.com/jakirkham) -- Host implementation of `to_arrow` using nanoarrow ([#16297](https://github.com/rapidsai/cudf/pull/16297)) [@zeroshade](https://github.com/zeroshade) -- Add ability to prefetch in `cudf.pandas` and change default to managed pool ([#16296](https://github.com/rapidsai/cudf/pull/16296)) [@galipremsagar](https://github.com/galipremsagar) -- Fix tests for polars 1.2 ([#16292](https://github.com/rapidsai/cudf/pull/16292)) [@lithomas1](https://github.com/lithomas1) -- Introduce dedicated options for low memory readers ([#16289](https://github.com/rapidsai/cudf/pull/16289)) [@galipremsagar](https://github.com/galipremsagar) -- Remove decimal/floating 64/128bit switches due to register pressure ([#16287](https://github.com/rapidsai/cudf/pull/16287)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Make ColumnAccessor strictly require a mapping of columns ([#16285](https://github.com/rapidsai/cudf/pull/16285)) [@mroeschke](https://github.com/mroeschke) -- Introduce version file so we can conditionally handle things in tests ([#16280](https://github.com/rapidsai/cudf/pull/16280)) [@wence-](https://github.com/wence-) -- Type & reduce cupy usage ([#16277](https://github.com/rapidsai/cudf/pull/16277)) [@mroeschke](https://github.com/mroeschke) -- Update cudf::detail::grid_1d to use thread_index_type ([#16276](https://github.com/rapidsai/cudf/pull/16276)) [@davidwendt](https://github.com/davidwendt) -- Replace np.isscalar/issubdtype checks with is_scalar/.kind checks ([#16275](https://github.com/rapidsai/cudf/pull/16275)) [@mroeschke](https://github.com/mroeschke) -- Remove xml from sort_ninja_log.py utility ([#16274](https://github.com/rapidsai/cudf/pull/16274)) [@davidwendt](https://github.com/davidwendt) -- Fix issue in horizontal concat implementation in cudf-polars ([#16271](https://github.com/rapidsai/cudf/pull/16271)) [@wence-](https://github.com/wence-) -- Preserve order in left join for cudf-polars ([#16268](https://github.com/rapidsai/cudf/pull/16268)) [@wence-](https://github.com/wence-) -- Replace is_datetime/timedelta_dtype checks with .kind checks ([#16262](https://github.com/rapidsai/cudf/pull/16262)) [@mroeschke](https://github.com/mroeschke) -- Replace is_float/integer_dtype checks with .kind checks ([#16261](https://github.com/rapidsai/cudf/pull/16261)) [@mroeschke](https://github.com/mroeschke) -- Build and test with CUDA 12.5.1 ([#16259](https://github.com/rapidsai/cudf/pull/16259)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Replace is_bool_type with checking .dtype.kind ([#16255](https://github.com/rapidsai/cudf/pull/16255)) [@mroeschke](https://github.com/mroeschke) -- remove `cuco_noexcept.diff` ([#16254](https://github.com/rapidsai/cudf/pull/16254)) [@trxcllnt](https://github.com/trxcllnt) -- Update contains_tests.cpp to use public cudf::slice ([#16253](https://github.com/rapidsai/cudf/pull/16253)) [@davidwendt](https://github.com/davidwendt) -- Improve the test data for pylibcudf I/O tests ([#16247](https://github.com/rapidsai/cudf/pull/16247)) [@lithomas1](https://github.com/lithomas1) -- Short circuit some Column methods ([#16246](https://github.com/rapidsai/cudf/pull/16246)) [@mroeschke](https://github.com/mroeschke) -- Make nvcomp adapter compatible with new version macros ([#16245](https://github.com/rapidsai/cudf/pull/16245)) [@vuule](https://github.com/vuule) -- Add Column.strftime/strptime instead of overloading `as_string/datetime/timedelta_column` ([#16243](https://github.com/rapidsai/cudf/pull/16243)) [@mroeschke](https://github.com/mroeschke) -- Remove temporary functor overloads required by cuco version bump ([#16242](https://github.com/rapidsai/cudf/pull/16242)) [@PointKernel](https://github.com/PointKernel) -- Remove hash_character_ngrams dependency from jaccard_index ([#16241](https://github.com/rapidsai/cudf/pull/16241)) [@davidwendt](https://github.com/davidwendt) -- Expose sorted groupby parameters to pylibcudf ([#16240](https://github.com/rapidsai/cudf/pull/16240)) [@wence-](https://github.com/wence-) -- Expose reflection to check if casting between two types is supported ([#16239](https://github.com/rapidsai/cudf/pull/16239)) [@wence-](https://github.com/wence-) -- Handle nans in groupby-aggregations in polars executor ([#16233](https://github.com/rapidsai/cudf/pull/16233)) [@wence-](https://github.com/wence-) -- Remove `mr` param from `write_csv` and `write_json` ([#16231](https://github.com/rapidsai/cudf/pull/16231)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) -- Support Literals in groupby-agg ([#16218](https://github.com/rapidsai/cudf/pull/16218)) [@wence-](https://github.com/wence-) -- Handler csv reader options in cudf-polars ([#16211](https://github.com/rapidsai/cudf/pull/16211)) [@wence-](https://github.com/wence-) -- Update vendored thread_pool implementation ([#16210](https://github.com/rapidsai/cudf/pull/16210)) [@wence-](https://github.com/wence-) -- Add low memory JSON reader for `cudf.pandas` ([#16204](https://github.com/rapidsai/cudf/pull/16204)) [@galipremsagar](https://github.com/galipremsagar) -- Clean up state variables in MultiIndex ([#16203](https://github.com/rapidsai/cudf/pull/16203)) [@mroeschke](https://github.com/mroeschke) -- skip CMake 3.30.0 ([#16202](https://github.com/rapidsai/cudf/pull/16202)) [@jameslamb](https://github.com/jameslamb) -- Assert valid metadata is passed in to_arrow for list_view ([#16198](https://github.com/rapidsai/cudf/pull/16198)) [@wence-](https://github.com/wence-) -- Expose type traits to pylibcudf ([#16197](https://github.com/rapidsai/cudf/pull/16197)) [@wence-](https://github.com/wence-) -- Report number of rows per file read by PQ reader when no row selection and fix segfault in chunked PQ reader when skip_rows > 0 ([#16195](https://github.com/rapidsai/cudf/pull/16195)) [@mhaseeb123](https://github.com/mhaseeb123) -- Cast count aggs to correct dtype in translation ([#16192](https://github.com/rapidsai/cudf/pull/16192)) [@wence-](https://github.com/wence-) -- Some small fixes in cudf-polars ([#16191](https://github.com/rapidsai/cudf/pull/16191)) [@wence-](https://github.com/wence-) -- split up CUDA-suffixed dependencies in dependencies.yaml ([#16183](https://github.com/rapidsai/cudf/pull/16183)) [@jameslamb](https://github.com/jameslamb) -- Define PTDS for the stream hook libs ([#16182](https://github.com/rapidsai/cudf/pull/16182)) [@trxcllnt](https://github.com/trxcllnt) -- Make `test_python_cudf_pandas` generate `requirements.txt` ([#16181](https://github.com/rapidsai/cudf/pull/16181)) [@trxcllnt](https://github.com/trxcllnt) -- Add environment-agnostic `ci/run_cudf_polars_pytest.sh` ([#16178](https://github.com/rapidsai/cudf/pull/16178)) [@trxcllnt](https://github.com/trxcllnt) -- Implement translation for some unary functions and a single datetime extraction ([#16173](https://github.com/rapidsai/cudf/pull/16173)) [@wence-](https://github.com/wence-) -- Remove size constraints on source files in batched JSON reading ([#16162](https://github.com/rapidsai/cudf/pull/16162)) [@shrshi](https://github.com/shrshi) -- CI: Build wheels for cudf-polars ([#16156](https://github.com/rapidsai/cudf/pull/16156)) [@lithomas1](https://github.com/lithomas1) -- Update cudf-polars for v1 release of polars ([#16149](https://github.com/rapidsai/cudf/pull/16149)) [@wence-](https://github.com/wence-) -- Use strings concatenate to support large strings in CSV writer ([#16148](https://github.com/rapidsai/cudf/pull/16148)) [@davidwendt](https://github.com/davidwendt) -- Use verify-alpha-spec hook ([#16144](https://github.com/rapidsai/cudf/pull/16144)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Adds write-coalescing code path optimization to FST ([#16143](https://github.com/rapidsai/cudf/pull/16143)) [@elstehle](https://github.com/elstehle) -- MAINT: Adapt to NumPy 2 promotion changes ([#16141](https://github.com/rapidsai/cudf/pull/16141)) [@seberg](https://github.com/seberg) -- API: Check for integer overflows when creating scalar form python int ([#16140](https://github.com/rapidsai/cudf/pull/16140)) [@seberg](https://github.com/seberg) -- Remove the (unused) implementation of `host_parse_nested_json` ([#16135](https://github.com/rapidsai/cudf/pull/16135)) [@vuule](https://github.com/vuule) -- Deprecate Arrow support in I/O ([#16132](https://github.com/rapidsai/cudf/pull/16132)) [@lithomas1](https://github.com/lithomas1) -- Disable dict support for split-page kernel in the parquet reader. ([#16128](https://github.com/rapidsai/cudf/pull/16128)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add throughput metrics for REDUCTION_BENCH/REDUCTION_NVBENCH benchmarks ([#16126](https://github.com/rapidsai/cudf/pull/16126)) [@jihoonson](https://github.com/jihoonson) -- Add ensure_index to not unnecessarily shallow copy cudf.Index ([#16117](https://github.com/rapidsai/cudf/pull/16117)) [@mroeschke](https://github.com/mroeschke) -- Make binary operators work between fixed-point and floating args ([#16116](https://github.com/rapidsai/cudf/pull/16116)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Implement Ternary copy_if_else ([#16114](https://github.com/rapidsai/cudf/pull/16114)) [@wence-](https://github.com/wence-) -- Implement handlers for series literal in cudf-polars ([#16113](https://github.com/rapidsai/cudf/pull/16113)) [@wence-](https://github.com/wence-) -- Fix dtype errors in `StringArrays` ([#16111](https://github.com/rapidsai/cudf/pull/16111)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure MultiIndex.to_frame deep copies columns ([#16110](https://github.com/rapidsai/cudf/pull/16110)) [@mroeschke](https://github.com/mroeschke) -- Parallelize `gpuInitStringDescriptors` for fixed length byte array data ([#16109](https://github.com/rapidsai/cudf/pull/16109)) [@mhaseeb123](https://github.com/mhaseeb123) -- Finish implementation of cudf-polars boolean function handlers ([#16098](https://github.com/rapidsai/cudf/pull/16098)) [@wence-](https://github.com/wence-) -- Expose and then implement support for cross joins in cudf-polars ([#16097](https://github.com/rapidsai/cudf/pull/16097)) [@wence-](https://github.com/wence-) -- Defer copying in Column.astype(copy=True) ([#16095](https://github.com/rapidsai/cudf/pull/16095)) [@mroeschke](https://github.com/mroeschke) -- Fix segfault in conditional join ([#16094](https://github.com/rapidsai/cudf/pull/16094)) [@bdice](https://github.com/bdice) -- Free temp memory no longer needed in multibyte_split processing ([#16091](https://github.com/rapidsai/cudf/pull/16091)) [@davidwendt](https://github.com/davidwendt) -- Rename gather/scatter benchmarks to clarify coalesced behavior. ([#16083](https://github.com/rapidsai/cudf/pull/16083)) [@bdice](https://github.com/bdice) -- Adapt to polars upstream changes and turn on CI testing ([#16081](https://github.com/rapidsai/cudf/pull/16081)) [@wence-](https://github.com/wence-) -- Reduce/clean copy usage in Series, reshaping ([#16080](https://github.com/rapidsai/cudf/pull/16080)) [@mroeschke](https://github.com/mroeschke) -- Account for FIXED_LEN_BYTE_ARRAY when calculating fragment sizes in Parquet writer ([#16064](https://github.com/rapidsai/cudf/pull/16064)) [@etseidl](https://github.com/etseidl) -- Reduce (shallow) copies in DataFrame ops ([#16060](https://github.com/rapidsai/cudf/pull/16060)) [@mroeschke](https://github.com/mroeschke) -- Add multi-file support to `dask_cudf.read_json` ([#16057](https://github.com/rapidsai/cudf/pull/16057)) [@rjzamora](https://github.com/rjzamora) -- Reduce deep copies in Index ops ([#16054](https://github.com/rapidsai/cudf/pull/16054)) [@mroeschke](https://github.com/mroeschke) -- Implement chunked column wise concat in chunked parquet reader ([#16052](https://github.com/rapidsai/cudf/pull/16052)) [@galipremsagar](https://github.com/galipremsagar) -- Add exception when trying to create large strings with cudf::test::strings_column_wrapper ([#16049](https://github.com/rapidsai/cudf/pull/16049)) [@davidwendt](https://github.com/davidwendt) -- Return `FrozenList` for `Index.names` ([#16047](https://github.com/rapidsai/cudf/pull/16047)) [@galipremsagar](https://github.com/galipremsagar) -- Add ast cast test ([#16045](https://github.com/rapidsai/cudf/pull/16045)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Remove `override_dtypes` and `include_index` from `Frame._copy_type_metadata` ([#16043](https://github.com/rapidsai/cudf/pull/16043)) [@mroeschke](https://github.com/mroeschke) -- Add ruff rules to avoid importing from typing ([#16040](https://github.com/rapidsai/cudf/pull/16040)) [@mroeschke](https://github.com/mroeschke) -- Fix decimal -> float cast in ast code ([#16038](https://github.com/rapidsai/cudf/pull/16038)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Add compile option to enable large strings support ([#16037](https://github.com/rapidsai/cudf/pull/16037)) [@davidwendt](https://github.com/davidwendt) -- Reduce conditional_join nvbench configurations ([#16036](https://github.com/rapidsai/cudf/pull/16036)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Project automation update: skip if not in project ([#16035](https://github.com/rapidsai/cudf/pull/16035)) [@jarmak-nv](https://github.com/jarmak-nv) -- Add stream parameter to cudf::io::text::multibyte_split ([#16034](https://github.com/rapidsai/cudf/pull/16034)) [@davidwendt](https://github.com/davidwendt) -- Delete unused code from stringfunction evaluator ([#16032](https://github.com/rapidsai/cudf/pull/16032)) [@wence-](https://github.com/wence-) -- Fix exclude regex in pre-commit clang-format hook ([#16030](https://github.com/rapidsai/cudf/pull/16030)) [@wence-](https://github.com/wence-) -- Refactor rmm usage in `cudf.pandas` ([#16021](https://github.com/rapidsai/cudf/pull/16021)) [@galipremsagar](https://github.com/galipremsagar) -- Enable ruff TCH: typing imports under if TYPE_CHECKING ([#16015](https://github.com/rapidsai/cudf/pull/16015)) [@mroeschke](https://github.com/mroeschke) -- Restrict the allowed pandas timezone objects in cudf ([#16013](https://github.com/rapidsai/cudf/pull/16013)) [@mroeschke](https://github.com/mroeschke) -- orc multithreaded benchmark ([#16009](https://github.com/rapidsai/cudf/pull/16009)) [@zpuller](https://github.com/zpuller) -- Add tests of expression-based sort and sort-by ([#16008](https://github.com/rapidsai/cudf/pull/16008)) [@wence-](https://github.com/wence-) -- Add tests of implemented StringFunctions ([#16007](https://github.com/rapidsai/cudf/pull/16007)) [@wence-](https://github.com/wence-) -- Add test that diagonal concat with mismatching schemas raises ([#16006](https://github.com/rapidsai/cudf/pull/16006)) [@wence-](https://github.com/wence-) -- Add coverage selecting len from a dataframe (number of rows) ([#16005](https://github.com/rapidsai/cudf/pull/16005)) [@wence-](https://github.com/wence-) -- Add basic tests of dataframe scan ([#16003](https://github.com/rapidsai/cudf/pull/16003)) [@wence-](https://github.com/wence-) -- Add coverage for both expression and dataframe filter ([#16002](https://github.com/rapidsai/cudf/pull/16002)) [@wence-](https://github.com/wence-) -- Remove deprecated ExtContext node ([#16001](https://github.com/rapidsai/cudf/pull/16001)) [@wence-](https://github.com/wence-) -- Fix typo bug in gather implementation ([#16000](https://github.com/rapidsai/cudf/pull/16000)) [@wence-](https://github.com/wence-) -- Extend coverage of groupby and rolling window nodes ([#15999](https://github.com/rapidsai/cudf/pull/15999)) [@wence-](https://github.com/wence-) -- Coverage of binops where one or both operands are a scalar ([#15998](https://github.com/rapidsai/cudf/pull/15998)) [@wence-](https://github.com/wence-) -- Add full coverage for whole-frame Agg expressions ([#15997](https://github.com/rapidsai/cudf/pull/15997)) [@wence-](https://github.com/wence-) -- Add tests covering magic methods of Expr objects ([#15996](https://github.com/rapidsai/cudf/pull/15996)) [@wence-](https://github.com/wence-) -- Add full coverage of utility functions ([#15995](https://github.com/rapidsai/cudf/pull/15995)) [@wence-](https://github.com/wence-) -- Test behaviour of containers ([#15994](https://github.com/rapidsai/cudf/pull/15994)) [@wence-](https://github.com/wence-) -- Fix implemention of any, all, and isbetween ([#15993](https://github.com/rapidsai/cudf/pull/15993)) [@wence-](https://github.com/wence-) -- Raise early on unhandled PythonScan node ([#15992](https://github.com/rapidsai/cudf/pull/15992)) [@wence-](https://github.com/wence-) -- Remove mapfunction nodes that don't exist/aren't supported ([#15991](https://github.com/rapidsai/cudf/pull/15991)) [@wence-](https://github.com/wence-) -- Add test coverage for slicing with "out of bounds" negative indices ([#15990](https://github.com/rapidsai/cudf/pull/15990)) [@wence-](https://github.com/wence-) -- Standardize and type `Series.dt` methods ([#15987](https://github.com/rapidsai/cudf/pull/15987)) [@mroeschke](https://github.com/mroeschke) -- Refactor distinct with hashset-based algorithms ([#15984](https://github.com/rapidsai/cudf/pull/15984)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- resolve dependency-file-generator warning, remove unnecessary rapids-build-backend configuration ([#15980](https://github.com/rapidsai/cudf/pull/15980)) [@jameslamb](https://github.com/jameslamb) -- Project automation bug fixes ([#15971](https://github.com/rapidsai/cudf/pull/15971)) [@jarmak-nv](https://github.com/jarmak-nv) -- Add typing to single_column_frame ([#15965](https://github.com/rapidsai/cudf/pull/15965)) [@mroeschke](https://github.com/mroeschke) -- Move some misc Frame methods to appropriate locations ([#15963](https://github.com/rapidsai/cudf/pull/15963)) [@mroeschke](https://github.com/mroeschke) -- Condense pylibcudf data fixtures ([#15958](https://github.com/rapidsai/cudf/pull/15958)) [@lithomas1](https://github.com/lithomas1) -- Refactor fillna logic to push specifics toward Frame subclasses and Column subclasses ([#15957](https://github.com/rapidsai/cudf/pull/15957)) [@mroeschke](https://github.com/mroeschke) -- Remove unused parsing utilities ([#15955](https://github.com/rapidsai/cudf/pull/15955)) [@vuule](https://github.com/vuule) -- Remove `Scalar` container type from polars interpreter ([#15953](https://github.com/rapidsai/cudf/pull/15953)) [@wence-](https://github.com/wence-) -- Support arbitrary CUDA versions in UDF code ([#15950](https://github.com/rapidsai/cudf/pull/15950)) [@bdice](https://github.com/bdice) -- Support large strings in cudf::io::text::multibyte_split ([#15947](https://github.com/rapidsai/cudf/pull/15947)) [@davidwendt](https://github.com/davidwendt) -- Add external issue label and project automation ([#15945](https://github.com/rapidsai/cudf/pull/15945)) [@jarmak-nv](https://github.com/jarmak-nv) -- Enable round-tripping of large strings in `cudf` ([#15944](https://github.com/rapidsai/cudf/pull/15944)) [@galipremsagar](https://github.com/galipremsagar) -- Add more complete type annotations in polars interpreter ([#15942](https://github.com/rapidsai/cudf/pull/15942)) [@wence-](https://github.com/wence-) -- Update implementations to build with the latest cuco ([#15938](https://github.com/rapidsai/cudf/pull/15938)) [@PointKernel](https://github.com/PointKernel) -- Support timezone aware pandas inputs in cudf ([#15935](https://github.com/rapidsai/cudf/pull/15935)) [@mroeschke](https://github.com/mroeschke) -- Define Column.nan_as_null to return self ([#15923](https://github.com/rapidsai/cudf/pull/15923)) [@mroeschke](https://github.com/mroeschke) -- Make Frame._dtype an iterator instead of a dict ([#15920](https://github.com/rapidsai/cudf/pull/15920)) [@mroeschke](https://github.com/mroeschke) -- Port start of datetime.hpp to pylibcudf ([#15916](https://github.com/rapidsai/cudf/pull/15916)) [@wence-](https://github.com/wence-) -- Introduce `NamedColumn` concept in cudf-polars ([#15914](https://github.com/rapidsai/cudf/pull/15914)) [@wence-](https://github.com/wence-) -- Avoid redefining Frame._get_columns_by_label in subclasses ([#15912](https://github.com/rapidsai/cudf/pull/15912)) [@mroeschke](https://github.com/mroeschke) -- Templatization of fixed-width parquet decoding kernels. ([#15911](https://github.com/rapidsai/cudf/pull/15911)) [@nvdbaranec](https://github.com/nvdbaranec) -- New Decimal <--> Floating conversion ([#15905](https://github.com/rapidsai/cudf/pull/15905)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Use Arrow C Data Interface functions for Python interop ([#15904](https://github.com/rapidsai/cudf/pull/15904)) [@vyasr](https://github.com/vyasr) -- Use offsetalator in cudf::io::json::detail::parse_string ([#15900](https://github.com/rapidsai/cudf/pull/15900)) [@davidwendt](https://github.com/davidwendt) -- Rename strings multiple target replace API ([#15898](https://github.com/rapidsai/cudf/pull/15898)) [@davidwendt](https://github.com/davidwendt) -- Apply clang-tidy autofixes ([#15894](https://github.com/rapidsai/cudf/pull/15894)) [@vyasr](https://github.com/vyasr) -- Update Python labels and remove unnecessary ones ([#15893](https://github.com/rapidsai/cudf/pull/15893)) [@vyasr](https://github.com/vyasr) -- Clean up pylibcudf test assertations ([#15892](https://github.com/rapidsai/cudf/pull/15892)) [@lithomas1](https://github.com/lithomas1) -- Use offsetalator in orc rowgroup_char_counts_kernel ([#15891](https://github.com/rapidsai/cudf/pull/15891)) [@davidwendt](https://github.com/davidwendt) -- Ensure literals have correct dtype ([#15890](https://github.com/rapidsai/cudf/pull/15890)) [@wence-](https://github.com/wence-) -- Add overflow check when converting large strings to lists columns ([#15887](https://github.com/rapidsai/cudf/pull/15887)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in nvtext::tokenize_with_vocabulary ([#15878](https://github.com/rapidsai/cudf/pull/15878)) [@davidwendt](https://github.com/davidwendt) -- Update interleave lists column for large strings ([#15877](https://github.com/rapidsai/cudf/pull/15877)) [@davidwendt](https://github.com/davidwendt) -- Simple NumPy 2 fixes that are clearly no behavior change ([#15876](https://github.com/rapidsai/cudf/pull/15876)) [@seberg](https://github.com/seberg) -- Support `arrow:schema` in Parquet writer to faithfully roundtrip `duration` types with Arrow ([#15875](https://github.com/rapidsai/cudf/pull/15875)) [@mhaseeb123](https://github.com/mhaseeb123) -- Refactor join benchmarks to target public APIs with the default stream ([#15873](https://github.com/rapidsai/cudf/pull/15873)) [@PointKernel](https://github.com/PointKernel) -- Fix url-decode benchmark to use offsetalator ([#15871](https://github.com/rapidsai/cudf/pull/15871)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in strings shift functor ([#15870](https://github.com/rapidsai/cudf/pull/15870)) [@davidwendt](https://github.com/davidwendt) -- Memory Profiling ([#15866](https://github.com/rapidsai/cudf/pull/15866)) [@madsbk](https://github.com/madsbk) -- Expose stream parameter to public rolling APIs ([#15865](https://github.com/rapidsai/cudf/pull/15865)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- Make Frame.astype return Self instead of a ColumnAccessor ([#15861](https://github.com/rapidsai/cudf/pull/15861)) [@mroeschke](https://github.com/mroeschke) -- Use ColumnAccessor row and column length attributes more consistently ([#15857](https://github.com/rapidsai/cudf/pull/15857)) [@mroeschke](https://github.com/mroeschke) -- add unit test setup for cudf_kafka ([#15853](https://github.com/rapidsai/cudf/pull/15853)) [@jameslamb](https://github.com/jameslamb) -- Remove internal usage of core.index.as_index in favor of cudf.Index ([#15851](https://github.com/rapidsai/cudf/pull/15851)) [@mroeschke](https://github.com/mroeschke) -- Ensure cudf.Series(cudf.Series(...)) creates a reference to the same index ([#15845](https://github.com/rapidsai/cudf/pull/15845)) [@mroeschke](https://github.com/mroeschke) -- Remove benchmark-specific use of pinned-pooled memory in Parquet multithreaded benchmark. ([#15838](https://github.com/rapidsai/cudf/pull/15838)) [@nvdbaranec](https://github.com/nvdbaranec) -- Implement `on_bad_lines` in json reader ([#15834](https://github.com/rapidsai/cudf/pull/15834)) [@galipremsagar](https://github.com/galipremsagar) -- Make Column.to_pandas return Index instead of Series ([#15833](https://github.com/rapidsai/cudf/pull/15833)) [@mroeschke](https://github.com/mroeschke) -- Add test of interoperability of cuDF and arrow BYTE_STREAM_SPLIT encoders ([#15832](https://github.com/rapidsai/cudf/pull/15832)) [@etseidl](https://github.com/etseidl) -- Refactor Parquet writer options and builders ([#15831](https://github.com/rapidsai/cudf/pull/15831)) [@etseidl](https://github.com/etseidl) -- Migrate reshape.pxd to pylibcudf ([#15827](https://github.com/rapidsai/cudf/pull/15827)) [@lithomas1](https://github.com/lithomas1) -- Remove legacy JSON reader and concurrent_unordered_map.cuh. ([#15813](https://github.com/rapidsai/cudf/pull/15813)) [@bdice](https://github.com/bdice) -- Switch cuIO benchmarks to use pinned-pool host allocations by default. ([#15805](https://github.com/rapidsai/cudf/pull/15805)) [@nvdbaranec](https://github.com/nvdbaranec) -- Change thrust::count_if call to raw kernel in strings split APIs ([#15762](https://github.com/rapidsai/cudf/pull/15762)) [@davidwendt](https://github.com/davidwendt) -- Improve performance for long strings for nvtext::replace_tokens ([#15756](https://github.com/rapidsai/cudf/pull/15756)) [@davidwendt](https://github.com/davidwendt) -- Implement chunked parquet reader in cudf-python ([#15728](https://github.com/rapidsai/cudf/pull/15728)) [@galipremsagar](https://github.com/galipremsagar) -- Add `from_arrow_host` functions for cudf interop with nanoarrow ([#15645](https://github.com/rapidsai/cudf/pull/15645)) [@zeroshade](https://github.com/zeroshade) -- Add ability to enable rmm pool on `cudf.pandas` import ([#15628](https://github.com/rapidsai/cudf/pull/15628)) [@galipremsagar](https://github.com/galipremsagar) -- Executor for polars logical plans ([#15504](https://github.com/rapidsai/cudf/pull/15504)) [@wence-](https://github.com/wence-) -- Implement day_name and month_name to match pandas ([#15479](https://github.com/rapidsai/cudf/pull/15479)) [@btepera](https://github.com/btepera) -- Utilities for decimal <--> floating conversion ([#15359](https://github.com/rapidsai/cudf/pull/15359)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- For powers of 10, replace ipow with switch ([#15353](https://github.com/rapidsai/cudf/pull/15353)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Use rapids-build-backend. ([#15245](https://github.com/rapidsai/cudf/pull/15245)) [@vyasr](https://github.com/vyasr) -- Add `codecov` coverage for `pandas_tests` ([#14513](https://github.com/rapidsai/cudf/pull/14513)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure objects with __interface__ are converted to cupy/numpy arrays ([#16436](https://github.com/NVIDIA/cudf/pull/16436)) [@mroeschke](https://github.com/mroeschke) +- Add about rmm modes in `cudf.pandas` docs ([#16404](https://github.com/NVIDIA/cudf/pull/16404)) [@galipremsagar](https://github.com/galipremsagar) +- Gracefully CUDF_FAIL when `skip_rows > 0` in Chunked Parquet reader ([#16385](https://github.com/NVIDIA/cudf/pull/16385)) [@mhaseeb123](https://github.com/mhaseeb123) +- Make C++ compilation warning free after #16297 ([#16379](https://github.com/NVIDIA/cudf/pull/16379)) [@wence-](https://github.com/wence-) +- Align Index __init__ APIs with pandas 2.x ([#16362](https://github.com/NVIDIA/cudf/pull/16362)) [@mroeschke](https://github.com/mroeschke) +- Use rapids_cpm_bs_thread_pool() ([#16360](https://github.com/NVIDIA/cudf/pull/16360)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Rename PrefetchConfig to prefetch_config. ([#16358](https://github.com/NVIDIA/cudf/pull/16358)) [@bdice](https://github.com/bdice) +- Implement parquet reading using pylibcudf in cudf-polars ([#16346](https://github.com/NVIDIA/cudf/pull/16346)) [@lithomas1](https://github.com/lithomas1) +- Fix compile warnings with `jni_utils.hpp` ([#16336](https://github.com/NVIDIA/cudf/pull/16336)) [@ttnghia](https://github.com/ttnghia) +- Align Series APIs with pandas 2.x ([#16333](https://github.com/NVIDIA/cudf/pull/16333)) [@mroeschke](https://github.com/mroeschke) +- Add missing `stream` param to dictionary factory APIs ([#16319](https://github.com/NVIDIA/cudf/pull/16319)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Mark cudf._typing as a typing module in ruff ([#16318](https://github.com/NVIDIA/cudf/pull/16318)) [@mroeschke](https://github.com/mroeschke) +- Add `stream` param to list explode APIs ([#16317](https://github.com/NVIDIA/cudf/pull/16317)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Fix polars for 1.2.1 ([#16316](https://github.com/NVIDIA/cudf/pull/16316)) [@lithomas1](https://github.com/lithomas1) +- Use workflow branch 24.08 again ([#16314](https://github.com/NVIDIA/cudf/pull/16314)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Deprecate dtype= parameter in reduction methods ([#16313](https://github.com/NVIDIA/cudf/pull/16313)) [@mroeschke](https://github.com/mroeschke) +- Remove squeeze argument from groupby ([#16312](https://github.com/NVIDIA/cudf/pull/16312)) [@mroeschke](https://github.com/mroeschke) +- Align more DataFrame APIs with pandas ([#16310](https://github.com/NVIDIA/cudf/pull/16310)) [@mroeschke](https://github.com/mroeschke) +- Clean unneeded/redudant dtype utils ([#16309](https://github.com/NVIDIA/cudf/pull/16309)) [@mroeschke](https://github.com/mroeschke) +- Implement read_csv in cudf-polars using pylibcudf ([#16307](https://github.com/NVIDIA/cudf/pull/16307)) [@lithomas1](https://github.com/lithomas1) +- Use Column.can_cast_safely instead of some ad-hoc dtype functions in .where ([#16303](https://github.com/NVIDIA/cudf/pull/16303)) [@mroeschke](https://github.com/mroeschke) +- Drop `{{ pin_compatible('numpy', max_pin='x') }}` ([#16301](https://github.com/NVIDIA/cudf/pull/16301)) [@jakirkham](https://github.com/jakirkham) +- Host implementation of `to_arrow` using nanoarrow ([#16297](https://github.com/NVIDIA/cudf/pull/16297)) [@zeroshade](https://github.com/zeroshade) +- Add ability to prefetch in `cudf.pandas` and change default to managed pool ([#16296](https://github.com/NVIDIA/cudf/pull/16296)) [@galipremsagar](https://github.com/galipremsagar) +- Fix tests for polars 1.2 ([#16292](https://github.com/NVIDIA/cudf/pull/16292)) [@lithomas1](https://github.com/lithomas1) +- Introduce dedicated options for low memory readers ([#16289](https://github.com/NVIDIA/cudf/pull/16289)) [@galipremsagar](https://github.com/galipremsagar) +- Remove decimal/floating 64/128bit switches due to register pressure ([#16287](https://github.com/NVIDIA/cudf/pull/16287)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Make ColumnAccessor strictly require a mapping of columns ([#16285](https://github.com/NVIDIA/cudf/pull/16285)) [@mroeschke](https://github.com/mroeschke) +- Introduce version file so we can conditionally handle things in tests ([#16280](https://github.com/NVIDIA/cudf/pull/16280)) [@wence-](https://github.com/wence-) +- Type & reduce cupy usage ([#16277](https://github.com/NVIDIA/cudf/pull/16277)) [@mroeschke](https://github.com/mroeschke) +- Update cudf::detail::grid_1d to use thread_index_type ([#16276](https://github.com/NVIDIA/cudf/pull/16276)) [@davidwendt](https://github.com/davidwendt) +- Replace np.isscalar/issubdtype checks with is_scalar/.kind checks ([#16275](https://github.com/NVIDIA/cudf/pull/16275)) [@mroeschke](https://github.com/mroeschke) +- Remove xml from sort_ninja_log.py utility ([#16274](https://github.com/NVIDIA/cudf/pull/16274)) [@davidwendt](https://github.com/davidwendt) +- Fix issue in horizontal concat implementation in cudf-polars ([#16271](https://github.com/NVIDIA/cudf/pull/16271)) [@wence-](https://github.com/wence-) +- Preserve order in left join for cudf-polars ([#16268](https://github.com/NVIDIA/cudf/pull/16268)) [@wence-](https://github.com/wence-) +- Replace is_datetime/timedelta_dtype checks with .kind checks ([#16262](https://github.com/NVIDIA/cudf/pull/16262)) [@mroeschke](https://github.com/mroeschke) +- Replace is_float/integer_dtype checks with .kind checks ([#16261](https://github.com/NVIDIA/cudf/pull/16261)) [@mroeschke](https://github.com/mroeschke) +- Build and test with CUDA 12.5.1 ([#16259](https://github.com/NVIDIA/cudf/pull/16259)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Replace is_bool_type with checking .dtype.kind ([#16255](https://github.com/NVIDIA/cudf/pull/16255)) [@mroeschke](https://github.com/mroeschke) +- remove `cuco_noexcept.diff` ([#16254](https://github.com/NVIDIA/cudf/pull/16254)) [@trxcllnt](https://github.com/trxcllnt) +- Update contains_tests.cpp to use public cudf::slice ([#16253](https://github.com/NVIDIA/cudf/pull/16253)) [@davidwendt](https://github.com/davidwendt) +- Improve the test data for pylibcudf I/O tests ([#16247](https://github.com/NVIDIA/cudf/pull/16247)) [@lithomas1](https://github.com/lithomas1) +- Short circuit some Column methods ([#16246](https://github.com/NVIDIA/cudf/pull/16246)) [@mroeschke](https://github.com/mroeschke) +- Make nvcomp adapter compatible with new version macros ([#16245](https://github.com/NVIDIA/cudf/pull/16245)) [@vuule](https://github.com/vuule) +- Add Column.strftime/strptime instead of overloading `as_string/datetime/timedelta_column` ([#16243](https://github.com/NVIDIA/cudf/pull/16243)) [@mroeschke](https://github.com/mroeschke) +- Remove temporary functor overloads required by cuco version bump ([#16242](https://github.com/NVIDIA/cudf/pull/16242)) [@PointKernel](https://github.com/PointKernel) +- Remove hash_character_ngrams dependency from jaccard_index ([#16241](https://github.com/NVIDIA/cudf/pull/16241)) [@davidwendt](https://github.com/davidwendt) +- Expose sorted groupby parameters to pylibcudf ([#16240](https://github.com/NVIDIA/cudf/pull/16240)) [@wence-](https://github.com/wence-) +- Expose reflection to check if casting between two types is supported ([#16239](https://github.com/NVIDIA/cudf/pull/16239)) [@wence-](https://github.com/wence-) +- Handle nans in groupby-aggregations in polars executor ([#16233](https://github.com/NVIDIA/cudf/pull/16233)) [@wence-](https://github.com/wence-) +- Remove `mr` param from `write_csv` and `write_json` ([#16231](https://github.com/NVIDIA/cudf/pull/16231)) [@JayjeetAtGithub](https://github.com/JayjeetAtGithub) +- Support Literals in groupby-agg ([#16218](https://github.com/NVIDIA/cudf/pull/16218)) [@wence-](https://github.com/wence-) +- Handler csv reader options in cudf-polars ([#16211](https://github.com/NVIDIA/cudf/pull/16211)) [@wence-](https://github.com/wence-) +- Update vendored thread_pool implementation ([#16210](https://github.com/NVIDIA/cudf/pull/16210)) [@wence-](https://github.com/wence-) +- Add low memory JSON reader for `cudf.pandas` ([#16204](https://github.com/NVIDIA/cudf/pull/16204)) [@galipremsagar](https://github.com/galipremsagar) +- Clean up state variables in MultiIndex ([#16203](https://github.com/NVIDIA/cudf/pull/16203)) [@mroeschke](https://github.com/mroeschke) +- skip CMake 3.30.0 ([#16202](https://github.com/NVIDIA/cudf/pull/16202)) [@jameslamb](https://github.com/jameslamb) +- Assert valid metadata is passed in to_arrow for list_view ([#16198](https://github.com/NVIDIA/cudf/pull/16198)) [@wence-](https://github.com/wence-) +- Expose type traits to pylibcudf ([#16197](https://github.com/NVIDIA/cudf/pull/16197)) [@wence-](https://github.com/wence-) +- Report number of rows per file read by PQ reader when no row selection and fix segfault in chunked PQ reader when skip_rows > 0 ([#16195](https://github.com/NVIDIA/cudf/pull/16195)) [@mhaseeb123](https://github.com/mhaseeb123) +- Cast count aggs to correct dtype in translation ([#16192](https://github.com/NVIDIA/cudf/pull/16192)) [@wence-](https://github.com/wence-) +- Some small fixes in cudf-polars ([#16191](https://github.com/NVIDIA/cudf/pull/16191)) [@wence-](https://github.com/wence-) +- split up CUDA-suffixed dependencies in dependencies.yaml ([#16183](https://github.com/NVIDIA/cudf/pull/16183)) [@jameslamb](https://github.com/jameslamb) +- Define PTDS for the stream hook libs ([#16182](https://github.com/NVIDIA/cudf/pull/16182)) [@trxcllnt](https://github.com/trxcllnt) +- Make `test_python_cudf_pandas` generate `requirements.txt` ([#16181](https://github.com/NVIDIA/cudf/pull/16181)) [@trxcllnt](https://github.com/trxcllnt) +- Add environment-agnostic `ci/run_cudf_polars_pytest.sh` ([#16178](https://github.com/NVIDIA/cudf/pull/16178)) [@trxcllnt](https://github.com/trxcllnt) +- Implement translation for some unary functions and a single datetime extraction ([#16173](https://github.com/NVIDIA/cudf/pull/16173)) [@wence-](https://github.com/wence-) +- Remove size constraints on source files in batched JSON reading ([#16162](https://github.com/NVIDIA/cudf/pull/16162)) [@shrshi](https://github.com/shrshi) +- CI: Build wheels for cudf-polars ([#16156](https://github.com/NVIDIA/cudf/pull/16156)) [@lithomas1](https://github.com/lithomas1) +- Update cudf-polars for v1 release of polars ([#16149](https://github.com/NVIDIA/cudf/pull/16149)) [@wence-](https://github.com/wence-) +- Use strings concatenate to support large strings in CSV writer ([#16148](https://github.com/NVIDIA/cudf/pull/16148)) [@davidwendt](https://github.com/davidwendt) +- Use verify-alpha-spec hook ([#16144](https://github.com/NVIDIA/cudf/pull/16144)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Adds write-coalescing code path optimization to FST ([#16143](https://github.com/NVIDIA/cudf/pull/16143)) [@elstehle](https://github.com/elstehle) +- MAINT: Adapt to NumPy 2 promotion changes ([#16141](https://github.com/NVIDIA/cudf/pull/16141)) [@seberg](https://github.com/seberg) +- API: Check for integer overflows when creating scalar form python int ([#16140](https://github.com/NVIDIA/cudf/pull/16140)) [@seberg](https://github.com/seberg) +- Remove the (unused) implementation of `host_parse_nested_json` ([#16135](https://github.com/NVIDIA/cudf/pull/16135)) [@vuule](https://github.com/vuule) +- Deprecate Arrow support in I/O ([#16132](https://github.com/NVIDIA/cudf/pull/16132)) [@lithomas1](https://github.com/lithomas1) +- Disable dict support for split-page kernel in the parquet reader. ([#16128](https://github.com/NVIDIA/cudf/pull/16128)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add throughput metrics for REDUCTION_BENCH/REDUCTION_NVBENCH benchmarks ([#16126](https://github.com/NVIDIA/cudf/pull/16126)) [@jihoonson](https://github.com/jihoonson) +- Add ensure_index to not unnecessarily shallow copy cudf.Index ([#16117](https://github.com/NVIDIA/cudf/pull/16117)) [@mroeschke](https://github.com/mroeschke) +- Make binary operators work between fixed-point and floating args ([#16116](https://github.com/NVIDIA/cudf/pull/16116)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Implement Ternary copy_if_else ([#16114](https://github.com/NVIDIA/cudf/pull/16114)) [@wence-](https://github.com/wence-) +- Implement handlers for series literal in cudf-polars ([#16113](https://github.com/NVIDIA/cudf/pull/16113)) [@wence-](https://github.com/wence-) +- Fix dtype errors in `StringArrays` ([#16111](https://github.com/NVIDIA/cudf/pull/16111)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure MultiIndex.to_frame deep copies columns ([#16110](https://github.com/NVIDIA/cudf/pull/16110)) [@mroeschke](https://github.com/mroeschke) +- Parallelize `gpuInitStringDescriptors` for fixed length byte array data ([#16109](https://github.com/NVIDIA/cudf/pull/16109)) [@mhaseeb123](https://github.com/mhaseeb123) +- Finish implementation of cudf-polars boolean function handlers ([#16098](https://github.com/NVIDIA/cudf/pull/16098)) [@wence-](https://github.com/wence-) +- Expose and then implement support for cross joins in cudf-polars ([#16097](https://github.com/NVIDIA/cudf/pull/16097)) [@wence-](https://github.com/wence-) +- Defer copying in Column.astype(copy=True) ([#16095](https://github.com/NVIDIA/cudf/pull/16095)) [@mroeschke](https://github.com/mroeschke) +- Fix segfault in conditional join ([#16094](https://github.com/NVIDIA/cudf/pull/16094)) [@bdice](https://github.com/bdice) +- Free temp memory no longer needed in multibyte_split processing ([#16091](https://github.com/NVIDIA/cudf/pull/16091)) [@davidwendt](https://github.com/davidwendt) +- Rename gather/scatter benchmarks to clarify coalesced behavior. ([#16083](https://github.com/NVIDIA/cudf/pull/16083)) [@bdice](https://github.com/bdice) +- Adapt to polars upstream changes and turn on CI testing ([#16081](https://github.com/NVIDIA/cudf/pull/16081)) [@wence-](https://github.com/wence-) +- Reduce/clean copy usage in Series, reshaping ([#16080](https://github.com/NVIDIA/cudf/pull/16080)) [@mroeschke](https://github.com/mroeschke) +- Account for FIXED_LEN_BYTE_ARRAY when calculating fragment sizes in Parquet writer ([#16064](https://github.com/NVIDIA/cudf/pull/16064)) [@etseidl](https://github.com/etseidl) +- Reduce (shallow) copies in DataFrame ops ([#16060](https://github.com/NVIDIA/cudf/pull/16060)) [@mroeschke](https://github.com/mroeschke) +- Add multi-file support to `dask_cudf.read_json` ([#16057](https://github.com/NVIDIA/cudf/pull/16057)) [@rjzamora](https://github.com/rjzamora) +- Reduce deep copies in Index ops ([#16054](https://github.com/NVIDIA/cudf/pull/16054)) [@mroeschke](https://github.com/mroeschke) +- Implement chunked column wise concat in chunked parquet reader ([#16052](https://github.com/NVIDIA/cudf/pull/16052)) [@galipremsagar](https://github.com/galipremsagar) +- Add exception when trying to create large strings with cudf::test::strings_column_wrapper ([#16049](https://github.com/NVIDIA/cudf/pull/16049)) [@davidwendt](https://github.com/davidwendt) +- Return `FrozenList` for `Index.names` ([#16047](https://github.com/NVIDIA/cudf/pull/16047)) [@galipremsagar](https://github.com/galipremsagar) +- Add ast cast test ([#16045](https://github.com/NVIDIA/cudf/pull/16045)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Remove `override_dtypes` and `include_index` from `Frame._copy_type_metadata` ([#16043](https://github.com/NVIDIA/cudf/pull/16043)) [@mroeschke](https://github.com/mroeschke) +- Add ruff rules to avoid importing from typing ([#16040](https://github.com/NVIDIA/cudf/pull/16040)) [@mroeschke](https://github.com/mroeschke) +- Fix decimal -> float cast in ast code ([#16038](https://github.com/NVIDIA/cudf/pull/16038)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Add compile option to enable large strings support ([#16037](https://github.com/NVIDIA/cudf/pull/16037)) [@davidwendt](https://github.com/davidwendt) +- Reduce conditional_join nvbench configurations ([#16036](https://github.com/NVIDIA/cudf/pull/16036)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Project automation update: skip if not in project ([#16035](https://github.com/NVIDIA/cudf/pull/16035)) [@jarmak-nv](https://github.com/jarmak-nv) +- Add stream parameter to cudf::io::text::multibyte_split ([#16034](https://github.com/NVIDIA/cudf/pull/16034)) [@davidwendt](https://github.com/davidwendt) +- Delete unused code from stringfunction evaluator ([#16032](https://github.com/NVIDIA/cudf/pull/16032)) [@wence-](https://github.com/wence-) +- Fix exclude regex in pre-commit clang-format hook ([#16030](https://github.com/NVIDIA/cudf/pull/16030)) [@wence-](https://github.com/wence-) +- Refactor rmm usage in `cudf.pandas` ([#16021](https://github.com/NVIDIA/cudf/pull/16021)) [@galipremsagar](https://github.com/galipremsagar) +- Enable ruff TCH: typing imports under if TYPE_CHECKING ([#16015](https://github.com/NVIDIA/cudf/pull/16015)) [@mroeschke](https://github.com/mroeschke) +- Restrict the allowed pandas timezone objects in cudf ([#16013](https://github.com/NVIDIA/cudf/pull/16013)) [@mroeschke](https://github.com/mroeschke) +- orc multithreaded benchmark ([#16009](https://github.com/NVIDIA/cudf/pull/16009)) [@zpuller](https://github.com/zpuller) +- Add tests of expression-based sort and sort-by ([#16008](https://github.com/NVIDIA/cudf/pull/16008)) [@wence-](https://github.com/wence-) +- Add tests of implemented StringFunctions ([#16007](https://github.com/NVIDIA/cudf/pull/16007)) [@wence-](https://github.com/wence-) +- Add test that diagonal concat with mismatching schemas raises ([#16006](https://github.com/NVIDIA/cudf/pull/16006)) [@wence-](https://github.com/wence-) +- Add coverage selecting len from a dataframe (number of rows) ([#16005](https://github.com/NVIDIA/cudf/pull/16005)) [@wence-](https://github.com/wence-) +- Add basic tests of dataframe scan ([#16003](https://github.com/NVIDIA/cudf/pull/16003)) [@wence-](https://github.com/wence-) +- Add coverage for both expression and dataframe filter ([#16002](https://github.com/NVIDIA/cudf/pull/16002)) [@wence-](https://github.com/wence-) +- Remove deprecated ExtContext node ([#16001](https://github.com/NVIDIA/cudf/pull/16001)) [@wence-](https://github.com/wence-) +- Fix typo bug in gather implementation ([#16000](https://github.com/NVIDIA/cudf/pull/16000)) [@wence-](https://github.com/wence-) +- Extend coverage of groupby and rolling window nodes ([#15999](https://github.com/NVIDIA/cudf/pull/15999)) [@wence-](https://github.com/wence-) +- Coverage of binops where one or both operands are a scalar ([#15998](https://github.com/NVIDIA/cudf/pull/15998)) [@wence-](https://github.com/wence-) +- Add full coverage for whole-frame Agg expressions ([#15997](https://github.com/NVIDIA/cudf/pull/15997)) [@wence-](https://github.com/wence-) +- Add tests covering magic methods of Expr objects ([#15996](https://github.com/NVIDIA/cudf/pull/15996)) [@wence-](https://github.com/wence-) +- Add full coverage of utility functions ([#15995](https://github.com/NVIDIA/cudf/pull/15995)) [@wence-](https://github.com/wence-) +- Test behaviour of containers ([#15994](https://github.com/NVIDIA/cudf/pull/15994)) [@wence-](https://github.com/wence-) +- Fix implemention of any, all, and isbetween ([#15993](https://github.com/NVIDIA/cudf/pull/15993)) [@wence-](https://github.com/wence-) +- Raise early on unhandled PythonScan node ([#15992](https://github.com/NVIDIA/cudf/pull/15992)) [@wence-](https://github.com/wence-) +- Remove mapfunction nodes that don't exist/aren't supported ([#15991](https://github.com/NVIDIA/cudf/pull/15991)) [@wence-](https://github.com/wence-) +- Add test coverage for slicing with "out of bounds" negative indices ([#15990](https://github.com/NVIDIA/cudf/pull/15990)) [@wence-](https://github.com/wence-) +- Standardize and type `Series.dt` methods ([#15987](https://github.com/NVIDIA/cudf/pull/15987)) [@mroeschke](https://github.com/mroeschke) +- Refactor distinct with hashset-based algorithms ([#15984](https://github.com/NVIDIA/cudf/pull/15984)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- resolve dependency-file-generator warning, remove unnecessary rapids-build-backend configuration ([#15980](https://github.com/NVIDIA/cudf/pull/15980)) [@jameslamb](https://github.com/jameslamb) +- Project automation bug fixes ([#15971](https://github.com/NVIDIA/cudf/pull/15971)) [@jarmak-nv](https://github.com/jarmak-nv) +- Add typing to single_column_frame ([#15965](https://github.com/NVIDIA/cudf/pull/15965)) [@mroeschke](https://github.com/mroeschke) +- Move some misc Frame methods to appropriate locations ([#15963](https://github.com/NVIDIA/cudf/pull/15963)) [@mroeschke](https://github.com/mroeschke) +- Condense pylibcudf data fixtures ([#15958](https://github.com/NVIDIA/cudf/pull/15958)) [@lithomas1](https://github.com/lithomas1) +- Refactor fillna logic to push specifics toward Frame subclasses and Column subclasses ([#15957](https://github.com/NVIDIA/cudf/pull/15957)) [@mroeschke](https://github.com/mroeschke) +- Remove unused parsing utilities ([#15955](https://github.com/NVIDIA/cudf/pull/15955)) [@vuule](https://github.com/vuule) +- Remove `Scalar` container type from polars interpreter ([#15953](https://github.com/NVIDIA/cudf/pull/15953)) [@wence-](https://github.com/wence-) +- Support arbitrary CUDA versions in UDF code ([#15950](https://github.com/NVIDIA/cudf/pull/15950)) [@bdice](https://github.com/bdice) +- Support large strings in cudf::io::text::multibyte_split ([#15947](https://github.com/NVIDIA/cudf/pull/15947)) [@davidwendt](https://github.com/davidwendt) +- Add external issue label and project automation ([#15945](https://github.com/NVIDIA/cudf/pull/15945)) [@jarmak-nv](https://github.com/jarmak-nv) +- Enable round-tripping of large strings in `cudf` ([#15944](https://github.com/NVIDIA/cudf/pull/15944)) [@galipremsagar](https://github.com/galipremsagar) +- Add more complete type annotations in polars interpreter ([#15942](https://github.com/NVIDIA/cudf/pull/15942)) [@wence-](https://github.com/wence-) +- Update implementations to build with the latest cuco ([#15938](https://github.com/NVIDIA/cudf/pull/15938)) [@PointKernel](https://github.com/PointKernel) +- Support timezone aware pandas inputs in cudf ([#15935](https://github.com/NVIDIA/cudf/pull/15935)) [@mroeschke](https://github.com/mroeschke) +- Define Column.nan_as_null to return self ([#15923](https://github.com/NVIDIA/cudf/pull/15923)) [@mroeschke](https://github.com/mroeschke) +- Make Frame._dtype an iterator instead of a dict ([#15920](https://github.com/NVIDIA/cudf/pull/15920)) [@mroeschke](https://github.com/mroeschke) +- Port start of datetime.hpp to pylibcudf ([#15916](https://github.com/NVIDIA/cudf/pull/15916)) [@wence-](https://github.com/wence-) +- Introduce `NamedColumn` concept in cudf-polars ([#15914](https://github.com/NVIDIA/cudf/pull/15914)) [@wence-](https://github.com/wence-) +- Avoid redefining Frame._get_columns_by_label in subclasses ([#15912](https://github.com/NVIDIA/cudf/pull/15912)) [@mroeschke](https://github.com/mroeschke) +- Templatization of fixed-width parquet decoding kernels. ([#15911](https://github.com/NVIDIA/cudf/pull/15911)) [@nvdbaranec](https://github.com/nvdbaranec) +- New Decimal <--> Floating conversion ([#15905](https://github.com/NVIDIA/cudf/pull/15905)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Use Arrow C Data Interface functions for Python interop ([#15904](https://github.com/NVIDIA/cudf/pull/15904)) [@vyasr](https://github.com/vyasr) +- Use offsetalator in cudf::io::json::detail::parse_string ([#15900](https://github.com/NVIDIA/cudf/pull/15900)) [@davidwendt](https://github.com/davidwendt) +- Rename strings multiple target replace API ([#15898](https://github.com/NVIDIA/cudf/pull/15898)) [@davidwendt](https://github.com/davidwendt) +- Apply clang-tidy autofixes ([#15894](https://github.com/NVIDIA/cudf/pull/15894)) [@vyasr](https://github.com/vyasr) +- Update Python labels and remove unnecessary ones ([#15893](https://github.com/NVIDIA/cudf/pull/15893)) [@vyasr](https://github.com/vyasr) +- Clean up pylibcudf test assertations ([#15892](https://github.com/NVIDIA/cudf/pull/15892)) [@lithomas1](https://github.com/lithomas1) +- Use offsetalator in orc rowgroup_char_counts_kernel ([#15891](https://github.com/NVIDIA/cudf/pull/15891)) [@davidwendt](https://github.com/davidwendt) +- Ensure literals have correct dtype ([#15890](https://github.com/NVIDIA/cudf/pull/15890)) [@wence-](https://github.com/wence-) +- Add overflow check when converting large strings to lists columns ([#15887](https://github.com/NVIDIA/cudf/pull/15887)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in nvtext::tokenize_with_vocabulary ([#15878](https://github.com/NVIDIA/cudf/pull/15878)) [@davidwendt](https://github.com/davidwendt) +- Update interleave lists column for large strings ([#15877](https://github.com/NVIDIA/cudf/pull/15877)) [@davidwendt](https://github.com/davidwendt) +- Simple NumPy 2 fixes that are clearly no behavior change ([#15876](https://github.com/NVIDIA/cudf/pull/15876)) [@seberg](https://github.com/seberg) +- Support `arrow:schema` in Parquet writer to faithfully roundtrip `duration` types with Arrow ([#15875](https://github.com/NVIDIA/cudf/pull/15875)) [@mhaseeb123](https://github.com/mhaseeb123) +- Refactor join benchmarks to target public APIs with the default stream ([#15873](https://github.com/NVIDIA/cudf/pull/15873)) [@PointKernel](https://github.com/PointKernel) +- Fix url-decode benchmark to use offsetalator ([#15871](https://github.com/NVIDIA/cudf/pull/15871)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in strings shift functor ([#15870](https://github.com/NVIDIA/cudf/pull/15870)) [@davidwendt](https://github.com/davidwendt) +- Memory Profiling ([#15866](https://github.com/NVIDIA/cudf/pull/15866)) [@madsbk](https://github.com/madsbk) +- Expose stream parameter to public rolling APIs ([#15865](https://github.com/NVIDIA/cudf/pull/15865)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- Make Frame.astype return Self instead of a ColumnAccessor ([#15861](https://github.com/NVIDIA/cudf/pull/15861)) [@mroeschke](https://github.com/mroeschke) +- Use ColumnAccessor row and column length attributes more consistently ([#15857](https://github.com/NVIDIA/cudf/pull/15857)) [@mroeschke](https://github.com/mroeschke) +- add unit test setup for cudf_kafka ([#15853](https://github.com/NVIDIA/cudf/pull/15853)) [@jameslamb](https://github.com/jameslamb) +- Remove internal usage of core.index.as_index in favor of cudf.Index ([#15851](https://github.com/NVIDIA/cudf/pull/15851)) [@mroeschke](https://github.com/mroeschke) +- Ensure cudf.Series(cudf.Series(...)) creates a reference to the same index ([#15845](https://github.com/NVIDIA/cudf/pull/15845)) [@mroeschke](https://github.com/mroeschke) +- Remove benchmark-specific use of pinned-pooled memory in Parquet multithreaded benchmark. ([#15838](https://github.com/NVIDIA/cudf/pull/15838)) [@nvdbaranec](https://github.com/nvdbaranec) +- Implement `on_bad_lines` in json reader ([#15834](https://github.com/NVIDIA/cudf/pull/15834)) [@galipremsagar](https://github.com/galipremsagar) +- Make Column.to_pandas return Index instead of Series ([#15833](https://github.com/NVIDIA/cudf/pull/15833)) [@mroeschke](https://github.com/mroeschke) +- Add test of interoperability of cuDF and arrow BYTE_STREAM_SPLIT encoders ([#15832](https://github.com/NVIDIA/cudf/pull/15832)) [@etseidl](https://github.com/etseidl) +- Refactor Parquet writer options and builders ([#15831](https://github.com/NVIDIA/cudf/pull/15831)) [@etseidl](https://github.com/etseidl) +- Migrate reshape.pxd to pylibcudf ([#15827](https://github.com/NVIDIA/cudf/pull/15827)) [@lithomas1](https://github.com/lithomas1) +- Remove legacy JSON reader and concurrent_unordered_map.cuh. ([#15813](https://github.com/NVIDIA/cudf/pull/15813)) [@bdice](https://github.com/bdice) +- Switch cuIO benchmarks to use pinned-pool host allocations by default. ([#15805](https://github.com/NVIDIA/cudf/pull/15805)) [@nvdbaranec](https://github.com/nvdbaranec) +- Change thrust::count_if call to raw kernel in strings split APIs ([#15762](https://github.com/NVIDIA/cudf/pull/15762)) [@davidwendt](https://github.com/davidwendt) +- Improve performance for long strings for nvtext::replace_tokens ([#15756](https://github.com/NVIDIA/cudf/pull/15756)) [@davidwendt](https://github.com/davidwendt) +- Implement chunked parquet reader in cudf-python ([#15728](https://github.com/NVIDIA/cudf/pull/15728)) [@galipremsagar](https://github.com/galipremsagar) +- Add `from_arrow_host` functions for cudf interop with nanoarrow ([#15645](https://github.com/NVIDIA/cudf/pull/15645)) [@zeroshade](https://github.com/zeroshade) +- Add ability to enable rmm pool on `cudf.pandas` import ([#15628](https://github.com/NVIDIA/cudf/pull/15628)) [@galipremsagar](https://github.com/galipremsagar) +- Executor for polars logical plans ([#15504](https://github.com/NVIDIA/cudf/pull/15504)) [@wence-](https://github.com/wence-) +- Implement day_name and month_name to match pandas ([#15479](https://github.com/NVIDIA/cudf/pull/15479)) [@btepera](https://github.com/btepera) +- Utilities for decimal <--> floating conversion ([#15359](https://github.com/NVIDIA/cudf/pull/15359)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- For powers of 10, replace ipow with switch ([#15353](https://github.com/NVIDIA/cudf/pull/15353)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Use rapids-build-backend. ([#15245](https://github.com/NVIDIA/cudf/pull/15245)) [@vyasr](https://github.com/vyasr) +- Add `codecov` coverage for `pandas_tests` ([#14513](https://github.com/NVIDIA/cudf/pull/14513)) [@galipremsagar](https://github.com/galipremsagar) # cudf 24.06.00 (5 Jun 2024) ## 🚨 Breaking Changes -- Deprecate `Groupby.collect` ([#15808](https://github.com/rapidsai/cudf/pull/15808)) [@galipremsagar](https://github.com/galipremsagar) -- Raise FileNotFoundError when a literal JSON string that looks like a json filename is passed ([#15806](https://github.com/rapidsai/cudf/pull/15806)) [@lithomas1](https://github.com/lithomas1) -- Support filtered I/O in `chunked_parquet_reader` and simplify the use of `parquet_reader_options` ([#15764](https://github.com/rapidsai/cudf/pull/15764)) [@mhaseeb123](https://github.com/mhaseeb123) -- Raise errors for unsupported operations on certain types ([#15712](https://github.com/rapidsai/cudf/pull/15712)) [@galipremsagar](https://github.com/galipremsagar) -- Support `DurationType` in cudf parquet reader via `arrow:schema` ([#15617](https://github.com/rapidsai/cudf/pull/15617)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove protobuf and use parsed ORC statistics from libcudf ([#15564](https://github.com/rapidsai/cudf/pull/15564)) [@bdice](https://github.com/bdice) -- Remove legacy JSON reader from Python ([#15538](https://github.com/rapidsai/cudf/pull/15538)) [@bdice](https://github.com/bdice) -- Removing all batching code from parquet writer ([#15528](https://github.com/rapidsai/cudf/pull/15528)) [@mhaseeb123](https://github.com/mhaseeb123) -- Convert libcudf resource parameters to rmm::device_async_resource_ref ([#15507](https://github.com/rapidsai/cudf/pull/15507)) [@harrism](https://github.com/harrism) -- Remove deprecated strings offsets_begin ([#15454](https://github.com/rapidsai/cudf/pull/15454)) [@davidwendt](https://github.com/davidwendt) -- Floating <--> fixed-point conversion must now be called explicitly ([#15438](https://github.com/rapidsai/cudf/pull/15438)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Bind `read_parquet_metadata` API to libcudf instead of pyarrow and extract `RowGroup` information ([#15398](https://github.com/rapidsai/cudf/pull/15398)) [@mhaseeb123](https://github.com/mhaseeb123) -- Remove deprecated hash() and spark_murmurhash3_x86_32() ([#15375](https://github.com/rapidsai/cudf/pull/15375)) [@davidwendt](https://github.com/davidwendt) -- Remove empty elements from exploded character-ngrams output ([#15371](https://github.com/rapidsai/cudf/pull/15371)) [@davidwendt](https://github.com/davidwendt) -- [FEA] Performance improvement for mixed left semi/anti join ([#15288](https://github.com/rapidsai/cudf/pull/15288)) [@tgujar](https://github.com/tgujar) -- Align date_range defaults with pandas, support tz ([#15139](https://github.com/rapidsai/cudf/pull/15139)) [@mroeschke](https://github.com/mroeschke) +- Deprecate `Groupby.collect` ([#15808](https://github.com/NVIDIA/cudf/pull/15808)) [@galipremsagar](https://github.com/galipremsagar) +- Raise FileNotFoundError when a literal JSON string that looks like a json filename is passed ([#15806](https://github.com/NVIDIA/cudf/pull/15806)) [@lithomas1](https://github.com/lithomas1) +- Support filtered I/O in `chunked_parquet_reader` and simplify the use of `parquet_reader_options` ([#15764](https://github.com/NVIDIA/cudf/pull/15764)) [@mhaseeb123](https://github.com/mhaseeb123) +- Raise errors for unsupported operations on certain types ([#15712](https://github.com/NVIDIA/cudf/pull/15712)) [@galipremsagar](https://github.com/galipremsagar) +- Support `DurationType` in cudf parquet reader via `arrow:schema` ([#15617](https://github.com/NVIDIA/cudf/pull/15617)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove protobuf and use parsed ORC statistics from libcudf ([#15564](https://github.com/NVIDIA/cudf/pull/15564)) [@bdice](https://github.com/bdice) +- Remove legacy JSON reader from Python ([#15538](https://github.com/NVIDIA/cudf/pull/15538)) [@bdice](https://github.com/bdice) +- Removing all batching code from parquet writer ([#15528](https://github.com/NVIDIA/cudf/pull/15528)) [@mhaseeb123](https://github.com/mhaseeb123) +- Convert libcudf resource parameters to rmm::device_async_resource_ref ([#15507](https://github.com/NVIDIA/cudf/pull/15507)) [@harrism](https://github.com/harrism) +- Remove deprecated strings offsets_begin ([#15454](https://github.com/NVIDIA/cudf/pull/15454)) [@davidwendt](https://github.com/davidwendt) +- Floating <--> fixed-point conversion must now be called explicitly ([#15438](https://github.com/NVIDIA/cudf/pull/15438)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Bind `read_parquet_metadata` API to libcudf instead of pyarrow and extract `RowGroup` information ([#15398](https://github.com/NVIDIA/cudf/pull/15398)) [@mhaseeb123](https://github.com/mhaseeb123) +- Remove deprecated hash() and spark_murmurhash3_x86_32() ([#15375](https://github.com/NVIDIA/cudf/pull/15375)) [@davidwendt](https://github.com/davidwendt) +- Remove empty elements from exploded character-ngrams output ([#15371](https://github.com/NVIDIA/cudf/pull/15371)) [@davidwendt](https://github.com/davidwendt) +- [FEA] Performance improvement for mixed left semi/anti join ([#15288](https://github.com/NVIDIA/cudf/pull/15288)) [@tgujar](https://github.com/tgujar) +- Align date_range defaults with pandas, support tz ([#15139](https://github.com/NVIDIA/cudf/pull/15139)) [@mroeschke](https://github.com/mroeschke) ## 🐛 Bug Fixes -- Revert "Fix docs for IO readers and strings_convert" ([#15872](https://github.com/rapidsai/cudf/pull/15872)) [@vyasr](https://github.com/vyasr) -- Remove problematic call of index setter to unblock dask-cuda CI ([#15844](https://github.com/rapidsai/cudf/pull/15844)) [@charlesbluca](https://github.com/charlesbluca) -- Use rapids_cpm_nvtx3 to get same nvtx3 target state as rmm ([#15840](https://github.com/rapidsai/cudf/pull/15840)) [@robertmaynard](https://github.com/robertmaynard) -- Return boolean from config_host_memory_resource instead of throwing ([#15815](https://github.com/rapidsai/cudf/pull/15815)) [@abellina](https://github.com/abellina) -- Add temporary dask-cudf workaround for categorical sorting ([#15801](https://github.com/rapidsai/cudf/pull/15801)) [@rjzamora](https://github.com/rjzamora) -- Fix row group alignment in ORC writer ([#15789](https://github.com/rapidsai/cudf/pull/15789)) [@vuule](https://github.com/vuule) -- Raise error when sorting by categorical column in dask-cudf ([#15788](https://github.com/rapidsai/cudf/pull/15788)) [@rjzamora](https://github.com/rjzamora) -- Upgrade `arrow` to 16.1 ([#15787](https://github.com/rapidsai/cudf/pull/15787)) [@galipremsagar](https://github.com/galipremsagar) -- Add support for `PandasArray` for `pandas<2.1.0` ([#15786](https://github.com/rapidsai/cudf/pull/15786)) [@galipremsagar](https://github.com/galipremsagar) -- Limit runtime dependency to `libarrow>=16.0.0,<16.1.0a0` ([#15782](https://github.com/rapidsai/cudf/pull/15782)) [@pentschev](https://github.com/pentschev) -- Fix cat.as_ordered not propogating correct size ([#15780](https://github.com/rapidsai/cudf/pull/15780)) [@mroeschke](https://github.com/mroeschke) -- Handle mixed-like homogeneous types in `isin` ([#15771](https://github.com/rapidsai/cudf/pull/15771)) [@galipremsagar](https://github.com/galipremsagar) -- Fix id_vars and value_vars not accepting string scalars in melt ([#15765](https://github.com/rapidsai/cudf/pull/15765)) [@mroeschke](https://github.com/mroeschke) -- Fix `DatetimeIndex.loc` for all types of ordering cases ([#15761](https://github.com/rapidsai/cudf/pull/15761)) [@galipremsagar](https://github.com/galipremsagar) -- Fix arrow versioning logic ([#15755](https://github.com/rapidsai/cudf/pull/15755)) [@vyasr](https://github.com/vyasr) -- Avoid running sanitizer on Java test designed to cause an error ([#15753](https://github.com/rapidsai/cudf/pull/15753)) [@jlowe](https://github.com/jlowe) -- Handle empty dataframe object with index present in setitem of `loc` ([#15752](https://github.com/rapidsai/cudf/pull/15752)) [@galipremsagar](https://github.com/galipremsagar) -- Eliminate circular reference in DataFrame/Series.iloc/loc ([#15749](https://github.com/rapidsai/cudf/pull/15749)) [@mroeschke](https://github.com/mroeschke) -- Cap the absolute row index per pass in parquet chunked reader. ([#15735](https://github.com/rapidsai/cudf/pull/15735)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix `Index.repeat` for `datetime64` types ([#15722](https://github.com/rapidsai/cudf/pull/15722)) [@galipremsagar](https://github.com/galipremsagar) -- Fix multibyte check for case convert for large strings ([#15721](https://github.com/rapidsai/cudf/pull/15721)) [@davidwendt](https://github.com/davidwendt) -- Fix `get_loc` to properly fetch results from an index that is in decreasing order ([#15719](https://github.com/rapidsai/cudf/pull/15719)) [@galipremsagar](https://github.com/galipremsagar) -- Return same type as the original index for `.loc` operations ([#15717](https://github.com/rapidsai/cudf/pull/15717)) [@galipremsagar](https://github.com/galipremsagar) -- Correct static builds + static arrow ([#15715](https://github.com/rapidsai/cudf/pull/15715)) [@robertmaynard](https://github.com/robertmaynard) -- Raise errors for unsupported operations on certain types ([#15712](https://github.com/rapidsai/cudf/pull/15712)) [@galipremsagar](https://github.com/galipremsagar) -- Fix ColumnAccessor caching of nrows if empty previously ([#15710](https://github.com/rapidsai/cudf/pull/15710)) [@mroeschke](https://github.com/mroeschke) -- Allow `None` when `nan_as_null=False` in column constructor ([#15709](https://github.com/rapidsai/cudf/pull/15709)) [@galipremsagar](https://github.com/galipremsagar) -- Refine `CudaTest.testCudaException` in case throwing wrong type of CudaError under aarch64 ([#15706](https://github.com/rapidsai/cudf/pull/15706)) [@sperlingxx](https://github.com/sperlingxx) -- Fix maxima of categorical column ([#15701](https://github.com/rapidsai/cudf/pull/15701)) [@rjzamora](https://github.com/rjzamora) -- Add proxy for inplace operations in `cudf.pandas` ([#15695](https://github.com/rapidsai/cudf/pull/15695)) [@galipremsagar](https://github.com/galipremsagar) -- Make `nan_as_null` behavior consistent across all APIs ([#15692](https://github.com/rapidsai/cudf/pull/15692)) [@galipremsagar](https://github.com/galipremsagar) -- Fix CI s3 api command to fetch latest results ([#15687](https://github.com/rapidsai/cudf/pull/15687)) [@galipremsagar](https://github.com/galipremsagar) -- Add `NumpyExtensionArray` proxy type in `cudf.pandas` ([#15686](https://github.com/rapidsai/cudf/pull/15686)) [@galipremsagar](https://github.com/galipremsagar) -- Properly implement binaryops for proxy types ([#15684](https://github.com/rapidsai/cudf/pull/15684)) [@galipremsagar](https://github.com/galipremsagar) -- Fix copy assignment and the comparison operator of `rmm_host_allocator` ([#15677](https://github.com/rapidsai/cudf/pull/15677)) [@vuule](https://github.com/vuule) -- Fix multi-source reading in JSON byte range reader ([#15671](https://github.com/rapidsai/cudf/pull/15671)) [@shrshi](https://github.com/shrshi) -- Return `int64` when pandas compatible mode is turned on for `get_indexer` ([#15659](https://github.com/rapidsai/cudf/pull/15659)) [@galipremsagar](https://github.com/galipremsagar) -- Fix Index contains for error validations and float vs int comparisons ([#15657](https://github.com/rapidsai/cudf/pull/15657)) [@galipremsagar](https://github.com/galipremsagar) -- Preserve sub-second data for time scalars in column construction ([#15655](https://github.com/rapidsai/cudf/pull/15655)) [@galipremsagar](https://github.com/galipremsagar) -- Check row limit size in cudf::strings::join_strings ([#15643](https://github.com/rapidsai/cudf/pull/15643)) [@davidwendt](https://github.com/davidwendt) -- Enable sorting on column with nulls using query-planning ([#15639](https://github.com/rapidsai/cudf/pull/15639)) [@rjzamora](https://github.com/rjzamora) -- Fix operator precedence problem in Parquet reader ([#15638](https://github.com/rapidsai/cudf/pull/15638)) [@etseidl](https://github.com/etseidl) -- Fix decoding of dictionary encoded FIXED_LEN_BYTE_ARRAY data in Parquet reader ([#15601](https://github.com/rapidsai/cudf/pull/15601)) [@etseidl](https://github.com/etseidl) -- Fix debug warnings/errors in from_arrow_device_test.cpp ([#15596](https://github.com/rapidsai/cudf/pull/15596)) [@davidwendt](https://github.com/davidwendt) -- Add "collect" aggregation support to dask-cudf ([#15593](https://github.com/rapidsai/cudf/pull/15593)) [@rjzamora](https://github.com/rjzamora) -- Fix categorical-accessor support and testing in dask-cudf ([#15591](https://github.com/rapidsai/cudf/pull/15591)) [@rjzamora](https://github.com/rjzamora) -- Disable compute-sanitizer usage in CI tests with CUDA<11.6 ([#15584](https://github.com/rapidsai/cudf/pull/15584)) [@davidwendt](https://github.com/davidwendt) -- Preserve RangeIndex.step in to_arrow/from_arrow ([#15581](https://github.com/rapidsai/cudf/pull/15581)) [@mroeschke](https://github.com/mroeschke) -- Ignore new cupy warning ([#15574](https://github.com/rapidsai/cudf/pull/15574)) [@vyasr](https://github.com/vyasr) -- Add cuda-sanitizer-api dependency for test-cpp matrix 11.4 ([#15573](https://github.com/rapidsai/cudf/pull/15573)) [@davidwendt](https://github.com/davidwendt) -- Allow apply udf to reference global modules in cudf.pandas ([#15569](https://github.com/rapidsai/cudf/pull/15569)) [@mroeschke](https://github.com/mroeschke) -- Fix deprecation warnings for json legacy reader ([#15563](https://github.com/rapidsai/cudf/pull/15563)) [@davidwendt](https://github.com/davidwendt) -- Fix millisecond resampling in cudf Python ([#15560](https://github.com/rapidsai/cudf/pull/15560)) [@mroeschke](https://github.com/mroeschke) -- Rename JSON_READER_OPTION to JSON_READER_OPTION_NVBENCH. ([#15553](https://github.com/rapidsai/cudf/pull/15553)) [@bdice](https://github.com/bdice) -- Fix a JNI bug in JSON parsing fixup ([#15550](https://github.com/rapidsai/cudf/pull/15550)) [@revans2](https://github.com/revans2) -- Remove conda channel setup from wheel CI image script. ([#15539](https://github.com/rapidsai/cudf/pull/15539)) [@bdice](https://github.com/bdice) -- cudf.pandas: Series dt accessor is CombinedDatetimelikeProperties ([#15523](https://github.com/rapidsai/cudf/pull/15523)) [@wence-](https://github.com/wence-) -- Fix for some compiler warnings in parquet/page_decode.cuh ([#15518](https://github.com/rapidsai/cudf/pull/15518)) [@etseidl](https://github.com/etseidl) -- Fix exponent overflow in strings-to-double conversion ([#15517](https://github.com/rapidsai/cudf/pull/15517)) [@davidwendt](https://github.com/davidwendt) -- nanoarrow uses package override for proper pinned versions generation ([#15515](https://github.com/rapidsai/cudf/pull/15515)) [@robertmaynard](https://github.com/robertmaynard) -- Remove index name overrides in dask-cudf pyarrow table dispatch ([#15514](https://github.com/rapidsai/cudf/pull/15514)) [@charlesbluca](https://github.com/charlesbluca) -- Fix async synchronization issues in json_column.cu ([#15497](https://github.com/rapidsai/cudf/pull/15497)) [@karthikeyann](https://github.com/karthikeyann) -- Add new patch to hide more CCCL APIs ([#15493](https://github.com/rapidsai/cudf/pull/15493)) [@vyasr](https://github.com/vyasr) -- Make improvements in pandas-test reporting ([#15485](https://github.com/rapidsai/cudf/pull/15485)) [@galipremsagar](https://github.com/galipremsagar) -- Fixed page data truncation in parquet writer under certain conditions. ([#15474](https://github.com/rapidsai/cudf/pull/15474)) [@nvdbaranec](https://github.com/nvdbaranec) -- Only use data_type constructor with scale for decimal types ([#15472](https://github.com/rapidsai/cudf/pull/15472)) [@wence-](https://github.com/wence-) -- Avoid "p2p" shuffle as a default when `dask_cudf` is imported ([#15469](https://github.com/rapidsai/cudf/pull/15469)) [@rjzamora](https://github.com/rjzamora) -- Fix debug build errors from to_arrow_device_test.cpp ([#15463](https://github.com/rapidsai/cudf/pull/15463)) [@davidwendt](https://github.com/davidwendt) -- Fix base_normalator::integer_sizeof_fn integer dispatch ([#15457](https://github.com/rapidsai/cudf/pull/15457)) [@davidwendt](https://github.com/davidwendt) -- Allow consumers of static builds to find nanoarrow ([#15456](https://github.com/rapidsai/cudf/pull/15456)) [@robertmaynard](https://github.com/robertmaynard) -- Allow jit compilation when using a splayed CUDA toolkit ([#15451](https://github.com/rapidsai/cudf/pull/15451)) [@robertmaynard](https://github.com/robertmaynard) -- Handle case of scan aggregation in groupby-transform ([#15450](https://github.com/rapidsai/cudf/pull/15450)) [@wence-](https://github.com/wence-) -- Test static builds in CI and fix nanoarrow configure ([#15437](https://github.com/rapidsai/cudf/pull/15437)) [@vyasr](https://github.com/vyasr) -- Fixes potential race in JSON parser when parsing JSON lines format and when recovering from invalid lines ([#15419](https://github.com/rapidsai/cudf/pull/15419)) [@elstehle](https://github.com/elstehle) -- Fix errors in chunked ORC writer when no tables were (successfully) written ([#15393](https://github.com/rapidsai/cudf/pull/15393)) [@vuule](https://github.com/vuule) -- Support implicit array conversion with query-planning enabled ([#15378](https://github.com/rapidsai/cudf/pull/15378)) [@rjzamora](https://github.com/rjzamora) -- Fix arrow-based round trip of empty dataframes ([#15373](https://github.com/rapidsai/cudf/pull/15373)) [@wence-](https://github.com/wence-) -- Remove empty elements from exploded character-ngrams output ([#15371](https://github.com/rapidsai/cudf/pull/15371)) [@davidwendt](https://github.com/davidwendt) -- Remove boundscheck=False setting in cython files ([#15362](https://github.com/rapidsai/cudf/pull/15362)) [@wence-](https://github.com/wence-) -- Patch dask-expr `var` logic in dask-cudf ([#15347](https://github.com/rapidsai/cudf/pull/15347)) [@rjzamora](https://github.com/rjzamora) -- Fix for logical and syntactical errors in libcudf c++ examples ([#15346](https://github.com/rapidsai/cudf/pull/15346)) [@mhaseeb123](https://github.com/mhaseeb123) -- Disable dask-expr in docs builds. ([#15343](https://github.com/rapidsai/cudf/pull/15343)) [@bdice](https://github.com/bdice) -- Apply the cuFile error work around to data_sink as well ([#15335](https://github.com/rapidsai/cudf/pull/15335)) [@vuule](https://github.com/vuule) -- Fix parquet predicate filtering with column projection ([#15113](https://github.com/rapidsai/cudf/pull/15113)) [@karthikeyann](https://github.com/karthikeyann) -- Check column type equality, handling nested types correctly. ([#14531](https://github.com/rapidsai/cudf/pull/14531)) [@bdice](https://github.com/bdice) +- Revert "Fix docs for IO readers and strings_convert" ([#15872](https://github.com/NVIDIA/cudf/pull/15872)) [@vyasr](https://github.com/vyasr) +- Remove problematic call of index setter to unblock dask-cuda CI ([#15844](https://github.com/NVIDIA/cudf/pull/15844)) [@charlesbluca](https://github.com/charlesbluca) +- Use rapids_cpm_nvtx3 to get same nvtx3 target state as rmm ([#15840](https://github.com/NVIDIA/cudf/pull/15840)) [@robertmaynard](https://github.com/robertmaynard) +- Return boolean from config_host_memory_resource instead of throwing ([#15815](https://github.com/NVIDIA/cudf/pull/15815)) [@abellina](https://github.com/abellina) +- Add temporary dask-cudf workaround for categorical sorting ([#15801](https://github.com/NVIDIA/cudf/pull/15801)) [@rjzamora](https://github.com/rjzamora) +- Fix row group alignment in ORC writer ([#15789](https://github.com/NVIDIA/cudf/pull/15789)) [@vuule](https://github.com/vuule) +- Raise error when sorting by categorical column in dask-cudf ([#15788](https://github.com/NVIDIA/cudf/pull/15788)) [@rjzamora](https://github.com/rjzamora) +- Upgrade `arrow` to 16.1 ([#15787](https://github.com/NVIDIA/cudf/pull/15787)) [@galipremsagar](https://github.com/galipremsagar) +- Add support for `PandasArray` for `pandas<2.1.0` ([#15786](https://github.com/NVIDIA/cudf/pull/15786)) [@galipremsagar](https://github.com/galipremsagar) +- Limit runtime dependency to `libarrow>=16.0.0,<16.1.0a0` ([#15782](https://github.com/NVIDIA/cudf/pull/15782)) [@pentschev](https://github.com/pentschev) +- Fix cat.as_ordered not propogating correct size ([#15780](https://github.com/NVIDIA/cudf/pull/15780)) [@mroeschke](https://github.com/mroeschke) +- Handle mixed-like homogeneous types in `isin` ([#15771](https://github.com/NVIDIA/cudf/pull/15771)) [@galipremsagar](https://github.com/galipremsagar) +- Fix id_vars and value_vars not accepting string scalars in melt ([#15765](https://github.com/NVIDIA/cudf/pull/15765)) [@mroeschke](https://github.com/mroeschke) +- Fix `DatetimeIndex.loc` for all types of ordering cases ([#15761](https://github.com/NVIDIA/cudf/pull/15761)) [@galipremsagar](https://github.com/galipremsagar) +- Fix arrow versioning logic ([#15755](https://github.com/NVIDIA/cudf/pull/15755)) [@vyasr](https://github.com/vyasr) +- Avoid running sanitizer on Java test designed to cause an error ([#15753](https://github.com/NVIDIA/cudf/pull/15753)) [@jlowe](https://github.com/jlowe) +- Handle empty dataframe object with index present in setitem of `loc` ([#15752](https://github.com/NVIDIA/cudf/pull/15752)) [@galipremsagar](https://github.com/galipremsagar) +- Eliminate circular reference in DataFrame/Series.iloc/loc ([#15749](https://github.com/NVIDIA/cudf/pull/15749)) [@mroeschke](https://github.com/mroeschke) +- Cap the absolute row index per pass in parquet chunked reader. ([#15735](https://github.com/NVIDIA/cudf/pull/15735)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix `Index.repeat` for `datetime64` types ([#15722](https://github.com/NVIDIA/cudf/pull/15722)) [@galipremsagar](https://github.com/galipremsagar) +- Fix multibyte check for case convert for large strings ([#15721](https://github.com/NVIDIA/cudf/pull/15721)) [@davidwendt](https://github.com/davidwendt) +- Fix `get_loc` to properly fetch results from an index that is in decreasing order ([#15719](https://github.com/NVIDIA/cudf/pull/15719)) [@galipremsagar](https://github.com/galipremsagar) +- Return same type as the original index for `.loc` operations ([#15717](https://github.com/NVIDIA/cudf/pull/15717)) [@galipremsagar](https://github.com/galipremsagar) +- Correct static builds + static arrow ([#15715](https://github.com/NVIDIA/cudf/pull/15715)) [@robertmaynard](https://github.com/robertmaynard) +- Raise errors for unsupported operations on certain types ([#15712](https://github.com/NVIDIA/cudf/pull/15712)) [@galipremsagar](https://github.com/galipremsagar) +- Fix ColumnAccessor caching of nrows if empty previously ([#15710](https://github.com/NVIDIA/cudf/pull/15710)) [@mroeschke](https://github.com/mroeschke) +- Allow `None` when `nan_as_null=False` in column constructor ([#15709](https://github.com/NVIDIA/cudf/pull/15709)) [@galipremsagar](https://github.com/galipremsagar) +- Refine `CudaTest.testCudaException` in case throwing wrong type of CudaError under aarch64 ([#15706](https://github.com/NVIDIA/cudf/pull/15706)) [@sperlingxx](https://github.com/sperlingxx) +- Fix maxima of categorical column ([#15701](https://github.com/NVIDIA/cudf/pull/15701)) [@rjzamora](https://github.com/rjzamora) +- Add proxy for inplace operations in `cudf.pandas` ([#15695](https://github.com/NVIDIA/cudf/pull/15695)) [@galipremsagar](https://github.com/galipremsagar) +- Make `nan_as_null` behavior consistent across all APIs ([#15692](https://github.com/NVIDIA/cudf/pull/15692)) [@galipremsagar](https://github.com/galipremsagar) +- Fix CI s3 api command to fetch latest results ([#15687](https://github.com/NVIDIA/cudf/pull/15687)) [@galipremsagar](https://github.com/galipremsagar) +- Add `NumpyExtensionArray` proxy type in `cudf.pandas` ([#15686](https://github.com/NVIDIA/cudf/pull/15686)) [@galipremsagar](https://github.com/galipremsagar) +- Properly implement binaryops for proxy types ([#15684](https://github.com/NVIDIA/cudf/pull/15684)) [@galipremsagar](https://github.com/galipremsagar) +- Fix copy assignment and the comparison operator of `rmm_host_allocator` ([#15677](https://github.com/NVIDIA/cudf/pull/15677)) [@vuule](https://github.com/vuule) +- Fix multi-source reading in JSON byte range reader ([#15671](https://github.com/NVIDIA/cudf/pull/15671)) [@shrshi](https://github.com/shrshi) +- Return `int64` when pandas compatible mode is turned on for `get_indexer` ([#15659](https://github.com/NVIDIA/cudf/pull/15659)) [@galipremsagar](https://github.com/galipremsagar) +- Fix Index contains for error validations and float vs int comparisons ([#15657](https://github.com/NVIDIA/cudf/pull/15657)) [@galipremsagar](https://github.com/galipremsagar) +- Preserve sub-second data for time scalars in column construction ([#15655](https://github.com/NVIDIA/cudf/pull/15655)) [@galipremsagar](https://github.com/galipremsagar) +- Check row limit size in cudf::strings::join_strings ([#15643](https://github.com/NVIDIA/cudf/pull/15643)) [@davidwendt](https://github.com/davidwendt) +- Enable sorting on column with nulls using query-planning ([#15639](https://github.com/NVIDIA/cudf/pull/15639)) [@rjzamora](https://github.com/rjzamora) +- Fix operator precedence problem in Parquet reader ([#15638](https://github.com/NVIDIA/cudf/pull/15638)) [@etseidl](https://github.com/etseidl) +- Fix decoding of dictionary encoded FIXED_LEN_BYTE_ARRAY data in Parquet reader ([#15601](https://github.com/NVIDIA/cudf/pull/15601)) [@etseidl](https://github.com/etseidl) +- Fix debug warnings/errors in from_arrow_device_test.cpp ([#15596](https://github.com/NVIDIA/cudf/pull/15596)) [@davidwendt](https://github.com/davidwendt) +- Add "collect" aggregation support to dask-cudf ([#15593](https://github.com/NVIDIA/cudf/pull/15593)) [@rjzamora](https://github.com/rjzamora) +- Fix categorical-accessor support and testing in dask-cudf ([#15591](https://github.com/NVIDIA/cudf/pull/15591)) [@rjzamora](https://github.com/rjzamora) +- Disable compute-sanitizer usage in CI tests with CUDA<11.6 ([#15584](https://github.com/NVIDIA/cudf/pull/15584)) [@davidwendt](https://github.com/davidwendt) +- Preserve RangeIndex.step in to_arrow/from_arrow ([#15581](https://github.com/NVIDIA/cudf/pull/15581)) [@mroeschke](https://github.com/mroeschke) +- Ignore new cupy warning ([#15574](https://github.com/NVIDIA/cudf/pull/15574)) [@vyasr](https://github.com/vyasr) +- Add cuda-sanitizer-api dependency for test-cpp matrix 11.4 ([#15573](https://github.com/NVIDIA/cudf/pull/15573)) [@davidwendt](https://github.com/davidwendt) +- Allow apply udf to reference global modules in cudf.pandas ([#15569](https://github.com/NVIDIA/cudf/pull/15569)) [@mroeschke](https://github.com/mroeschke) +- Fix deprecation warnings for json legacy reader ([#15563](https://github.com/NVIDIA/cudf/pull/15563)) [@davidwendt](https://github.com/davidwendt) +- Fix millisecond resampling in cudf Python ([#15560](https://github.com/NVIDIA/cudf/pull/15560)) [@mroeschke](https://github.com/mroeschke) +- Rename JSON_READER_OPTION to JSON_READER_OPTION_NVBENCH. ([#15553](https://github.com/NVIDIA/cudf/pull/15553)) [@bdice](https://github.com/bdice) +- Fix a JNI bug in JSON parsing fixup ([#15550](https://github.com/NVIDIA/cudf/pull/15550)) [@revans2](https://github.com/revans2) +- Remove conda channel setup from wheel CI image script. ([#15539](https://github.com/NVIDIA/cudf/pull/15539)) [@bdice](https://github.com/bdice) +- cudf.pandas: Series dt accessor is CombinedDatetimelikeProperties ([#15523](https://github.com/NVIDIA/cudf/pull/15523)) [@wence-](https://github.com/wence-) +- Fix for some compiler warnings in parquet/page_decode.cuh ([#15518](https://github.com/NVIDIA/cudf/pull/15518)) [@etseidl](https://github.com/etseidl) +- Fix exponent overflow in strings-to-double conversion ([#15517](https://github.com/NVIDIA/cudf/pull/15517)) [@davidwendt](https://github.com/davidwendt) +- nanoarrow uses package override for proper pinned versions generation ([#15515](https://github.com/NVIDIA/cudf/pull/15515)) [@robertmaynard](https://github.com/robertmaynard) +- Remove index name overrides in dask-cudf pyarrow table dispatch ([#15514](https://github.com/NVIDIA/cudf/pull/15514)) [@charlesbluca](https://github.com/charlesbluca) +- Fix async synchronization issues in json_column.cu ([#15497](https://github.com/NVIDIA/cudf/pull/15497)) [@karthikeyann](https://github.com/karthikeyann) +- Add new patch to hide more CCCL APIs ([#15493](https://github.com/NVIDIA/cudf/pull/15493)) [@vyasr](https://github.com/vyasr) +- Make improvements in pandas-test reporting ([#15485](https://github.com/NVIDIA/cudf/pull/15485)) [@galipremsagar](https://github.com/galipremsagar) +- Fixed page data truncation in parquet writer under certain conditions. ([#15474](https://github.com/NVIDIA/cudf/pull/15474)) [@nvdbaranec](https://github.com/nvdbaranec) +- Only use data_type constructor with scale for decimal types ([#15472](https://github.com/NVIDIA/cudf/pull/15472)) [@wence-](https://github.com/wence-) +- Avoid "p2p" shuffle as a default when `dask_cudf` is imported ([#15469](https://github.com/NVIDIA/cudf/pull/15469)) [@rjzamora](https://github.com/rjzamora) +- Fix debug build errors from to_arrow_device_test.cpp ([#15463](https://github.com/NVIDIA/cudf/pull/15463)) [@davidwendt](https://github.com/davidwendt) +- Fix base_normalator::integer_sizeof_fn integer dispatch ([#15457](https://github.com/NVIDIA/cudf/pull/15457)) [@davidwendt](https://github.com/davidwendt) +- Allow consumers of static builds to find nanoarrow ([#15456](https://github.com/NVIDIA/cudf/pull/15456)) [@robertmaynard](https://github.com/robertmaynard) +- Allow jit compilation when using a splayed CUDA toolkit ([#15451](https://github.com/NVIDIA/cudf/pull/15451)) [@robertmaynard](https://github.com/robertmaynard) +- Handle case of scan aggregation in groupby-transform ([#15450](https://github.com/NVIDIA/cudf/pull/15450)) [@wence-](https://github.com/wence-) +- Test static builds in CI and fix nanoarrow configure ([#15437](https://github.com/NVIDIA/cudf/pull/15437)) [@vyasr](https://github.com/vyasr) +- Fixes potential race in JSON parser when parsing JSON lines format and when recovering from invalid lines ([#15419](https://github.com/NVIDIA/cudf/pull/15419)) [@elstehle](https://github.com/elstehle) +- Fix errors in chunked ORC writer when no tables were (successfully) written ([#15393](https://github.com/NVIDIA/cudf/pull/15393)) [@vuule](https://github.com/vuule) +- Support implicit array conversion with query-planning enabled ([#15378](https://github.com/NVIDIA/cudf/pull/15378)) [@rjzamora](https://github.com/rjzamora) +- Fix arrow-based round trip of empty dataframes ([#15373](https://github.com/NVIDIA/cudf/pull/15373)) [@wence-](https://github.com/wence-) +- Remove empty elements from exploded character-ngrams output ([#15371](https://github.com/NVIDIA/cudf/pull/15371)) [@davidwendt](https://github.com/davidwendt) +- Remove boundscheck=False setting in cython files ([#15362](https://github.com/NVIDIA/cudf/pull/15362)) [@wence-](https://github.com/wence-) +- Patch dask-expr `var` logic in dask-cudf ([#15347](https://github.com/NVIDIA/cudf/pull/15347)) [@rjzamora](https://github.com/rjzamora) +- Fix for logical and syntactical errors in libcudf c++ examples ([#15346](https://github.com/NVIDIA/cudf/pull/15346)) [@mhaseeb123](https://github.com/mhaseeb123) +- Disable dask-expr in docs builds. ([#15343](https://github.com/NVIDIA/cudf/pull/15343)) [@bdice](https://github.com/bdice) +- Apply the cuFile error work around to data_sink as well ([#15335](https://github.com/NVIDIA/cudf/pull/15335)) [@vuule](https://github.com/vuule) +- Fix parquet predicate filtering with column projection ([#15113](https://github.com/NVIDIA/cudf/pull/15113)) [@karthikeyann](https://github.com/karthikeyann) +- Check column type equality, handling nested types correctly. ([#14531](https://github.com/NVIDIA/cudf/pull/14531)) [@bdice](https://github.com/bdice) ## 📖 Documentation -- Fix docs for IO readers and strings_convert ([#15842](https://github.com/rapidsai/cudf/pull/15842)) [@bdice](https://github.com/bdice) -- Update cudf.pandas docs for GA ([#15744](https://github.com/rapidsai/cudf/pull/15744)) [@beckernick](https://github.com/beckernick) -- Add contributing warning about circular imports ([#15691](https://github.com/rapidsai/cudf/pull/15691)) [@er-eis](https://github.com/er-eis) -- Update libcudf developer guide for strings offsets column ([#15661](https://github.com/rapidsai/cudf/pull/15661)) [@davidwendt](https://github.com/davidwendt) -- Update developer guide with device_async_resource_ref guidelines ([#15562](https://github.com/rapidsai/cudf/pull/15562)) [@harrism](https://github.com/harrism) -- DOC: add pandas intersphinx mapping ([#15531](https://github.com/rapidsai/cudf/pull/15531)) [@raybellwaves](https://github.com/raybellwaves) -- rm-dup-doc in frame.py ([#15530](https://github.com/rapidsai/cudf/pull/15530)) [@raybellwaves](https://github.com/raybellwaves) -- Update CONTRIBUTING.md to use latest cuda env ([#15467](https://github.com/rapidsai/cudf/pull/15467)) [@raybellwaves](https://github.com/raybellwaves) -- Doc: interleave columns pandas compat ([#15383](https://github.com/rapidsai/cudf/pull/15383)) [@raybellwaves](https://github.com/raybellwaves) -- Simplified README Examples ([#15338](https://github.com/rapidsai/cudf/pull/15338)) [@wkaisertexas](https://github.com/wkaisertexas) -- Add debug tips section to libcudf developer guide ([#15329](https://github.com/rapidsai/cudf/pull/15329)) [@davidwendt](https://github.com/davidwendt) -- Fix and clarify notes on result ordering ([#13255](https://github.com/rapidsai/cudf/pull/13255)) [@shwina](https://github.com/shwina) +- Fix docs for IO readers and strings_convert ([#15842](https://github.com/NVIDIA/cudf/pull/15842)) [@bdice](https://github.com/bdice) +- Update cudf.pandas docs for GA ([#15744](https://github.com/NVIDIA/cudf/pull/15744)) [@beckernick](https://github.com/beckernick) +- Add contributing warning about circular imports ([#15691](https://github.com/NVIDIA/cudf/pull/15691)) [@er-eis](https://github.com/er-eis) +- Update libcudf developer guide for strings offsets column ([#15661](https://github.com/NVIDIA/cudf/pull/15661)) [@davidwendt](https://github.com/davidwendt) +- Update developer guide with device_async_resource_ref guidelines ([#15562](https://github.com/NVIDIA/cudf/pull/15562)) [@harrism](https://github.com/harrism) +- DOC: add pandas intersphinx mapping ([#15531](https://github.com/NVIDIA/cudf/pull/15531)) [@raybellwaves](https://github.com/raybellwaves) +- rm-dup-doc in frame.py ([#15530](https://github.com/NVIDIA/cudf/pull/15530)) [@raybellwaves](https://github.com/raybellwaves) +- Update CONTRIBUTING.md to use latest cuda env ([#15467](https://github.com/NVIDIA/cudf/pull/15467)) [@raybellwaves](https://github.com/raybellwaves) +- Doc: interleave columns pandas compat ([#15383](https://github.com/NVIDIA/cudf/pull/15383)) [@raybellwaves](https://github.com/raybellwaves) +- Simplified README Examples ([#15338](https://github.com/NVIDIA/cudf/pull/15338)) [@wkaisertexas](https://github.com/wkaisertexas) +- Add debug tips section to libcudf developer guide ([#15329](https://github.com/NVIDIA/cudf/pull/15329)) [@davidwendt](https://github.com/davidwendt) +- Fix and clarify notes on result ordering ([#13255](https://github.com/NVIDIA/cudf/pull/13255)) [@shwina](https://github.com/shwina) ## 🚀 New Features -- Add JNI bindings for zstd compression of NVCOMP. ([#15729](https://github.com/rapidsai/cudf/pull/15729)) [@firestarman](https://github.com/firestarman) -- Fix spaces around CSV quoted strings ([#15727](https://github.com/rapidsai/cudf/pull/15727)) [@thabetx](https://github.com/thabetx) -- Add default pinned pool that falls back to new pinned allocations ([#15665](https://github.com/rapidsai/cudf/pull/15665)) [@vuule](https://github.com/vuule) -- Overhaul ops-codeowners coverage ([#15660](https://github.com/rapidsai/cudf/pull/15660)) [@raydouglass](https://github.com/raydouglass) -- Concatenate dictionary of objects along axis=1 ([#15623](https://github.com/rapidsai/cudf/pull/15623)) [@er-eis](https://github.com/er-eis) -- Construct `pylibcudf` columns from objects supporting `__cuda_array_interface__` ([#15615](https://github.com/rapidsai/cudf/pull/15615)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Expose some Parquet per-column configuration options via the python API ([#15613](https://github.com/rapidsai/cudf/pull/15613)) [@etseidl](https://github.com/etseidl) -- Migrate string `find` operations to `pylibcudf` ([#15604](https://github.com/rapidsai/cudf/pull/15604)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Round trip FIXED_LEN_BYTE_ARRAY data properly in Parquet writer ([#15600](https://github.com/rapidsai/cudf/pull/15600)) [@etseidl](https://github.com/etseidl) -- Reading multi-line JSON in string columns using runtime configurable delimiter ([#15556](https://github.com/rapidsai/cudf/pull/15556)) [@shrshi](https://github.com/shrshi) -- Remove public gtest dependency from libcudf conda package ([#15534](https://github.com/rapidsai/cudf/pull/15534)) [@robertmaynard](https://github.com/robertmaynard) -- Fea/move to latest nanoarrow ([#15526](https://github.com/rapidsai/cudf/pull/15526)) [@robertmaynard](https://github.com/robertmaynard) -- Migrate string `case` operations to `pylibcudf` ([#15489](https://github.com/rapidsai/cudf/pull/15489)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add Parquet encoding statistics to column chunk metadata ([#15452](https://github.com/rapidsai/cudf/pull/15452)) [@etseidl](https://github.com/etseidl) -- Implement JNI for chunked ORC reader ([#15446](https://github.com/rapidsai/cudf/pull/15446)) [@ttnghia](https://github.com/ttnghia) -- Add some missing optional fields to the Parquet RowGroup metadata ([#15421](https://github.com/rapidsai/cudf/pull/15421)) [@etseidl](https://github.com/etseidl) -- Adding parquet transcoding example ([#15420](https://github.com/rapidsai/cudf/pull/15420)) [@mhaseeb123](https://github.com/mhaseeb123) -- Add fields to Parquet Statistics structure that were added in parquet-format 2.10 ([#15412](https://github.com/rapidsai/cudf/pull/15412)) [@etseidl](https://github.com/etseidl) -- Add option to Parquet writer to skip compressing individual columns ([#15411](https://github.com/rapidsai/cudf/pull/15411)) [@etseidl](https://github.com/etseidl) -- Add BYTE_STREAM_SPLIT support to Parquet ([#15311](https://github.com/rapidsai/cudf/pull/15311)) [@etseidl](https://github.com/etseidl) -- Introduce benchmark suite for JSON reader options ([#15124](https://github.com/rapidsai/cudf/pull/15124)) [@shrshi](https://github.com/shrshi) -- Implement ORC chunked reader ([#15094](https://github.com/rapidsai/cudf/pull/15094)) [@ttnghia](https://github.com/ttnghia) -- Extend cudf devcontainers to specify jitify2 kernel cache ([#15068](https://github.com/rapidsai/cudf/pull/15068)) [@robertmaynard](https://github.com/robertmaynard) -- Add `to_arrow_device` function to cudf interop using nanoarrow ([#15047](https://github.com/rapidsai/cudf/pull/15047)) [@zeroshade](https://github.com/zeroshade) -- Add JSON option to prune columns ([#14996](https://github.com/rapidsai/cudf/pull/14996)) [@karthikeyann](https://github.com/karthikeyann) +- Add JNI bindings for zstd compression of NVCOMP. ([#15729](https://github.com/NVIDIA/cudf/pull/15729)) [@firestarman](https://github.com/firestarman) +- Fix spaces around CSV quoted strings ([#15727](https://github.com/NVIDIA/cudf/pull/15727)) [@thabetx](https://github.com/thabetx) +- Add default pinned pool that falls back to new pinned allocations ([#15665](https://github.com/NVIDIA/cudf/pull/15665)) [@vuule](https://github.com/vuule) +- Overhaul ops-codeowners coverage ([#15660](https://github.com/NVIDIA/cudf/pull/15660)) [@raydouglass](https://github.com/raydouglass) +- Concatenate dictionary of objects along axis=1 ([#15623](https://github.com/NVIDIA/cudf/pull/15623)) [@er-eis](https://github.com/er-eis) +- Construct `pylibcudf` columns from objects supporting `__cuda_array_interface__` ([#15615](https://github.com/NVIDIA/cudf/pull/15615)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Expose some Parquet per-column configuration options via the python API ([#15613](https://github.com/NVIDIA/cudf/pull/15613)) [@etseidl](https://github.com/etseidl) +- Migrate string `find` operations to `pylibcudf` ([#15604](https://github.com/NVIDIA/cudf/pull/15604)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Round trip FIXED_LEN_BYTE_ARRAY data properly in Parquet writer ([#15600](https://github.com/NVIDIA/cudf/pull/15600)) [@etseidl](https://github.com/etseidl) +- Reading multi-line JSON in string columns using runtime configurable delimiter ([#15556](https://github.com/NVIDIA/cudf/pull/15556)) [@shrshi](https://github.com/shrshi) +- Remove public gtest dependency from libcudf conda package ([#15534](https://github.com/NVIDIA/cudf/pull/15534)) [@robertmaynard](https://github.com/robertmaynard) +- Fea/move to latest nanoarrow ([#15526](https://github.com/NVIDIA/cudf/pull/15526)) [@robertmaynard](https://github.com/robertmaynard) +- Migrate string `case` operations to `pylibcudf` ([#15489](https://github.com/NVIDIA/cudf/pull/15489)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add Parquet encoding statistics to column chunk metadata ([#15452](https://github.com/NVIDIA/cudf/pull/15452)) [@etseidl](https://github.com/etseidl) +- Implement JNI for chunked ORC reader ([#15446](https://github.com/NVIDIA/cudf/pull/15446)) [@ttnghia](https://github.com/ttnghia) +- Add some missing optional fields to the Parquet RowGroup metadata ([#15421](https://github.com/NVIDIA/cudf/pull/15421)) [@etseidl](https://github.com/etseidl) +- Adding parquet transcoding example ([#15420](https://github.com/NVIDIA/cudf/pull/15420)) [@mhaseeb123](https://github.com/mhaseeb123) +- Add fields to Parquet Statistics structure that were added in parquet-format 2.10 ([#15412](https://github.com/NVIDIA/cudf/pull/15412)) [@etseidl](https://github.com/etseidl) +- Add option to Parquet writer to skip compressing individual columns ([#15411](https://github.com/NVIDIA/cudf/pull/15411)) [@etseidl](https://github.com/etseidl) +- Add BYTE_STREAM_SPLIT support to Parquet ([#15311](https://github.com/NVIDIA/cudf/pull/15311)) [@etseidl](https://github.com/etseidl) +- Introduce benchmark suite for JSON reader options ([#15124](https://github.com/NVIDIA/cudf/pull/15124)) [@shrshi](https://github.com/shrshi) +- Implement ORC chunked reader ([#15094](https://github.com/NVIDIA/cudf/pull/15094)) [@ttnghia](https://github.com/ttnghia) +- Extend cudf devcontainers to specify jitify2 kernel cache ([#15068](https://github.com/NVIDIA/cudf/pull/15068)) [@robertmaynard](https://github.com/robertmaynard) +- Add `to_arrow_device` function to cudf interop using nanoarrow ([#15047](https://github.com/NVIDIA/cudf/pull/15047)) [@zeroshade](https://github.com/zeroshade) +- Add JSON option to prune columns ([#14996](https://github.com/NVIDIA/cudf/pull/14996)) [@karthikeyann](https://github.com/karthikeyann) ## 🛠️ Improvements -- Deprecate `Groupby.collect` ([#15808](https://github.com/rapidsai/cudf/pull/15808)) [@galipremsagar](https://github.com/galipremsagar) -- Raise FileNotFoundError when a literal JSON string that looks like a json filename is passed ([#15806](https://github.com/rapidsai/cudf/pull/15806)) [@lithomas1](https://github.com/lithomas1) -- Deprecate `divisions='quantile'` support in `set_index` ([#15804](https://github.com/rapidsai/cudf/pull/15804)) [@rjzamora](https://github.com/rjzamora) -- Improve performance of Series.to_numpy/to_cupy ([#15792](https://github.com/rapidsai/cudf/pull/15792)) [@mroeschke](https://github.com/mroeschke) -- Access `self.index` instead of `self._index` where possible ([#15781](https://github.com/rapidsai/cudf/pull/15781)) [@mroeschke](https://github.com/mroeschke) -- Support filtered I/O in `chunked_parquet_reader` and simplify the use of `parquet_reader_options` ([#15764](https://github.com/rapidsai/cudf/pull/15764)) [@mhaseeb123](https://github.com/mhaseeb123) -- Avoid index-to-column conversion in some DataFrame ops ([#15763](https://github.com/rapidsai/cudf/pull/15763)) [@mroeschke](https://github.com/mroeschke) -- Fix `chunked_parquet_reader` behavior when input has no more rows to read ([#15757](https://github.com/rapidsai/cudf/pull/15757)) [@mhaseeb123](https://github.com/mhaseeb123) -- [JNI] Expose java API for cudf::io::config_host_memory_resource ([#15745](https://github.com/rapidsai/cudf/pull/15745)) [@abellina](https://github.com/abellina) -- Migrate all cpp pxd files into pylibcudf ([#15740](https://github.com/rapidsai/cudf/pull/15740)) [@vyasr](https://github.com/vyasr) -- Validate and materialize iterators earlier in as_column ([#15739](https://github.com/rapidsai/cudf/pull/15739)) [@mroeschke](https://github.com/mroeschke) -- Push some as_column arrow logic to ColumnBase.from_arrow ([#15738](https://github.com/rapidsai/cudf/pull/15738)) [@mroeschke](https://github.com/mroeschke) -- Expose stream parameter in public reduction APIs ([#15737](https://github.com/rapidsai/cudf/pull/15737)) [@srinivasyadav18](https://github.com/srinivasyadav18) -- remove unnecessary 'setuptools' host dependency, simplify dependencies.yaml ([#15736](https://github.com/rapidsai/cudf/pull/15736)) [@jameslamb](https://github.com/jameslamb) -- Defer to C++ equality and hashing for pylibcudf DataType and Aggregation objects ([#15732](https://github.com/rapidsai/cudf/pull/15732)) [@wence-](https://github.com/wence-) -- Implement null-aware NOT_EQUALS binop ([#15731](https://github.com/rapidsai/cudf/pull/15731)) [@wence-](https://github.com/wence-) -- Fix split-record result list column offset type ([#15707](https://github.com/rapidsai/cudf/pull/15707)) [@davidwendt](https://github.com/davidwendt) -- Upgrade `arrow` to `16` ([#15703](https://github.com/rapidsai/cudf/pull/15703)) [@galipremsagar](https://github.com/galipremsagar) -- Remove experimental namespace from make_strings_children ([#15702](https://github.com/rapidsai/cudf/pull/15702)) [@davidwendt](https://github.com/davidwendt) -- Rework get_json_object benchmark to use nvbench ([#15698](https://github.com/rapidsai/cudf/pull/15698)) [@davidwendt](https://github.com/davidwendt) -- Rework some python tests of Parquet delta encodings ([#15693](https://github.com/rapidsai/cudf/pull/15693)) [@etseidl](https://github.com/etseidl) -- Skeleton cudf polars package ([#15688](https://github.com/rapidsai/cudf/pull/15688)) [@wence-](https://github.com/wence-) -- Upgrade pre commit hooks ([#15685](https://github.com/rapidsai/cudf/pull/15685)) [@wence-](https://github.com/wence-) -- Allow `fillna` to validate for `CategoricalColumn.fillna` ([#15683](https://github.com/rapidsai/cudf/pull/15683)) [@galipremsagar](https://github.com/galipremsagar) -- Misc Column cleanups ([#15682](https://github.com/rapidsai/cudf/pull/15682)) [@mroeschke](https://github.com/mroeschke) -- Reducing runtime of JSON reader options benchmark ([#15681](https://github.com/rapidsai/cudf/pull/15681)) [@shrshi](https://github.com/shrshi) -- Add `Timestamp` and `Timedelta` proxy types ([#15680](https://github.com/rapidsai/cudf/pull/15680)) [@galipremsagar](https://github.com/galipremsagar) -- Remove host_parse_nested_json. ([#15674](https://github.com/rapidsai/cudf/pull/15674)) [@bdice](https://github.com/bdice) -- Reduce runtime for ParquetChunkedReaderInputLimitTest gtests ([#15672](https://github.com/rapidsai/cudf/pull/15672)) [@davidwendt](https://github.com/davidwendt) -- Add large-strings gtest for cudf::interleave_columns ([#15669](https://github.com/rapidsai/cudf/pull/15669)) [@davidwendt](https://github.com/davidwendt) -- Use experimental make_strings_children for multi-replace_re ([#15667](https://github.com/rapidsai/cudf/pull/15667)) [@davidwendt](https://github.com/davidwendt) -- Enabled `Holiday` types in `cudf.pandas` ([#15664](https://github.com/rapidsai/cudf/pull/15664)) [@galipremsagar](https://github.com/galipremsagar) -- Remove obsolete `XFAIL` markers for query-planning ([#15662](https://github.com/rapidsai/cudf/pull/15662)) [@rjzamora](https://github.com/rjzamora) -- Clean up join benchmarks ([#15644](https://github.com/rapidsai/cudf/pull/15644)) [@PointKernel](https://github.com/PointKernel) -- Enable warnings as errors in custreamz ([#15642](https://github.com/rapidsai/cudf/pull/15642)) [@mroeschke](https://github.com/mroeschke) -- Improve distinct join with set `retrieve` ([#15636](https://github.com/rapidsai/cudf/pull/15636)) [@PointKernel](https://github.com/PointKernel) -- Fix -Werror=type-limits. ([#15635](https://github.com/rapidsai/cudf/pull/15635)) [@bdice](https://github.com/bdice) -- Enable FutureWarnings/DeprecationWarnings as errors for dask_cudf ([#15634](https://github.com/rapidsai/cudf/pull/15634)) [@mroeschke](https://github.com/mroeschke) -- Remove NVBench SHA override. ([#15633](https://github.com/rapidsai/cudf/pull/15633)) [@alliepiper](https://github.com/alliepiper) -- Add support for large string columns to Parquet reader and writer ([#15632](https://github.com/rapidsai/cudf/pull/15632)) [@etseidl](https://github.com/etseidl) -- Large strings support in MD5 and SHA hashers ([#15631](https://github.com/rapidsai/cudf/pull/15631)) [@davidwendt](https://github.com/davidwendt) -- Fix make_offsets_child_column usage in cudf::strings::detail::shift ([#15630](https://github.com/rapidsai/cudf/pull/15630)) [@davidwendt](https://github.com/davidwendt) -- Use experimental make_strings_children for strings convert ([#15629](https://github.com/rapidsai/cudf/pull/15629)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-24.04 to branch-24.06 ([#15627](https://github.com/rapidsai/cudf/pull/15627)) [@bdice](https://github.com/bdice) -- Avoid accessing attributes via `_column` if not needed ([#15624](https://github.com/rapidsai/cudf/pull/15624)) [@mroeschke](https://github.com/mroeschke) -- Make ColumnBase.__cuda_array_interface__ opt out instead of opt in ([#15622](https://github.com/rapidsai/cudf/pull/15622)) [@mroeschke](https://github.com/mroeschke) -- Large strings support for cudf::gather ([#15621](https://github.com/rapidsai/cudf/pull/15621)) [@davidwendt](https://github.com/davidwendt) -- Remove jni-docker-build workflow ([#15619](https://github.com/rapidsai/cudf/pull/15619)) [@bdice](https://github.com/bdice) -- Support `DurationType` in cudf parquet reader via `arrow:schema` ([#15617](https://github.com/rapidsai/cudf/pull/15617)) [@mhaseeb123](https://github.com/mhaseeb123) -- Drop Centos7 support ([#15608](https://github.com/rapidsai/cudf/pull/15608)) [@NvTimLiu](https://github.com/NvTimLiu) -- Use experimental make_strings_children for json/csv writers ([#15599](https://github.com/rapidsai/cudf/pull/15599)) [@davidwendt](https://github.com/davidwendt) -- Use experimental make_strings_children for strings join/url_encode/slice ([#15598](https://github.com/rapidsai/cudf/pull/15598)) [@davidwendt](https://github.com/davidwendt) -- Use experimental make_strings_children in nvtext APIs ([#15595](https://github.com/rapidsai/cudf/pull/15595)) [@davidwendt](https://github.com/davidwendt) -- Migrate to `{{ stdlib("c") }}` ([#15594](https://github.com/rapidsai/cudf/pull/15594)) [@hcho3](https://github.com/hcho3) -- Deprecate `to/from_dask_dataframe` APIs in dask-cudf ([#15592](https://github.com/rapidsai/cudf/pull/15592)) [@rjzamora](https://github.com/rjzamora) -- Minor fixups for future NumPy 2 compatibility ([#15590](https://github.com/rapidsai/cudf/pull/15590)) [@seberg](https://github.com/seberg) -- Delay materializing RangeIndex in .reset_index ([#15588](https://github.com/rapidsai/cudf/pull/15588)) [@mroeschke](https://github.com/mroeschke) -- Use experimental make_strings_children for capitalize/case/pad functions ([#15587](https://github.com/rapidsai/cudf/pull/15587)) [@davidwendt](https://github.com/davidwendt) -- Use experimental make_strings_children for strings replace/filter/translate ([#15586](https://github.com/rapidsai/cudf/pull/15586)) [@davidwendt](https://github.com/davidwendt) -- Add multithreaded parquet reader benchmarks. ([#15585](https://github.com/rapidsai/cudf/pull/15585)) [@nvdbaranec](https://github.com/nvdbaranec) -- Don't materialize column during RangeIndex methods ([#15582](https://github.com/rapidsai/cudf/pull/15582)) [@mroeschke](https://github.com/mroeschke) -- Improve performance for cudf::strings::count_re ([#15578](https://github.com/rapidsai/cudf/pull/15578)) [@davidwendt](https://github.com/davidwendt) -- Replace RangeIndex._start/_stop/_step with _range ([#15576](https://github.com/rapidsai/cudf/pull/15576)) [@mroeschke](https://github.com/mroeschke) -- add --rm and --name to devcontainer run args ([#15572](https://github.com/rapidsai/cudf/pull/15572)) [@trxcllnt](https://github.com/trxcllnt) -- Change the default dictionary policy in Parquet writer from `ALWAYS` to `ADAPTIVE` ([#15570](https://github.com/rapidsai/cudf/pull/15570)) [@mhaseeb123](https://github.com/mhaseeb123) -- Rename experimental JSON tests. ([#15568](https://github.com/rapidsai/cudf/pull/15568)) [@bdice](https://github.com/bdice) -- Refactor JNI native dependency loading to allow returning of library path ([#15566](https://github.com/rapidsai/cudf/pull/15566)) [@jlowe](https://github.com/jlowe) -- Remove protobuf and use parsed ORC statistics from libcudf ([#15564](https://github.com/rapidsai/cudf/pull/15564)) [@bdice](https://github.com/bdice) -- Deprecate legacy JSON reader options. ([#15558](https://github.com/rapidsai/cudf/pull/15558)) [@bdice](https://github.com/bdice) -- Use same .clang-format in cuDF JNI ([#15557](https://github.com/rapidsai/cudf/pull/15557)) [@bdice](https://github.com/bdice) -- Large strings support for cudf::fill ([#15555](https://github.com/rapidsai/cudf/pull/15555)) [@davidwendt](https://github.com/davidwendt) -- Upgrade upper bound pinning to `pandas-2.2.2` ([#15554](https://github.com/rapidsai/cudf/pull/15554)) [@galipremsagar](https://github.com/galipremsagar) -- Work around issues with cccl main ([#15552](https://github.com/rapidsai/cudf/pull/15552)) [@miscco](https://github.com/miscco) -- Enable pandas plotting unit tests for cudf.pandas ([#15547](https://github.com/rapidsai/cudf/pull/15547)) [@mroeschke](https://github.com/mroeschke) -- Move timezone conversion logic to `DatetimeColumn` ([#15545](https://github.com/rapidsai/cudf/pull/15545)) [@mroeschke](https://github.com/mroeschke) -- Large strings support for cudf::interleave_columns ([#15544](https://github.com/rapidsai/cudf/pull/15544)) [@davidwendt](https://github.com/davidwendt) -- [skip ci] Switch back to 24.06 branch for pandas tests ([#15543](https://github.com/rapidsai/cudf/pull/15543)) [@galipremsagar](https://github.com/galipremsagar) -- Remove checks dependency from static-configure test job. ([#15542](https://github.com/rapidsai/cudf/pull/15542)) [@bdice](https://github.com/bdice) -- Remove legacy JSON reader from Python ([#15538](https://github.com/rapidsai/cudf/pull/15538)) [@bdice](https://github.com/bdice) -- Enable more ignored pandas unit tests for cudf.pandas ([#15535](https://github.com/rapidsai/cudf/pull/15535)) [@mroeschke](https://github.com/mroeschke) -- Large strings support for cudf::clamp ([#15533](https://github.com/rapidsai/cudf/pull/15533)) [@davidwendt](https://github.com/davidwendt) -- Remove version hard-coding ([#15529](https://github.com/rapidsai/cudf/pull/15529)) [@galipremsagar](https://github.com/galipremsagar) -- Removing all batching code from parquet writer ([#15528](https://github.com/rapidsai/cudf/pull/15528)) [@mhaseeb123](https://github.com/mhaseeb123) -- Make some private class properties not settable ([#15527](https://github.com/rapidsai/cudf/pull/15527)) [@mroeschke](https://github.com/mroeschke) -- Large strings support in regex replace APIs ([#15524](https://github.com/rapidsai/cudf/pull/15524)) [@davidwendt](https://github.com/davidwendt) -- Skip pandas unit tests that crash pytest workers in `cudf.pandas` ([#15521](https://github.com/rapidsai/cudf/pull/15521)) [@mroeschke](https://github.com/mroeschke) -- Preserve column metadata during more DataFrame operations ([#15519](https://github.com/rapidsai/cudf/pull/15519)) [@mroeschke](https://github.com/mroeschke) -- Move to pandas-tests to a dedicated workflow file and trigger it from branch.yaml ([#15516](https://github.com/rapidsai/cudf/pull/15516)) [@galipremsagar](https://github.com/galipremsagar) -- Large strings gtest fixture and utilities ([#15513](https://github.com/rapidsai/cudf/pull/15513)) [@davidwendt](https://github.com/davidwendt) -- Convert libcudf resource parameters to rmm::device_async_resource_ref ([#15507](https://github.com/rapidsai/cudf/pull/15507)) [@harrism](https://github.com/harrism) -- Relax protobuf lower bound to 3.20. ([#15506](https://github.com/rapidsai/cudf/pull/15506)) [@bdice](https://github.com/bdice) -- Clean up index methods ([#15496](https://github.com/rapidsai/cudf/pull/15496)) [@mroeschke](https://github.com/mroeschke) -- Update strings contains benchmarks to nvbench ([#15495](https://github.com/rapidsai/cudf/pull/15495)) [@davidwendt](https://github.com/davidwendt) -- Update NVBench fixture to use new hooks, fix pinned memory segfault. ([#15492](https://github.com/rapidsai/cudf/pull/15492)) [@alliepiper](https://github.com/alliepiper) -- Enable tests/scalar and test/series in cudf.pandas tests ([#15486](https://github.com/rapidsai/cudf/pull/15486)) [@mroeschke](https://github.com/mroeschke) -- Clean up __cuda_array_interface__ handling in as_column ([#15477](https://github.com/rapidsai/cudf/pull/15477)) [@mroeschke](https://github.com/mroeschke) -- Avoid .ordered and .categories from being settable in CategoricalColumn and CategoricalDtype ([#15475](https://github.com/rapidsai/cudf/pull/15475)) [@mroeschke](https://github.com/mroeschke) -- Ignore pandas tests for cudf.pandas that need motoserver ([#15468](https://github.com/rapidsai/cudf/pull/15468)) [@mroeschke](https://github.com/mroeschke) -- Use cached_property for NumericColumn.nan_count instead of ._nan_count variable ([#15466](https://github.com/rapidsai/cudf/pull/15466)) [@mroeschke](https://github.com/mroeschke) -- Add to_arrow_device() functions that accept views ([#15465](https://github.com/rapidsai/cudf/pull/15465)) [@davidwendt](https://github.com/davidwendt) -- Add custom status check workflow ([#15464](https://github.com/rapidsai/cudf/pull/15464)) [@galipremsagar](https://github.com/galipremsagar) -- Disable pandas 2.x clipboard tests in cudf.pandas tests ([#15462](https://github.com/rapidsai/cudf/pull/15462)) [@mroeschke](https://github.com/mroeschke) -- Enable tests/strings/test_api.py and tests/io/pytables in cudf.pandas tests ([#15461](https://github.com/rapidsai/cudf/pull/15461)) [@mroeschke](https://github.com/mroeschke) -- Enable test_parsing in cudf.pandas tests ([#15460](https://github.com/rapidsai/cudf/pull/15460)) [@mroeschke](https://github.com/mroeschke) -- Add `from_arrow_device` function to cudf interop using nanoarrow ([#15458](https://github.com/rapidsai/cudf/pull/15458)) [@zeroshade](https://github.com/zeroshade) -- Remove deprecated strings offsets_begin ([#15454](https://github.com/rapidsai/cudf/pull/15454)) [@davidwendt](https://github.com/davidwendt) -- Enable tests/windows/ in cudf.pandas tests ([#15444](https://github.com/rapidsai/cudf/pull/15444)) [@mroeschke](https://github.com/mroeschke) -- Enable tests/interchange/test_impl.py in cudf.pandas tests ([#15443](https://github.com/rapidsai/cudf/pull/15443)) [@mroeschke](https://github.com/mroeschke) -- Enable tests/io/test_user_agent.py in cudf pandas tests ([#15442](https://github.com/rapidsai/cudf/pull/15442)) [@mroeschke](https://github.com/mroeschke) -- Performance improvement in libcudf case conversion for long strings ([#15441](https://github.com/rapidsai/cudf/pull/15441)) [@davidwendt](https://github.com/davidwendt) -- Remove prior test skipping in run-pandas-tests with testing 2.2.1 ([#15440](https://github.com/rapidsai/cudf/pull/15440)) [@mroeschke](https://github.com/mroeschke) -- Support orc and text IO with dask-expr using legacy conversion ([#15439](https://github.com/rapidsai/cudf/pull/15439)) [@rjzamora](https://github.com/rjzamora) -- Floating <--> fixed-point conversion must now be called explicitly ([#15438](https://github.com/rapidsai/cudf/pull/15438)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Unify Copy-On-Write and Spilling ([#15436](https://github.com/rapidsai/cudf/pull/15436)) [@madsbk](https://github.com/madsbk) -- Enable ``dask_cudf`` json and s3 tests with query-planning on ([#15408](https://github.com/rapidsai/cudf/pull/15408)) [@rjzamora](https://github.com/rjzamora) -- Bump ruff and codespell pre-commit checks ([#15407](https://github.com/rapidsai/cudf/pull/15407)) [@mroeschke](https://github.com/mroeschke) -- Enable all tests for `arm` arch ([#15402](https://github.com/rapidsai/cudf/pull/15402)) [@galipremsagar](https://github.com/galipremsagar) -- Bind `read_parquet_metadata` API to libcudf instead of pyarrow and extract `RowGroup` information ([#15398](https://github.com/rapidsai/cudf/pull/15398)) [@mhaseeb123](https://github.com/mhaseeb123) -- Optimizing multi-source byte range reading in JSON reader ([#15396](https://github.com/rapidsai/cudf/pull/15396)) [@shrshi](https://github.com/shrshi) -- add correct labels to pandas_function_request.md ([#15381](https://github.com/rapidsai/cudf/pull/15381)) [@raybellwaves](https://github.com/raybellwaves) -- Remove deprecated hash() and spark_murmurhash3_x86_32() ([#15375](https://github.com/rapidsai/cudf/pull/15375)) [@davidwendt](https://github.com/davidwendt) -- Large strings support in cudf::merge ([#15374](https://github.com/rapidsai/cudf/pull/15374)) [@davidwendt](https://github.com/davidwendt) -- Enable test-reporting for pandas pytests in CI ([#15369](https://github.com/rapidsai/cudf/pull/15369)) [@galipremsagar](https://github.com/galipremsagar) -- Use logical types in Parquet reader ([#15365](https://github.com/rapidsai/cudf/pull/15365)) [@etseidl](https://github.com/etseidl) -- Add experimental make_strings_children utility ([#15363](https://github.com/rapidsai/cudf/pull/15363)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-24.04 to branch-24.06 ([#15349](https://github.com/rapidsai/cudf/pull/15349)) [@bdice](https://github.com/bdice) -- Fix CMake files in libcudf C++ examples to use existing libcudf build if present ([#15348](https://github.com/rapidsai/cudf/pull/15348)) [@mhaseeb123](https://github.com/mhaseeb123) -- Use ruff pydocstyle over pydocstyle pre-commit hook ([#15345](https://github.com/rapidsai/cudf/pull/15345)) [@mroeschke](https://github.com/mroeschke) -- Refactor stream mode setup for gtests ([#15337](https://github.com/rapidsai/cudf/pull/15337)) [@davidwendt](https://github.com/davidwendt) -- Benchmark decimal <--> floating conversions. ([#15334](https://github.com/rapidsai/cudf/pull/15334)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Avoid duplicate dask-cudf testing ([#15333](https://github.com/rapidsai/cudf/pull/15333)) [@rjzamora](https://github.com/rjzamora) -- Skip decode steps in Parquet reader when nullable columns have no nulls ([#15332](https://github.com/rapidsai/cudf/pull/15332)) [@etseidl](https://github.com/etseidl) -- Update udf_cpp to use rapids_cpm_cccl. ([#15331](https://github.com/rapidsai/cudf/pull/15331)) [@bdice](https://github.com/bdice) -- Forward-merge branch-24.04 into branch-24.06 [skip ci] ([#15330](https://github.com/rapidsai/cudf/pull/15330)) [@rapids-bot[bot]](https://github.com/rapids-bot[bot]) -- Allow ``numeric_only=True`` for simple groupby reductions ([#15326](https://github.com/rapidsai/cudf/pull/15326)) [@rjzamora](https://github.com/rjzamora) -- Drop CentOS 7 support. ([#15323](https://github.com/rapidsai/cudf/pull/15323)) [@bdice](https://github.com/bdice) -- Rework cudf::find_and_replace_all to use gather-based make_strings_column ([#15305](https://github.com/rapidsai/cudf/pull/15305)) [@davidwendt](https://github.com/davidwendt) -- First pass at adding testing for pylibcudf ([#15300](https://github.com/rapidsai/cudf/pull/15300)) [@vyasr](https://github.com/vyasr) -- [FEA] Performance improvement for mixed left semi/anti join ([#15288](https://github.com/rapidsai/cudf/pull/15288)) [@tgujar](https://github.com/tgujar) -- Rework cudf::replace_nulls to use strings::detail::copy_if_else ([#15286](https://github.com/rapidsai/cudf/pull/15286)) [@davidwendt](https://github.com/davidwendt) -- Clean up special casing in `as_column` for non-typed input ([#15276](https://github.com/rapidsai/cudf/pull/15276)) [@mroeschke](https://github.com/mroeschke) -- Large strings support in cudf::concatenate ([#15195](https://github.com/rapidsai/cudf/pull/15195)) [@davidwendt](https://github.com/davidwendt) -- Use less _is_categorical_dtype ([#15148](https://github.com/rapidsai/cudf/pull/15148)) [@mroeschke](https://github.com/mroeschke) -- Align date_range defaults with pandas, support tz ([#15139](https://github.com/rapidsai/cudf/pull/15139)) [@mroeschke](https://github.com/mroeschke) -- `ModuleAccelerator` performance: cache the result of checking if a caller is in the denylist ([#15056](https://github.com/rapidsai/cudf/pull/15056)) [@shwina](https://github.com/shwina) -- Use offsetalator in cudf::strings::replace functions ([#14824](https://github.com/rapidsai/cudf/pull/14824)) [@davidwendt](https://github.com/davidwendt) -- Cleanup some timedelta/datetime column logic ([#14715](https://github.com/rapidsai/cudf/pull/14715)) [@mroeschke](https://github.com/mroeschke) -- Refactor numpy array input in as_column ([#14651](https://github.com/rapidsai/cudf/pull/14651)) [@mroeschke](https://github.com/mroeschke) -- Refactor joins for conditional semis and antis ([#14646](https://github.com/rapidsai/cudf/pull/14646)) [@DanialJavady96](https://github.com/DanialJavady96) -- Eagerly populate the class dict for cudf.pandas proxy types ([#14534](https://github.com/rapidsai/cudf/pull/14534)) [@shwina](https://github.com/shwina) -- Some additional kernel thread index refactoring. ([#14107](https://github.com/rapidsai/cudf/pull/14107)) [@bdice](https://github.com/bdice) +- Deprecate `Groupby.collect` ([#15808](https://github.com/NVIDIA/cudf/pull/15808)) [@galipremsagar](https://github.com/galipremsagar) +- Raise FileNotFoundError when a literal JSON string that looks like a json filename is passed ([#15806](https://github.com/NVIDIA/cudf/pull/15806)) [@lithomas1](https://github.com/lithomas1) +- Deprecate `divisions='quantile'` support in `set_index` ([#15804](https://github.com/NVIDIA/cudf/pull/15804)) [@rjzamora](https://github.com/rjzamora) +- Improve performance of Series.to_numpy/to_cupy ([#15792](https://github.com/NVIDIA/cudf/pull/15792)) [@mroeschke](https://github.com/mroeschke) +- Access `self.index` instead of `self._index` where possible ([#15781](https://github.com/NVIDIA/cudf/pull/15781)) [@mroeschke](https://github.com/mroeschke) +- Support filtered I/O in `chunked_parquet_reader` and simplify the use of `parquet_reader_options` ([#15764](https://github.com/NVIDIA/cudf/pull/15764)) [@mhaseeb123](https://github.com/mhaseeb123) +- Avoid index-to-column conversion in some DataFrame ops ([#15763](https://github.com/NVIDIA/cudf/pull/15763)) [@mroeschke](https://github.com/mroeschke) +- Fix `chunked_parquet_reader` behavior when input has no more rows to read ([#15757](https://github.com/NVIDIA/cudf/pull/15757)) [@mhaseeb123](https://github.com/mhaseeb123) +- [JNI] Expose java API for cudf::io::config_host_memory_resource ([#15745](https://github.com/NVIDIA/cudf/pull/15745)) [@abellina](https://github.com/abellina) +- Migrate all cpp pxd files into pylibcudf ([#15740](https://github.com/NVIDIA/cudf/pull/15740)) [@vyasr](https://github.com/vyasr) +- Validate and materialize iterators earlier in as_column ([#15739](https://github.com/NVIDIA/cudf/pull/15739)) [@mroeschke](https://github.com/mroeschke) +- Push some as_column arrow logic to ColumnBase.from_arrow ([#15738](https://github.com/NVIDIA/cudf/pull/15738)) [@mroeschke](https://github.com/mroeschke) +- Expose stream parameter in public reduction APIs ([#15737](https://github.com/NVIDIA/cudf/pull/15737)) [@srinivasyadav18](https://github.com/srinivasyadav18) +- remove unnecessary 'setuptools' host dependency, simplify dependencies.yaml ([#15736](https://github.com/NVIDIA/cudf/pull/15736)) [@jameslamb](https://github.com/jameslamb) +- Defer to C++ equality and hashing for pylibcudf DataType and Aggregation objects ([#15732](https://github.com/NVIDIA/cudf/pull/15732)) [@wence-](https://github.com/wence-) +- Implement null-aware NOT_EQUALS binop ([#15731](https://github.com/NVIDIA/cudf/pull/15731)) [@wence-](https://github.com/wence-) +- Fix split-record result list column offset type ([#15707](https://github.com/NVIDIA/cudf/pull/15707)) [@davidwendt](https://github.com/davidwendt) +- Upgrade `arrow` to `16` ([#15703](https://github.com/NVIDIA/cudf/pull/15703)) [@galipremsagar](https://github.com/galipremsagar) +- Remove experimental namespace from make_strings_children ([#15702](https://github.com/NVIDIA/cudf/pull/15702)) [@davidwendt](https://github.com/davidwendt) +- Rework get_json_object benchmark to use nvbench ([#15698](https://github.com/NVIDIA/cudf/pull/15698)) [@davidwendt](https://github.com/davidwendt) +- Rework some python tests of Parquet delta encodings ([#15693](https://github.com/NVIDIA/cudf/pull/15693)) [@etseidl](https://github.com/etseidl) +- Skeleton cudf polars package ([#15688](https://github.com/NVIDIA/cudf/pull/15688)) [@wence-](https://github.com/wence-) +- Upgrade pre commit hooks ([#15685](https://github.com/NVIDIA/cudf/pull/15685)) [@wence-](https://github.com/wence-) +- Allow `fillna` to validate for `CategoricalColumn.fillna` ([#15683](https://github.com/NVIDIA/cudf/pull/15683)) [@galipremsagar](https://github.com/galipremsagar) +- Misc Column cleanups ([#15682](https://github.com/NVIDIA/cudf/pull/15682)) [@mroeschke](https://github.com/mroeschke) +- Reducing runtime of JSON reader options benchmark ([#15681](https://github.com/NVIDIA/cudf/pull/15681)) [@shrshi](https://github.com/shrshi) +- Add `Timestamp` and `Timedelta` proxy types ([#15680](https://github.com/NVIDIA/cudf/pull/15680)) [@galipremsagar](https://github.com/galipremsagar) +- Remove host_parse_nested_json. ([#15674](https://github.com/NVIDIA/cudf/pull/15674)) [@bdice](https://github.com/bdice) +- Reduce runtime for ParquetChunkedReaderInputLimitTest gtests ([#15672](https://github.com/NVIDIA/cudf/pull/15672)) [@davidwendt](https://github.com/davidwendt) +- Add large-strings gtest for cudf::interleave_columns ([#15669](https://github.com/NVIDIA/cudf/pull/15669)) [@davidwendt](https://github.com/davidwendt) +- Use experimental make_strings_children for multi-replace_re ([#15667](https://github.com/NVIDIA/cudf/pull/15667)) [@davidwendt](https://github.com/davidwendt) +- Enabled `Holiday` types in `cudf.pandas` ([#15664](https://github.com/NVIDIA/cudf/pull/15664)) [@galipremsagar](https://github.com/galipremsagar) +- Remove obsolete `XFAIL` markers for query-planning ([#15662](https://github.com/NVIDIA/cudf/pull/15662)) [@rjzamora](https://github.com/rjzamora) +- Clean up join benchmarks ([#15644](https://github.com/NVIDIA/cudf/pull/15644)) [@PointKernel](https://github.com/PointKernel) +- Enable warnings as errors in custreamz ([#15642](https://github.com/NVIDIA/cudf/pull/15642)) [@mroeschke](https://github.com/mroeschke) +- Improve distinct join with set `retrieve` ([#15636](https://github.com/NVIDIA/cudf/pull/15636)) [@PointKernel](https://github.com/PointKernel) +- Fix -Werror=type-limits. ([#15635](https://github.com/NVIDIA/cudf/pull/15635)) [@bdice](https://github.com/bdice) +- Enable FutureWarnings/DeprecationWarnings as errors for dask_cudf ([#15634](https://github.com/NVIDIA/cudf/pull/15634)) [@mroeschke](https://github.com/mroeschke) +- Remove NVBench SHA override. ([#15633](https://github.com/NVIDIA/cudf/pull/15633)) [@alliepiper](https://github.com/alliepiper) +- Add support for large string columns to Parquet reader and writer ([#15632](https://github.com/NVIDIA/cudf/pull/15632)) [@etseidl](https://github.com/etseidl) +- Large strings support in MD5 and SHA hashers ([#15631](https://github.com/NVIDIA/cudf/pull/15631)) [@davidwendt](https://github.com/davidwendt) +- Fix make_offsets_child_column usage in cudf::strings::detail::shift ([#15630](https://github.com/NVIDIA/cudf/pull/15630)) [@davidwendt](https://github.com/davidwendt) +- Use experimental make_strings_children for strings convert ([#15629](https://github.com/NVIDIA/cudf/pull/15629)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-24.04 to branch-24.06 ([#15627](https://github.com/NVIDIA/cudf/pull/15627)) [@bdice](https://github.com/bdice) +- Avoid accessing attributes via `_column` if not needed ([#15624](https://github.com/NVIDIA/cudf/pull/15624)) [@mroeschke](https://github.com/mroeschke) +- Make ColumnBase.__cuda_array_interface__ opt out instead of opt in ([#15622](https://github.com/NVIDIA/cudf/pull/15622)) [@mroeschke](https://github.com/mroeschke) +- Large strings support for cudf::gather ([#15621](https://github.com/NVIDIA/cudf/pull/15621)) [@davidwendt](https://github.com/davidwendt) +- Remove jni-docker-build workflow ([#15619](https://github.com/NVIDIA/cudf/pull/15619)) [@bdice](https://github.com/bdice) +- Support `DurationType` in cudf parquet reader via `arrow:schema` ([#15617](https://github.com/NVIDIA/cudf/pull/15617)) [@mhaseeb123](https://github.com/mhaseeb123) +- Drop Centos7 support ([#15608](https://github.com/NVIDIA/cudf/pull/15608)) [@NvTimLiu](https://github.com/NvTimLiu) +- Use experimental make_strings_children for json/csv writers ([#15599](https://github.com/NVIDIA/cudf/pull/15599)) [@davidwendt](https://github.com/davidwendt) +- Use experimental make_strings_children for strings join/url_encode/slice ([#15598](https://github.com/NVIDIA/cudf/pull/15598)) [@davidwendt](https://github.com/davidwendt) +- Use experimental make_strings_children in nvtext APIs ([#15595](https://github.com/NVIDIA/cudf/pull/15595)) [@davidwendt](https://github.com/davidwendt) +- Migrate to `{{ stdlib("c") }}` ([#15594](https://github.com/NVIDIA/cudf/pull/15594)) [@hcho3](https://github.com/hcho3) +- Deprecate `to/from_dask_dataframe` APIs in dask-cudf ([#15592](https://github.com/NVIDIA/cudf/pull/15592)) [@rjzamora](https://github.com/rjzamora) +- Minor fixups for future NumPy 2 compatibility ([#15590](https://github.com/NVIDIA/cudf/pull/15590)) [@seberg](https://github.com/seberg) +- Delay materializing RangeIndex in .reset_index ([#15588](https://github.com/NVIDIA/cudf/pull/15588)) [@mroeschke](https://github.com/mroeschke) +- Use experimental make_strings_children for capitalize/case/pad functions ([#15587](https://github.com/NVIDIA/cudf/pull/15587)) [@davidwendt](https://github.com/davidwendt) +- Use experimental make_strings_children for strings replace/filter/translate ([#15586](https://github.com/NVIDIA/cudf/pull/15586)) [@davidwendt](https://github.com/davidwendt) +- Add multithreaded parquet reader benchmarks. ([#15585](https://github.com/NVIDIA/cudf/pull/15585)) [@nvdbaranec](https://github.com/nvdbaranec) +- Don't materialize column during RangeIndex methods ([#15582](https://github.com/NVIDIA/cudf/pull/15582)) [@mroeschke](https://github.com/mroeschke) +- Improve performance for cudf::strings::count_re ([#15578](https://github.com/NVIDIA/cudf/pull/15578)) [@davidwendt](https://github.com/davidwendt) +- Replace RangeIndex._start/_stop/_step with _range ([#15576](https://github.com/NVIDIA/cudf/pull/15576)) [@mroeschke](https://github.com/mroeschke) +- add --rm and --name to devcontainer run args ([#15572](https://github.com/NVIDIA/cudf/pull/15572)) [@trxcllnt](https://github.com/trxcllnt) +- Change the default dictionary policy in Parquet writer from `ALWAYS` to `ADAPTIVE` ([#15570](https://github.com/NVIDIA/cudf/pull/15570)) [@mhaseeb123](https://github.com/mhaseeb123) +- Rename experimental JSON tests. ([#15568](https://github.com/NVIDIA/cudf/pull/15568)) [@bdice](https://github.com/bdice) +- Refactor JNI native dependency loading to allow returning of library path ([#15566](https://github.com/NVIDIA/cudf/pull/15566)) [@jlowe](https://github.com/jlowe) +- Remove protobuf and use parsed ORC statistics from libcudf ([#15564](https://github.com/NVIDIA/cudf/pull/15564)) [@bdice](https://github.com/bdice) +- Deprecate legacy JSON reader options. ([#15558](https://github.com/NVIDIA/cudf/pull/15558)) [@bdice](https://github.com/bdice) +- Use same .clang-format in cuDF JNI ([#15557](https://github.com/NVIDIA/cudf/pull/15557)) [@bdice](https://github.com/bdice) +- Large strings support for cudf::fill ([#15555](https://github.com/NVIDIA/cudf/pull/15555)) [@davidwendt](https://github.com/davidwendt) +- Upgrade upper bound pinning to `pandas-2.2.2` ([#15554](https://github.com/NVIDIA/cudf/pull/15554)) [@galipremsagar](https://github.com/galipremsagar) +- Work around issues with cccl main ([#15552](https://github.com/NVIDIA/cudf/pull/15552)) [@miscco](https://github.com/miscco) +- Enable pandas plotting unit tests for cudf.pandas ([#15547](https://github.com/NVIDIA/cudf/pull/15547)) [@mroeschke](https://github.com/mroeschke) +- Move timezone conversion logic to `DatetimeColumn` ([#15545](https://github.com/NVIDIA/cudf/pull/15545)) [@mroeschke](https://github.com/mroeschke) +- Large strings support for cudf::interleave_columns ([#15544](https://github.com/NVIDIA/cudf/pull/15544)) [@davidwendt](https://github.com/davidwendt) +- [skip ci] Switch back to 24.06 branch for pandas tests ([#15543](https://github.com/NVIDIA/cudf/pull/15543)) [@galipremsagar](https://github.com/galipremsagar) +- Remove checks dependency from static-configure test job. ([#15542](https://github.com/NVIDIA/cudf/pull/15542)) [@bdice](https://github.com/bdice) +- Remove legacy JSON reader from Python ([#15538](https://github.com/NVIDIA/cudf/pull/15538)) [@bdice](https://github.com/bdice) +- Enable more ignored pandas unit tests for cudf.pandas ([#15535](https://github.com/NVIDIA/cudf/pull/15535)) [@mroeschke](https://github.com/mroeschke) +- Large strings support for cudf::clamp ([#15533](https://github.com/NVIDIA/cudf/pull/15533)) [@davidwendt](https://github.com/davidwendt) +- Remove version hard-coding ([#15529](https://github.com/NVIDIA/cudf/pull/15529)) [@galipremsagar](https://github.com/galipremsagar) +- Removing all batching code from parquet writer ([#15528](https://github.com/NVIDIA/cudf/pull/15528)) [@mhaseeb123](https://github.com/mhaseeb123) +- Make some private class properties not settable ([#15527](https://github.com/NVIDIA/cudf/pull/15527)) [@mroeschke](https://github.com/mroeschke) +- Large strings support in regex replace APIs ([#15524](https://github.com/NVIDIA/cudf/pull/15524)) [@davidwendt](https://github.com/davidwendt) +- Skip pandas unit tests that crash pytest workers in `cudf.pandas` ([#15521](https://github.com/NVIDIA/cudf/pull/15521)) [@mroeschke](https://github.com/mroeschke) +- Preserve column metadata during more DataFrame operations ([#15519](https://github.com/NVIDIA/cudf/pull/15519)) [@mroeschke](https://github.com/mroeschke) +- Move to pandas-tests to a dedicated workflow file and trigger it from branch.yaml ([#15516](https://github.com/NVIDIA/cudf/pull/15516)) [@galipremsagar](https://github.com/galipremsagar) +- Large strings gtest fixture and utilities ([#15513](https://github.com/NVIDIA/cudf/pull/15513)) [@davidwendt](https://github.com/davidwendt) +- Convert libcudf resource parameters to rmm::device_async_resource_ref ([#15507](https://github.com/NVIDIA/cudf/pull/15507)) [@harrism](https://github.com/harrism) +- Relax protobuf lower bound to 3.20. ([#15506](https://github.com/NVIDIA/cudf/pull/15506)) [@bdice](https://github.com/bdice) +- Clean up index methods ([#15496](https://github.com/NVIDIA/cudf/pull/15496)) [@mroeschke](https://github.com/mroeschke) +- Update strings contains benchmarks to nvbench ([#15495](https://github.com/NVIDIA/cudf/pull/15495)) [@davidwendt](https://github.com/davidwendt) +- Update NVBench fixture to use new hooks, fix pinned memory segfault. ([#15492](https://github.com/NVIDIA/cudf/pull/15492)) [@alliepiper](https://github.com/alliepiper) +- Enable tests/scalar and test/series in cudf.pandas tests ([#15486](https://github.com/NVIDIA/cudf/pull/15486)) [@mroeschke](https://github.com/mroeschke) +- Clean up __cuda_array_interface__ handling in as_column ([#15477](https://github.com/NVIDIA/cudf/pull/15477)) [@mroeschke](https://github.com/mroeschke) +- Avoid .ordered and .categories from being settable in CategoricalColumn and CategoricalDtype ([#15475](https://github.com/NVIDIA/cudf/pull/15475)) [@mroeschke](https://github.com/mroeschke) +- Ignore pandas tests for cudf.pandas that need motoserver ([#15468](https://github.com/NVIDIA/cudf/pull/15468)) [@mroeschke](https://github.com/mroeschke) +- Use cached_property for NumericColumn.nan_count instead of ._nan_count variable ([#15466](https://github.com/NVIDIA/cudf/pull/15466)) [@mroeschke](https://github.com/mroeschke) +- Add to_arrow_device() functions that accept views ([#15465](https://github.com/NVIDIA/cudf/pull/15465)) [@davidwendt](https://github.com/davidwendt) +- Add custom status check workflow ([#15464](https://github.com/NVIDIA/cudf/pull/15464)) [@galipremsagar](https://github.com/galipremsagar) +- Disable pandas 2.x clipboard tests in cudf.pandas tests ([#15462](https://github.com/NVIDIA/cudf/pull/15462)) [@mroeschke](https://github.com/mroeschke) +- Enable tests/strings/test_api.py and tests/io/pytables in cudf.pandas tests ([#15461](https://github.com/NVIDIA/cudf/pull/15461)) [@mroeschke](https://github.com/mroeschke) +- Enable test_parsing in cudf.pandas tests ([#15460](https://github.com/NVIDIA/cudf/pull/15460)) [@mroeschke](https://github.com/mroeschke) +- Add `from_arrow_device` function to cudf interop using nanoarrow ([#15458](https://github.com/NVIDIA/cudf/pull/15458)) [@zeroshade](https://github.com/zeroshade) +- Remove deprecated strings offsets_begin ([#15454](https://github.com/NVIDIA/cudf/pull/15454)) [@davidwendt](https://github.com/davidwendt) +- Enable tests/windows/ in cudf.pandas tests ([#15444](https://github.com/NVIDIA/cudf/pull/15444)) [@mroeschke](https://github.com/mroeschke) +- Enable tests/interchange/test_impl.py in cudf.pandas tests ([#15443](https://github.com/NVIDIA/cudf/pull/15443)) [@mroeschke](https://github.com/mroeschke) +- Enable tests/io/test_user_agent.py in cudf pandas tests ([#15442](https://github.com/NVIDIA/cudf/pull/15442)) [@mroeschke](https://github.com/mroeschke) +- Performance improvement in libcudf case conversion for long strings ([#15441](https://github.com/NVIDIA/cudf/pull/15441)) [@davidwendt](https://github.com/davidwendt) +- Remove prior test skipping in run-pandas-tests with testing 2.2.1 ([#15440](https://github.com/NVIDIA/cudf/pull/15440)) [@mroeschke](https://github.com/mroeschke) +- Support orc and text IO with dask-expr using legacy conversion ([#15439](https://github.com/NVIDIA/cudf/pull/15439)) [@rjzamora](https://github.com/rjzamora) +- Floating <--> fixed-point conversion must now be called explicitly ([#15438](https://github.com/NVIDIA/cudf/pull/15438)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Unify Copy-On-Write and Spilling ([#15436](https://github.com/NVIDIA/cudf/pull/15436)) [@madsbk](https://github.com/madsbk) +- Enable ``dask_cudf`` json and s3 tests with query-planning on ([#15408](https://github.com/NVIDIA/cudf/pull/15408)) [@rjzamora](https://github.com/rjzamora) +- Bump ruff and codespell pre-commit checks ([#15407](https://github.com/NVIDIA/cudf/pull/15407)) [@mroeschke](https://github.com/mroeschke) +- Enable all tests for `arm` arch ([#15402](https://github.com/NVIDIA/cudf/pull/15402)) [@galipremsagar](https://github.com/galipremsagar) +- Bind `read_parquet_metadata` API to libcudf instead of pyarrow and extract `RowGroup` information ([#15398](https://github.com/NVIDIA/cudf/pull/15398)) [@mhaseeb123](https://github.com/mhaseeb123) +- Optimizing multi-source byte range reading in JSON reader ([#15396](https://github.com/NVIDIA/cudf/pull/15396)) [@shrshi](https://github.com/shrshi) +- add correct labels to pandas_function_request.md ([#15381](https://github.com/NVIDIA/cudf/pull/15381)) [@raybellwaves](https://github.com/raybellwaves) +- Remove deprecated hash() and spark_murmurhash3_x86_32() ([#15375](https://github.com/NVIDIA/cudf/pull/15375)) [@davidwendt](https://github.com/davidwendt) +- Large strings support in cudf::merge ([#15374](https://github.com/NVIDIA/cudf/pull/15374)) [@davidwendt](https://github.com/davidwendt) +- Enable test-reporting for pandas pytests in CI ([#15369](https://github.com/NVIDIA/cudf/pull/15369)) [@galipremsagar](https://github.com/galipremsagar) +- Use logical types in Parquet reader ([#15365](https://github.com/NVIDIA/cudf/pull/15365)) [@etseidl](https://github.com/etseidl) +- Add experimental make_strings_children utility ([#15363](https://github.com/NVIDIA/cudf/pull/15363)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-24.04 to branch-24.06 ([#15349](https://github.com/NVIDIA/cudf/pull/15349)) [@bdice](https://github.com/bdice) +- Fix CMake files in libcudf C++ examples to use existing libcudf build if present ([#15348](https://github.com/NVIDIA/cudf/pull/15348)) [@mhaseeb123](https://github.com/mhaseeb123) +- Use ruff pydocstyle over pydocstyle pre-commit hook ([#15345](https://github.com/NVIDIA/cudf/pull/15345)) [@mroeschke](https://github.com/mroeschke) +- Refactor stream mode setup for gtests ([#15337](https://github.com/NVIDIA/cudf/pull/15337)) [@davidwendt](https://github.com/davidwendt) +- Benchmark decimal <--> floating conversions. ([#15334](https://github.com/NVIDIA/cudf/pull/15334)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Avoid duplicate dask-cudf testing ([#15333](https://github.com/NVIDIA/cudf/pull/15333)) [@rjzamora](https://github.com/rjzamora) +- Skip decode steps in Parquet reader when nullable columns have no nulls ([#15332](https://github.com/NVIDIA/cudf/pull/15332)) [@etseidl](https://github.com/etseidl) +- Update udf_cpp to use rapids_cpm_cccl. ([#15331](https://github.com/NVIDIA/cudf/pull/15331)) [@bdice](https://github.com/bdice) +- Forward-merge branch-24.04 into branch-24.06 [skip ci] ([#15330](https://github.com/NVIDIA/cudf/pull/15330)) [@rapids-bot[bot]](https://github.com/rapids-bot[bot]) +- Allow ``numeric_only=True`` for simple groupby reductions ([#15326](https://github.com/NVIDIA/cudf/pull/15326)) [@rjzamora](https://github.com/rjzamora) +- Drop CentOS 7 support. ([#15323](https://github.com/NVIDIA/cudf/pull/15323)) [@bdice](https://github.com/bdice) +- Rework cudf::find_and_replace_all to use gather-based make_strings_column ([#15305](https://github.com/NVIDIA/cudf/pull/15305)) [@davidwendt](https://github.com/davidwendt) +- First pass at adding testing for pylibcudf ([#15300](https://github.com/NVIDIA/cudf/pull/15300)) [@vyasr](https://github.com/vyasr) +- [FEA] Performance improvement for mixed left semi/anti join ([#15288](https://github.com/NVIDIA/cudf/pull/15288)) [@tgujar](https://github.com/tgujar) +- Rework cudf::replace_nulls to use strings::detail::copy_if_else ([#15286](https://github.com/NVIDIA/cudf/pull/15286)) [@davidwendt](https://github.com/davidwendt) +- Clean up special casing in `as_column` for non-typed input ([#15276](https://github.com/NVIDIA/cudf/pull/15276)) [@mroeschke](https://github.com/mroeschke) +- Large strings support in cudf::concatenate ([#15195](https://github.com/NVIDIA/cudf/pull/15195)) [@davidwendt](https://github.com/davidwendt) +- Use less _is_categorical_dtype ([#15148](https://github.com/NVIDIA/cudf/pull/15148)) [@mroeschke](https://github.com/mroeschke) +- Align date_range defaults with pandas, support tz ([#15139](https://github.com/NVIDIA/cudf/pull/15139)) [@mroeschke](https://github.com/mroeschke) +- `ModuleAccelerator` performance: cache the result of checking if a caller is in the denylist ([#15056](https://github.com/NVIDIA/cudf/pull/15056)) [@shwina](https://github.com/shwina) +- Use offsetalator in cudf::strings::replace functions ([#14824](https://github.com/NVIDIA/cudf/pull/14824)) [@davidwendt](https://github.com/davidwendt) +- Cleanup some timedelta/datetime column logic ([#14715](https://github.com/NVIDIA/cudf/pull/14715)) [@mroeschke](https://github.com/mroeschke) +- Refactor numpy array input in as_column ([#14651](https://github.com/NVIDIA/cudf/pull/14651)) [@mroeschke](https://github.com/mroeschke) +- Refactor joins for conditional semis and antis ([#14646](https://github.com/NVIDIA/cudf/pull/14646)) [@DanialJavady96](https://github.com/DanialJavady96) +- Eagerly populate the class dict for cudf.pandas proxy types ([#14534](https://github.com/NVIDIA/cudf/pull/14534)) [@shwina](https://github.com/shwina) +- Some additional kernel thread index refactoring. ([#14107](https://github.com/NVIDIA/cudf/pull/14107)) [@bdice](https://github.com/bdice) # cuDF 24.04.00 (10 Apr 2024) ## 🚨 Breaking Changes -- Restructure pylibcudf/arrow interop facilities ([#15325](https://github.com/rapidsai/cudf/pull/15325)) [@vyasr](https://github.com/vyasr) -- Change exceptions thrown by copying APIs ([#15319](https://github.com/rapidsai/cudf/pull/15319)) [@vyasr](https://github.com/vyasr) -- Change strings_column_view::char_size to return int64 ([#15197](https://github.com/rapidsai/cudf/pull/15197)) [@davidwendt](https://github.com/davidwendt) -- Upgrade to `arrow-14.0.2` ([#15108](https://github.com/rapidsai/cudf/pull/15108)) [@galipremsagar](https://github.com/galipremsagar) -- Add support for `pandas-2.2` in `cudf` ([#15100](https://github.com/rapidsai/cudf/pull/15100)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate cudf::hashing::spark_murmurhash3_x86_32 ([#15074](https://github.com/rapidsai/cudf/pull/15074)) [@davidwendt](https://github.com/davidwendt) -- Align MultiIndex.get_indexder with pandas 2.2 change ([#15059](https://github.com/rapidsai/cudf/pull/15059)) [@mroeschke](https://github.com/mroeschke) -- Raise an error on import for unsupported GPUs. ([#15053](https://github.com/rapidsai/cudf/pull/15053)) [@bdice](https://github.com/bdice) -- Deprecate datelike isin casting strings to dates to match pandas 2.2 ([#15046](https://github.com/rapidsai/cudf/pull/15046)) [@mroeschke](https://github.com/mroeschke) -- Align concat Series name behavior in pandas 2.2 ([#15032](https://github.com/rapidsai/cudf/pull/15032)) [@mroeschke](https://github.com/mroeschke) -- Add `future_stack` to `DataFrame.stack` ([#15015](https://github.com/rapidsai/cudf/pull/15015)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate groupby fillna ([#15000](https://github.com/rapidsai/cudf/pull/15000)) [@mroeschke](https://github.com/mroeschke) -- Deprecate replace with categorical columns ([#14988](https://github.com/rapidsai/cudf/pull/14988)) [@mroeschke](https://github.com/mroeschke) -- Deprecate delim_whitespace in read_csv for pandas 2.2 ([#14986](https://github.com/rapidsai/cudf/pull/14986)) [@mroeschke](https://github.com/mroeschke) -- Deprecate parameters similar to pandas 2.2 ([#14984](https://github.com/rapidsai/cudf/pull/14984)) [@mroeschke](https://github.com/mroeschke) -- Add missing atomic operators, refactor atomic operators, move atomic operators to detail namespace. ([#14962](https://github.com/rapidsai/cudf/pull/14962)) [@bdice](https://github.com/bdice) -- Add `pandas-2.x` support in `cudf` ([#14916](https://github.com/rapidsai/cudf/pull/14916)) [@galipremsagar](https://github.com/galipremsagar) -- Use cuco::static_set in the hash-based groupby ([#14813](https://github.com/rapidsai/cudf/pull/14813)) [@PointKernel](https://github.com/PointKernel) +- Restructure pylibcudf/arrow interop facilities ([#15325](https://github.com/NVIDIA/cudf/pull/15325)) [@vyasr](https://github.com/vyasr) +- Change exceptions thrown by copying APIs ([#15319](https://github.com/NVIDIA/cudf/pull/15319)) [@vyasr](https://github.com/vyasr) +- Change strings_column_view::char_size to return int64 ([#15197](https://github.com/NVIDIA/cudf/pull/15197)) [@davidwendt](https://github.com/davidwendt) +- Upgrade to `arrow-14.0.2` ([#15108](https://github.com/NVIDIA/cudf/pull/15108)) [@galipremsagar](https://github.com/galipremsagar) +- Add support for `pandas-2.2` in `cudf` ([#15100](https://github.com/NVIDIA/cudf/pull/15100)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate cudf::hashing::spark_murmurhash3_x86_32 ([#15074](https://github.com/NVIDIA/cudf/pull/15074)) [@davidwendt](https://github.com/davidwendt) +- Align MultiIndex.get_indexder with pandas 2.2 change ([#15059](https://github.com/NVIDIA/cudf/pull/15059)) [@mroeschke](https://github.com/mroeschke) +- Raise an error on import for unsupported GPUs. ([#15053](https://github.com/NVIDIA/cudf/pull/15053)) [@bdice](https://github.com/bdice) +- Deprecate datelike isin casting strings to dates to match pandas 2.2 ([#15046](https://github.com/NVIDIA/cudf/pull/15046)) [@mroeschke](https://github.com/mroeschke) +- Align concat Series name behavior in pandas 2.2 ([#15032](https://github.com/NVIDIA/cudf/pull/15032)) [@mroeschke](https://github.com/mroeschke) +- Add `future_stack` to `DataFrame.stack` ([#15015](https://github.com/NVIDIA/cudf/pull/15015)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate groupby fillna ([#15000](https://github.com/NVIDIA/cudf/pull/15000)) [@mroeschke](https://github.com/mroeschke) +- Deprecate replace with categorical columns ([#14988](https://github.com/NVIDIA/cudf/pull/14988)) [@mroeschke](https://github.com/mroeschke) +- Deprecate delim_whitespace in read_csv for pandas 2.2 ([#14986](https://github.com/NVIDIA/cudf/pull/14986)) [@mroeschke](https://github.com/mroeschke) +- Deprecate parameters similar to pandas 2.2 ([#14984](https://github.com/NVIDIA/cudf/pull/14984)) [@mroeschke](https://github.com/mroeschke) +- Add missing atomic operators, refactor atomic operators, move atomic operators to detail namespace. ([#14962](https://github.com/NVIDIA/cudf/pull/14962)) [@bdice](https://github.com/bdice) +- Add `pandas-2.x` support in `cudf` ([#14916](https://github.com/NVIDIA/cudf/pull/14916)) [@galipremsagar](https://github.com/galipremsagar) +- Use cuco::static_set in the hash-based groupby ([#14813](https://github.com/NVIDIA/cudf/pull/14813)) [@PointKernel](https://github.com/PointKernel) ## 🐛 Bug Fixes -- Fix an issue with creating a series from scalar when `dtype='category'` ([#15476](https://github.com/rapidsai/cudf/pull/15476)) [@galipremsagar](https://github.com/galipremsagar) -- Update pre-commit-hooks to v0.0.3 ([#15355](https://github.com/rapidsai/cudf/pull/15355)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- [BUG][JNI] Trigger MemoryBuffer.onClosed after memory is freed ([#15351](https://github.com/rapidsai/cudf/pull/15351)) [@abellina](https://github.com/abellina) -- Fix an issue with multiple short list rowgroups using the Parquet chunked reader. ([#15342](https://github.com/rapidsai/cudf/pull/15342)) [@nvdbaranec](https://github.com/nvdbaranec) -- Avoid importing dask-expr if "query-planning" config is `False` ([#15340](https://github.com/rapidsai/cudf/pull/15340)) [@rjzamora](https://github.com/rjzamora) -- Fix gtests/ERROR_TEST errors when run in Debug ([#15317](https://github.com/rapidsai/cudf/pull/15317)) [@davidwendt](https://github.com/davidwendt) -- Fix OOB read in `inflate_kernel` ([#15309](https://github.com/rapidsai/cudf/pull/15309)) [@vuule](https://github.com/vuule) -- Work around a cuFile error when running CSV tests with memcheck ([#15293](https://github.com/rapidsai/cudf/pull/15293)) [@vuule](https://github.com/vuule) -- Fix Doxygen upload directory ([#15291](https://github.com/rapidsai/cudf/pull/15291)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Fix Doxygen check ([#15289](https://github.com/rapidsai/cudf/pull/15289)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Reintroduce PANDAS_GE_220 import ([#15287](https://github.com/rapidsai/cudf/pull/15287)) [@wence-](https://github.com/wence-) -- Fix mean computation for the geometric distribution in the data generator ([#15282](https://github.com/rapidsai/cudf/pull/15282)) [@vuule](https://github.com/vuule) -- Fix Parquet decimal64 stats ([#15281](https://github.com/rapidsai/cudf/pull/15281)) [@etseidl](https://github.com/etseidl) -- Make linking of nvtx3-cpp BUILD_LOCAL_INTERFACE ([#15271](https://github.com/rapidsai/cudf/pull/15271)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Workaround compute-sanitizer memcheck bug ([#15259](https://github.com/rapidsai/cudf/pull/15259)) [@davidwendt](https://github.com/davidwendt) -- Cleanup `hostdevice_vector` and add more APIs ([#15252](https://github.com/rapidsai/cudf/pull/15252)) [@ttnghia](https://github.com/ttnghia) -- Fix number of rows in randomly generated lists columns ([#15248](https://github.com/rapidsai/cudf/pull/15248)) [@vuule](https://github.com/vuule) -- Fix wrong output for `collect_list`/`collect_set` of lists column ([#15243](https://github.com/rapidsai/cudf/pull/15243)) [@ttnghia](https://github.com/ttnghia) -- Fix testchunkedPackTwoPasses to copy from the bounce buffer ([#15220](https://github.com/rapidsai/cudf/pull/15220)) [@abellina](https://github.com/abellina) -- Fix accessing `.columns` by an external API ([#15212](https://github.com/rapidsai/cudf/pull/15212)) [@galipremsagar](https://github.com/galipremsagar) -- [JNI] Disable testChunkedPackTwoPasses for now ([#15210](https://github.com/rapidsai/cudf/pull/15210)) [@abellina](https://github.com/abellina) -- Update labeler and codeowner configs for CMake files ([#15208](https://github.com/rapidsai/cudf/pull/15208)) [@PointKernel](https://github.com/PointKernel) -- Avoid dict normalization in ``__dask_tokenize__`` ([#15187](https://github.com/rapidsai/cudf/pull/15187)) [@rjzamora](https://github.com/rjzamora) -- Fix memcheck error in distinct inner join ([#15164](https://github.com/rapidsai/cudf/pull/15164)) [@PointKernel](https://github.com/PointKernel) -- Remove unneeded script parameters in test_cpp_memcheck.sh ([#15158](https://github.com/rapidsai/cudf/pull/15158)) [@davidwendt](https://github.com/davidwendt) -- Fix `ListColumn.to_pandas()` to retain `list` type ([#15155](https://github.com/rapidsai/cudf/pull/15155)) [@galipremsagar](https://github.com/galipremsagar) -- Avoid factorization in MultiIndex.to_pandas ([#15150](https://github.com/rapidsai/cudf/pull/15150)) [@mroeschke](https://github.com/mroeschke) -- Fix GroupBy.get_group and GroupBy.indices ([#15143](https://github.com/rapidsai/cudf/pull/15143)) [@wence-](https://github.com/wence-) -- Remove `const` from `range_window_bounds::_extent`. ([#15138](https://github.com/rapidsai/cudf/pull/15138)) [@mythrocks](https://github.com/mythrocks) -- DataFrame.columns = ... retains RangeIndex & set dtype ([#15129](https://github.com/rapidsai/cudf/pull/15129)) [@mroeschke](https://github.com/mroeschke) -- Correctly handle output for `GroupBy.apply` when chunk results are reindexed series ([#15109](https://github.com/rapidsai/cudf/pull/15109)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix Series.groupby.shift with a MultiIndex ([#15098](https://github.com/rapidsai/cudf/pull/15098)) [@mroeschke](https://github.com/mroeschke) -- Fix reductions when DataFrame has MulitIndex columns ([#15097](https://github.com/rapidsai/cudf/pull/15097)) [@mroeschke](https://github.com/mroeschke) -- Fix deprecation warnings for deprecated hash() calls ([#15095](https://github.com/rapidsai/cudf/pull/15095)) [@davidwendt](https://github.com/davidwendt) -- Add support for arrow `large_string` in `cudf` ([#15093](https://github.com/rapidsai/cudf/pull/15093)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `sort_values` pytest failure with pandas-2.x regression ([#15092](https://github.com/rapidsai/cudf/pull/15092)) [@galipremsagar](https://github.com/galipremsagar) -- Resolve path parsing issues in `get_json_object` ([#15082](https://github.com/rapidsai/cudf/pull/15082)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Fix bugs in handling of delta encodings ([#15075](https://github.com/rapidsai/cudf/pull/15075)) [@etseidl](https://github.com/etseidl) -- Fix `is_device_write_preferred` in `void_sink` and `user_sink_wrapper` ([#15064](https://github.com/rapidsai/cudf/pull/15064)) [@vuule](https://github.com/vuule) -- Eliminate duplicate allocation of nested string columns ([#15061](https://github.com/rapidsai/cudf/pull/15061)) [@vuule](https://github.com/vuule) -- Raise an error on import for unsupported GPUs. ([#15053](https://github.com/rapidsai/cudf/pull/15053)) [@bdice](https://github.com/bdice) -- Align concat Series name behavior in pandas 2.2 ([#15032](https://github.com/rapidsai/cudf/pull/15032)) [@mroeschke](https://github.com/mroeschke) -- Fix `Index.difference` to handle duplicate values when one of the inputs is empty ([#15016](https://github.com/rapidsai/cudf/pull/15016)) [@galipremsagar](https://github.com/galipremsagar) -- Add `future_stack` to `DataFrame.stack` ([#15015](https://github.com/rapidsai/cudf/pull/15015)) [@galipremsagar](https://github.com/galipremsagar) -- Fix handling of values=None in pylibcudf GroupBy.get_groups ([#14998](https://github.com/rapidsai/cudf/pull/14998)) [@shwina](https://github.com/shwina) -- Fix `DataFrame.sort_index` to respect `ignore_index` on all axis ([#14995](https://github.com/rapidsai/cudf/pull/14995)) [@galipremsagar](https://github.com/galipremsagar) -- Raise for pyarrow array that is tz-aware ([#14980](https://github.com/rapidsai/cudf/pull/14980)) [@mroeschke](https://github.com/mroeschke) -- Direct ``SeriesGroupBy.aggregate`` to ``SeriesGroupBy.agg`` ([#14971](https://github.com/rapidsai/cudf/pull/14971)) [@rjzamora](https://github.com/rjzamora) -- Respect IntervalDtype and CategoricalDtype objects passed by users ([#14961](https://github.com/rapidsai/cudf/pull/14961)) [@mroeschke](https://github.com/mroeschke) -- unset `CUDF_SPILL` after a pytest ([#14958](https://github.com/rapidsai/cudf/pull/14958)) [@galipremsagar](https://github.com/galipremsagar) -- Fix Null literals to be not parsed as string when mixed types as string is enabled in JSON reader ([#14939](https://github.com/rapidsai/cudf/pull/14939)) [@karthikeyann](https://github.com/karthikeyann) -- Fix chunked reads of Parquet delta encoded pages ([#14921](https://github.com/rapidsai/cudf/pull/14921)) [@etseidl](https://github.com/etseidl) -- Fix reading offset for data stream in ORC reader ([#14911](https://github.com/rapidsai/cudf/pull/14911)) [@ttnghia](https://github.com/ttnghia) -- Enable sanitizer check for a test case testORCReadAndWriteForDecimal128 ([#14897](https://github.com/rapidsai/cudf/pull/14897)) [@res-life](https://github.com/res-life) -- Fix dask token normalization ([#14829](https://github.com/rapidsai/cudf/pull/14829)) [@rjzamora](https://github.com/rjzamora) -- Fix 24.04 versions ([#14825](https://github.com/rapidsai/cudf/pull/14825)) [@raydouglass](https://github.com/raydouglass) -- Ensure slow private attrs are maybe proxies ([#14380](https://github.com/rapidsai/cudf/pull/14380)) [@mroeschke](https://github.com/mroeschke) +- Fix an issue with creating a series from scalar when `dtype='category'` ([#15476](https://github.com/NVIDIA/cudf/pull/15476)) [@galipremsagar](https://github.com/galipremsagar) +- Update pre-commit-hooks to v0.0.3 ([#15355](https://github.com/NVIDIA/cudf/pull/15355)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- [BUG][JNI] Trigger MemoryBuffer.onClosed after memory is freed ([#15351](https://github.com/NVIDIA/cudf/pull/15351)) [@abellina](https://github.com/abellina) +- Fix an issue with multiple short list rowgroups using the Parquet chunked reader. ([#15342](https://github.com/NVIDIA/cudf/pull/15342)) [@nvdbaranec](https://github.com/nvdbaranec) +- Avoid importing dask-expr if "query-planning" config is `False` ([#15340](https://github.com/NVIDIA/cudf/pull/15340)) [@rjzamora](https://github.com/rjzamora) +- Fix gtests/ERROR_TEST errors when run in Debug ([#15317](https://github.com/NVIDIA/cudf/pull/15317)) [@davidwendt](https://github.com/davidwendt) +- Fix OOB read in `inflate_kernel` ([#15309](https://github.com/NVIDIA/cudf/pull/15309)) [@vuule](https://github.com/vuule) +- Work around a cuFile error when running CSV tests with memcheck ([#15293](https://github.com/NVIDIA/cudf/pull/15293)) [@vuule](https://github.com/vuule) +- Fix Doxygen upload directory ([#15291](https://github.com/NVIDIA/cudf/pull/15291)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Fix Doxygen check ([#15289](https://github.com/NVIDIA/cudf/pull/15289)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Reintroduce PANDAS_GE_220 import ([#15287](https://github.com/NVIDIA/cudf/pull/15287)) [@wence-](https://github.com/wence-) +- Fix mean computation for the geometric distribution in the data generator ([#15282](https://github.com/NVIDIA/cudf/pull/15282)) [@vuule](https://github.com/vuule) +- Fix Parquet decimal64 stats ([#15281](https://github.com/NVIDIA/cudf/pull/15281)) [@etseidl](https://github.com/etseidl) +- Make linking of nvtx3-cpp BUILD_LOCAL_INTERFACE ([#15271](https://github.com/NVIDIA/cudf/pull/15271)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Workaround compute-sanitizer memcheck bug ([#15259](https://github.com/NVIDIA/cudf/pull/15259)) [@davidwendt](https://github.com/davidwendt) +- Cleanup `hostdevice_vector` and add more APIs ([#15252](https://github.com/NVIDIA/cudf/pull/15252)) [@ttnghia](https://github.com/ttnghia) +- Fix number of rows in randomly generated lists columns ([#15248](https://github.com/NVIDIA/cudf/pull/15248)) [@vuule](https://github.com/vuule) +- Fix wrong output for `collect_list`/`collect_set` of lists column ([#15243](https://github.com/NVIDIA/cudf/pull/15243)) [@ttnghia](https://github.com/ttnghia) +- Fix testchunkedPackTwoPasses to copy from the bounce buffer ([#15220](https://github.com/NVIDIA/cudf/pull/15220)) [@abellina](https://github.com/abellina) +- Fix accessing `.columns` by an external API ([#15212](https://github.com/NVIDIA/cudf/pull/15212)) [@galipremsagar](https://github.com/galipremsagar) +- [JNI] Disable testChunkedPackTwoPasses for now ([#15210](https://github.com/NVIDIA/cudf/pull/15210)) [@abellina](https://github.com/abellina) +- Update labeler and codeowner configs for CMake files ([#15208](https://github.com/NVIDIA/cudf/pull/15208)) [@PointKernel](https://github.com/PointKernel) +- Avoid dict normalization in ``__dask_tokenize__`` ([#15187](https://github.com/NVIDIA/cudf/pull/15187)) [@rjzamora](https://github.com/rjzamora) +- Fix memcheck error in distinct inner join ([#15164](https://github.com/NVIDIA/cudf/pull/15164)) [@PointKernel](https://github.com/PointKernel) +- Remove unneeded script parameters in test_cpp_memcheck.sh ([#15158](https://github.com/NVIDIA/cudf/pull/15158)) [@davidwendt](https://github.com/davidwendt) +- Fix `ListColumn.to_pandas()` to retain `list` type ([#15155](https://github.com/NVIDIA/cudf/pull/15155)) [@galipremsagar](https://github.com/galipremsagar) +- Avoid factorization in MultiIndex.to_pandas ([#15150](https://github.com/NVIDIA/cudf/pull/15150)) [@mroeschke](https://github.com/mroeschke) +- Fix GroupBy.get_group and GroupBy.indices ([#15143](https://github.com/NVIDIA/cudf/pull/15143)) [@wence-](https://github.com/wence-) +- Remove `const` from `range_window_bounds::_extent`. ([#15138](https://github.com/NVIDIA/cudf/pull/15138)) [@mythrocks](https://github.com/mythrocks) +- DataFrame.columns = ... retains RangeIndex & set dtype ([#15129](https://github.com/NVIDIA/cudf/pull/15129)) [@mroeschke](https://github.com/mroeschke) +- Correctly handle output for `GroupBy.apply` when chunk results are reindexed series ([#15109](https://github.com/NVIDIA/cudf/pull/15109)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix Series.groupby.shift with a MultiIndex ([#15098](https://github.com/NVIDIA/cudf/pull/15098)) [@mroeschke](https://github.com/mroeschke) +- Fix reductions when DataFrame has MulitIndex columns ([#15097](https://github.com/NVIDIA/cudf/pull/15097)) [@mroeschke](https://github.com/mroeschke) +- Fix deprecation warnings for deprecated hash() calls ([#15095](https://github.com/NVIDIA/cudf/pull/15095)) [@davidwendt](https://github.com/davidwendt) +- Add support for arrow `large_string` in `cudf` ([#15093](https://github.com/NVIDIA/cudf/pull/15093)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `sort_values` pytest failure with pandas-2.x regression ([#15092](https://github.com/NVIDIA/cudf/pull/15092)) [@galipremsagar](https://github.com/galipremsagar) +- Resolve path parsing issues in `get_json_object` ([#15082](https://github.com/NVIDIA/cudf/pull/15082)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Fix bugs in handling of delta encodings ([#15075](https://github.com/NVIDIA/cudf/pull/15075)) [@etseidl](https://github.com/etseidl) +- Fix `is_device_write_preferred` in `void_sink` and `user_sink_wrapper` ([#15064](https://github.com/NVIDIA/cudf/pull/15064)) [@vuule](https://github.com/vuule) +- Eliminate duplicate allocation of nested string columns ([#15061](https://github.com/NVIDIA/cudf/pull/15061)) [@vuule](https://github.com/vuule) +- Raise an error on import for unsupported GPUs. ([#15053](https://github.com/NVIDIA/cudf/pull/15053)) [@bdice](https://github.com/bdice) +- Align concat Series name behavior in pandas 2.2 ([#15032](https://github.com/NVIDIA/cudf/pull/15032)) [@mroeschke](https://github.com/mroeschke) +- Fix `Index.difference` to handle duplicate values when one of the inputs is empty ([#15016](https://github.com/NVIDIA/cudf/pull/15016)) [@galipremsagar](https://github.com/galipremsagar) +- Add `future_stack` to `DataFrame.stack` ([#15015](https://github.com/NVIDIA/cudf/pull/15015)) [@galipremsagar](https://github.com/galipremsagar) +- Fix handling of values=None in pylibcudf GroupBy.get_groups ([#14998](https://github.com/NVIDIA/cudf/pull/14998)) [@shwina](https://github.com/shwina) +- Fix `DataFrame.sort_index` to respect `ignore_index` on all axis ([#14995](https://github.com/NVIDIA/cudf/pull/14995)) [@galipremsagar](https://github.com/galipremsagar) +- Raise for pyarrow array that is tz-aware ([#14980](https://github.com/NVIDIA/cudf/pull/14980)) [@mroeschke](https://github.com/mroeschke) +- Direct ``SeriesGroupBy.aggregate`` to ``SeriesGroupBy.agg`` ([#14971](https://github.com/NVIDIA/cudf/pull/14971)) [@rjzamora](https://github.com/rjzamora) +- Respect IntervalDtype and CategoricalDtype objects passed by users ([#14961](https://github.com/NVIDIA/cudf/pull/14961)) [@mroeschke](https://github.com/mroeschke) +- unset `CUDF_SPILL` after a pytest ([#14958](https://github.com/NVIDIA/cudf/pull/14958)) [@galipremsagar](https://github.com/galipremsagar) +- Fix Null literals to be not parsed as string when mixed types as string is enabled in JSON reader ([#14939](https://github.com/NVIDIA/cudf/pull/14939)) [@karthikeyann](https://github.com/karthikeyann) +- Fix chunked reads of Parquet delta encoded pages ([#14921](https://github.com/NVIDIA/cudf/pull/14921)) [@etseidl](https://github.com/etseidl) +- Fix reading offset for data stream in ORC reader ([#14911](https://github.com/NVIDIA/cudf/pull/14911)) [@ttnghia](https://github.com/ttnghia) +- Enable sanitizer check for a test case testORCReadAndWriteForDecimal128 ([#14897](https://github.com/NVIDIA/cudf/pull/14897)) [@res-life](https://github.com/res-life) +- Fix dask token normalization ([#14829](https://github.com/NVIDIA/cudf/pull/14829)) [@rjzamora](https://github.com/rjzamora) +- Fix 24.04 versions ([#14825](https://github.com/NVIDIA/cudf/pull/14825)) [@raydouglass](https://github.com/raydouglass) +- Ensure slow private attrs are maybe proxies ([#14380](https://github.com/NVIDIA/cudf/pull/14380)) [@mroeschke](https://github.com/mroeschke) ## 📖 Documentation -- Ignore DLManagedTensor in the docs build ([#15392](https://github.com/rapidsai/cudf/pull/15392)) [@davidwendt](https://github.com/davidwendt) -- Revert "Temporarily disable docs errors. ([#15265)" (#15269](https://github.com/rapidsai/cudf/pull/15265)" (#15269)) [@bdice](https://github.com/bdice) -- Temporarily disable docs errors. ([#15265](https://github.com/rapidsai/cudf/pull/15265)) [@bdice](https://github.com/bdice) -- Update `developer_guide.md` with new guidance on quoted internal includes ([#15238](https://github.com/rapidsai/cudf/pull/15238)) [@harrism](https://github.com/harrism) -- Fix broken link for developer guide ([#15025](https://github.com/rapidsai/cudf/pull/15025)) [@sanjana098](https://github.com/sanjana098) -- [DOC] Update typo in docs example of structs_column_wrapper ([#14949](https://github.com/rapidsai/cudf/pull/14949)) [@karthikeyann](https://github.com/karthikeyann) -- Update cudf.pandas FAQ. ([#14940](https://github.com/rapidsai/cudf/pull/14940)) [@bdice](https://github.com/bdice) -- Optimize doc builds ([#14856](https://github.com/rapidsai/cudf/pull/14856)) [@vyasr](https://github.com/vyasr) -- Add developer guideline to use east const. ([#14836](https://github.com/rapidsai/cudf/pull/14836)) [@bdice](https://github.com/bdice) -- Document how cuDF is pronounced ([#14753](https://github.com/rapidsai/cudf/pull/14753)) [@pentschev](https://github.com/pentschev) -- Notes convert to Pandas-compat ([#12641](https://github.com/rapidsai/cudf/pull/12641)) [@Touutae-lab](https://github.com/Touutae-lab) +- Ignore DLManagedTensor in the docs build ([#15392](https://github.com/NVIDIA/cudf/pull/15392)) [@davidwendt](https://github.com/davidwendt) +- Revert "Temporarily disable docs errors. ([#15265)" (#15269](https://github.com/NVIDIA/cudf/pull/15265)" (#15269)) [@bdice](https://github.com/bdice) +- Temporarily disable docs errors. ([#15265](https://github.com/NVIDIA/cudf/pull/15265)) [@bdice](https://github.com/bdice) +- Update `developer_guide.md` with new guidance on quoted internal includes ([#15238](https://github.com/NVIDIA/cudf/pull/15238)) [@harrism](https://github.com/harrism) +- Fix broken link for developer guide ([#15025](https://github.com/NVIDIA/cudf/pull/15025)) [@sanjana098](https://github.com/sanjana098) +- [DOC] Update typo in docs example of structs_column_wrapper ([#14949](https://github.com/NVIDIA/cudf/pull/14949)) [@karthikeyann](https://github.com/karthikeyann) +- Update cudf.pandas FAQ. ([#14940](https://github.com/NVIDIA/cudf/pull/14940)) [@bdice](https://github.com/bdice) +- Optimize doc builds ([#14856](https://github.com/NVIDIA/cudf/pull/14856)) [@vyasr](https://github.com/vyasr) +- Add developer guideline to use east const. ([#14836](https://github.com/NVIDIA/cudf/pull/14836)) [@bdice](https://github.com/bdice) +- Document how cuDF is pronounced ([#14753](https://github.com/NVIDIA/cudf/pull/14753)) [@pentschev](https://github.com/pentschev) +- Notes convert to Pandas-compat ([#12641](https://github.com/NVIDIA/cudf/pull/12641)) [@Touutae-lab](https://github.com/Touutae-lab) ## 🚀 New Features -- Address inconsistency in single quote normalization in JSON reader ([#15324](https://github.com/rapidsai/cudf/pull/15324)) [@shrshi](https://github.com/shrshi) -- Use JNI pinned pool resource with cuIO ([#15255](https://github.com/rapidsai/cudf/pull/15255)) [@abellina](https://github.com/abellina) -- Add DELTA_BYTE_ARRAY encoder for Parquet ([#15239](https://github.com/rapidsai/cudf/pull/15239)) [@etseidl](https://github.com/etseidl) -- Migrate filling operations to pylibcudf ([#15225](https://github.com/rapidsai/cudf/pull/15225)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- [JNI] rmm based pinned pool ([#15219](https://github.com/rapidsai/cudf/pull/15219)) [@abellina](https://github.com/abellina) -- Implement zero-copy host buffer source instead of using an arrow implementation ([#15189](https://github.com/rapidsai/cudf/pull/15189)) [@vuule](https://github.com/vuule) -- Enable creation of columns from scalar ([#15181](https://github.com/rapidsai/cudf/pull/15181)) [@vyasr](https://github.com/vyasr) -- Use NVTX from GitHub. ([#15178](https://github.com/rapidsai/cudf/pull/15178)) [@bdice](https://github.com/bdice) -- Implement `segmented_row_bit_count` for computing row sizes by segments of rows ([#15169](https://github.com/rapidsai/cudf/pull/15169)) [@ttnghia](https://github.com/ttnghia) -- Implement search using pylibcudf ([#15166](https://github.com/rapidsai/cudf/pull/15166)) [@vyasr](https://github.com/vyasr) -- Add distinct left join ([#15149](https://github.com/rapidsai/cudf/pull/15149)) [@PointKernel](https://github.com/PointKernel) -- Add cardinality control for groupby benchs with flat types ([#15134](https://github.com/rapidsai/cudf/pull/15134)) [@PointKernel](https://github.com/PointKernel) -- Add ability to request Parquet encodings on a per-column basis ([#15081](https://github.com/rapidsai/cudf/pull/15081)) [@etseidl](https://github.com/etseidl) -- Automate include grouping order in .clang-format ([#15063](https://github.com/rapidsai/cudf/pull/15063)) [@harrism](https://github.com/harrism) -- Requesting a clean build directory also clears Jitify cache ([#15052](https://github.com/rapidsai/cudf/pull/15052)) [@robertmaynard](https://github.com/robertmaynard) -- API for JSON unquoted whitespace normalization ([#15033](https://github.com/rapidsai/cudf/pull/15033)) [@shrshi](https://github.com/shrshi) -- Implement concatenate, lists.explode, merge, sorting, and stream compaction in pylibcudf ([#15011](https://github.com/rapidsai/cudf/pull/15011)) [@vyasr](https://github.com/vyasr) -- Implement replace in pylibcudf ([#15005](https://github.com/rapidsai/cudf/pull/15005)) [@vyasr](https://github.com/vyasr) -- Add distinct key inner join ([#14990](https://github.com/rapidsai/cudf/pull/14990)) [@PointKernel](https://github.com/PointKernel) -- Implement rolling in pylibcudf ([#14982](https://github.com/rapidsai/cudf/pull/14982)) [@vyasr](https://github.com/vyasr) -- Implement joins in pylibcudf ([#14972](https://github.com/rapidsai/cudf/pull/14972)) [@vyasr](https://github.com/vyasr) -- Implement scans and reductions in pylibcudf ([#14970](https://github.com/rapidsai/cudf/pull/14970)) [@vyasr](https://github.com/vyasr) -- Rewrite cudf internals using pylibcudf groupby ([#14946](https://github.com/rapidsai/cudf/pull/14946)) [@vyasr](https://github.com/vyasr) -- Implement groupby in pylibcudf ([#14945](https://github.com/rapidsai/cudf/pull/14945)) [@vyasr](https://github.com/vyasr) -- Support casting of Map type to string in JSON reader ([#14936](https://github.com/rapidsai/cudf/pull/14936)) [@karthikeyann](https://github.com/karthikeyann) -- POC for whitespace removal in input JSON data using FST ([#14931](https://github.com/rapidsai/cudf/pull/14931)) [@shrshi](https://github.com/shrshi) -- Support for LZ4 compression in ORC and Parquet ([#14906](https://github.com/rapidsai/cudf/pull/14906)) [@vuule](https://github.com/vuule) -- Remove supports_streams from cuDF custom memory resources. ([#14857](https://github.com/rapidsai/cudf/pull/14857)) [@harrism](https://github.com/harrism) -- Migrate unary operations to pylibcudf ([#14850](https://github.com/rapidsai/cudf/pull/14850)) [@vyasr](https://github.com/vyasr) -- Migrate binary operations to pylibcudf ([#14821](https://github.com/rapidsai/cudf/pull/14821)) [@vyasr](https://github.com/vyasr) -- Add row index and stripe size options to Python ORC chunked writer ([#14785](https://github.com/rapidsai/cudf/pull/14785)) [@vuule](https://github.com/vuule) -- Support CUDA 12.2 ([#14712](https://github.com/rapidsai/cudf/pull/14712)) [@jameslamb](https://github.com/jameslamb) +- Address inconsistency in single quote normalization in JSON reader ([#15324](https://github.com/NVIDIA/cudf/pull/15324)) [@shrshi](https://github.com/shrshi) +- Use JNI pinned pool resource with cuIO ([#15255](https://github.com/NVIDIA/cudf/pull/15255)) [@abellina](https://github.com/abellina) +- Add DELTA_BYTE_ARRAY encoder for Parquet ([#15239](https://github.com/NVIDIA/cudf/pull/15239)) [@etseidl](https://github.com/etseidl) +- Migrate filling operations to pylibcudf ([#15225](https://github.com/NVIDIA/cudf/pull/15225)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- [JNI] rmm based pinned pool ([#15219](https://github.com/NVIDIA/cudf/pull/15219)) [@abellina](https://github.com/abellina) +- Implement zero-copy host buffer source instead of using an arrow implementation ([#15189](https://github.com/NVIDIA/cudf/pull/15189)) [@vuule](https://github.com/vuule) +- Enable creation of columns from scalar ([#15181](https://github.com/NVIDIA/cudf/pull/15181)) [@vyasr](https://github.com/vyasr) +- Use NVTX from GitHub. ([#15178](https://github.com/NVIDIA/cudf/pull/15178)) [@bdice](https://github.com/bdice) +- Implement `segmented_row_bit_count` for computing row sizes by segments of rows ([#15169](https://github.com/NVIDIA/cudf/pull/15169)) [@ttnghia](https://github.com/ttnghia) +- Implement search using pylibcudf ([#15166](https://github.com/NVIDIA/cudf/pull/15166)) [@vyasr](https://github.com/vyasr) +- Add distinct left join ([#15149](https://github.com/NVIDIA/cudf/pull/15149)) [@PointKernel](https://github.com/PointKernel) +- Add cardinality control for groupby benchs with flat types ([#15134](https://github.com/NVIDIA/cudf/pull/15134)) [@PointKernel](https://github.com/PointKernel) +- Add ability to request Parquet encodings on a per-column basis ([#15081](https://github.com/NVIDIA/cudf/pull/15081)) [@etseidl](https://github.com/etseidl) +- Automate include grouping order in .clang-format ([#15063](https://github.com/NVIDIA/cudf/pull/15063)) [@harrism](https://github.com/harrism) +- Requesting a clean build directory also clears Jitify cache ([#15052](https://github.com/NVIDIA/cudf/pull/15052)) [@robertmaynard](https://github.com/robertmaynard) +- API for JSON unquoted whitespace normalization ([#15033](https://github.com/NVIDIA/cudf/pull/15033)) [@shrshi](https://github.com/shrshi) +- Implement concatenate, lists.explode, merge, sorting, and stream compaction in pylibcudf ([#15011](https://github.com/NVIDIA/cudf/pull/15011)) [@vyasr](https://github.com/vyasr) +- Implement replace in pylibcudf ([#15005](https://github.com/NVIDIA/cudf/pull/15005)) [@vyasr](https://github.com/vyasr) +- Add distinct key inner join ([#14990](https://github.com/NVIDIA/cudf/pull/14990)) [@PointKernel](https://github.com/PointKernel) +- Implement rolling in pylibcudf ([#14982](https://github.com/NVIDIA/cudf/pull/14982)) [@vyasr](https://github.com/vyasr) +- Implement joins in pylibcudf ([#14972](https://github.com/NVIDIA/cudf/pull/14972)) [@vyasr](https://github.com/vyasr) +- Implement scans and reductions in pylibcudf ([#14970](https://github.com/NVIDIA/cudf/pull/14970)) [@vyasr](https://github.com/vyasr) +- Rewrite cudf internals using pylibcudf groupby ([#14946](https://github.com/NVIDIA/cudf/pull/14946)) [@vyasr](https://github.com/vyasr) +- Implement groupby in pylibcudf ([#14945](https://github.com/NVIDIA/cudf/pull/14945)) [@vyasr](https://github.com/vyasr) +- Support casting of Map type to string in JSON reader ([#14936](https://github.com/NVIDIA/cudf/pull/14936)) [@karthikeyann](https://github.com/karthikeyann) +- POC for whitespace removal in input JSON data using FST ([#14931](https://github.com/NVIDIA/cudf/pull/14931)) [@shrshi](https://github.com/shrshi) +- Support for LZ4 compression in ORC and Parquet ([#14906](https://github.com/NVIDIA/cudf/pull/14906)) [@vuule](https://github.com/vuule) +- Remove supports_streams from cuDF custom memory resources. ([#14857](https://github.com/NVIDIA/cudf/pull/14857)) [@harrism](https://github.com/harrism) +- Migrate unary operations to pylibcudf ([#14850](https://github.com/NVIDIA/cudf/pull/14850)) [@vyasr](https://github.com/vyasr) +- Migrate binary operations to pylibcudf ([#14821](https://github.com/NVIDIA/cudf/pull/14821)) [@vyasr](https://github.com/vyasr) +- Add row index and stripe size options to Python ORC chunked writer ([#14785](https://github.com/NVIDIA/cudf/pull/14785)) [@vuule](https://github.com/vuule) +- Support CUDA 12.2 ([#14712](https://github.com/NVIDIA/cudf/pull/14712)) [@jameslamb](https://github.com/jameslamb) ## 🛠️ Improvements -- Use `conda env create --yes` instead of `--force` ([#15403](https://github.com/rapidsai/cudf/pull/15403)) [@bdice](https://github.com/bdice) -- Restructure pylibcudf/arrow interop facilities ([#15325](https://github.com/rapidsai/cudf/pull/15325)) [@vyasr](https://github.com/vyasr) -- Change exceptions thrown by copying APIs ([#15319](https://github.com/rapidsai/cudf/pull/15319)) [@vyasr](https://github.com/vyasr) -- Enable branch testing for `cudf.pandas` ([#15316](https://github.com/rapidsai/cudf/pull/15316)) [@galipremsagar](https://github.com/galipremsagar) -- Replace black with ruff-format ([#15312](https://github.com/rapidsai/cudf/pull/15312)) [@mroeschke](https://github.com/mroeschke) -- This fixes an NPE when trying to read empty JSON data by adding a new API for missing information ([#15307](https://github.com/rapidsai/cudf/pull/15307)) [@revans2](https://github.com/revans2) -- Address poor performance of Parquet string decoding ([#15304](https://github.com/rapidsai/cudf/pull/15304)) [@etseidl](https://github.com/etseidl) -- Update script input name ([#15301](https://github.com/rapidsai/cudf/pull/15301)) [@AyodeAwe](https://github.com/AyodeAwe) -- Make test_read_parquet_partitioned_filtered data deterministic ([#15296](https://github.com/rapidsai/cudf/pull/15296)) [@mroeschke](https://github.com/mroeschke) -- Add timeout for `cudf.pandas` pandas tests ([#15284](https://github.com/rapidsai/cudf/pull/15284)) [@galipremsagar](https://github.com/galipremsagar) -- Add upper bound to prevent usage of NumPy 2 ([#15283](https://github.com/rapidsai/cudf/pull/15283)) [@bdice](https://github.com/bdice) -- Fix cudf::test::to_host return of host_vector ([#15263](https://github.com/rapidsai/cudf/pull/15263)) [@davidwendt](https://github.com/davidwendt) -- Implement grouped product scan ([#15254](https://github.com/rapidsai/cudf/pull/15254)) [@wence-](https://github.com/wence-) -- Add CUDA 12.4 to supported PTX versions ([#15247](https://github.com/rapidsai/cudf/pull/15247)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Implement DataFrame|Series.squeeze ([#15244](https://github.com/rapidsai/cudf/pull/15244)) [@mroeschke](https://github.com/mroeschke) -- Roll back ipow changes due to register pressure. ([#15242](https://github.com/rapidsai/cudf/pull/15242)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Remove create_chars_child_column utility ([#15241](https://github.com/rapidsai/cudf/pull/15241)) [@davidwendt](https://github.com/davidwendt) -- Update dlpack to version 0.8 ([#15237](https://github.com/rapidsai/cudf/pull/15237)) [@dantegd](https://github.com/dantegd) -- Improve performance in JSON reader when `mixed_types_as_string` option is enabled ([#15236](https://github.com/rapidsai/cudf/pull/15236)) [@shrshi](https://github.com/shrshi) -- Remove row conversion code from libcudf ([#15234](https://github.com/rapidsai/cudf/pull/15234)) [@ttnghia](https://github.com/ttnghia) -- Use variable substitution for RAPIDS version in Doxyfile ([#15231](https://github.com/rapidsai/cudf/pull/15231)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Add ListColumns.to_pandas(arrow_type=) ([#15228](https://github.com/rapidsai/cudf/pull/15228)) [@mroeschke](https://github.com/mroeschke) -- Treat dask-cudf CI artifacts as pure wheels ([#15223](https://github.com/rapidsai/cudf/pull/15223)) [@bdice](https://github.com/bdice) -- Clean up usage of __CUDA_ARCH__ and other macros. ([#15218](https://github.com/rapidsai/cudf/pull/15218)) [@bdice](https://github.com/bdice) -- DOC: use constants in performance-comparisons.ipynb ([#15215](https://github.com/rapidsai/cudf/pull/15215)) [@raybellwaves](https://github.com/raybellwaves) -- Rewrite conversion in terms of column ([#15213](https://github.com/rapidsai/cudf/pull/15213)) [@vyasr](https://github.com/vyasr) -- Switch `pytest-xdist` algo to `worksteal` ([#15207](https://github.com/rapidsai/cudf/pull/15207)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate strings_column_view::offsets_begin() ([#15205](https://github.com/rapidsai/cudf/pull/15205)) [@davidwendt](https://github.com/davidwendt) -- Add `get_upstream_resource` method to `stream_checking_resource_adaptor` ([#15203](https://github.com/rapidsai/cudf/pull/15203)) [@miscco](https://github.com/miscco) -- Tune up row size estimation in the data generator ([#15202](https://github.com/rapidsai/cudf/pull/15202)) [@vuule](https://github.com/vuule) -- Fix `offset` value for generating test data in `parquet_chunked_reader_test.cu` ([#15200](https://github.com/rapidsai/cudf/pull/15200)) [@ttnghia](https://github.com/ttnghia) -- Change strings_column_view::char_size to return int64 ([#15197](https://github.com/rapidsai/cudf/pull/15197)) [@davidwendt](https://github.com/davidwendt) -- Fix includes for row_operators.cuh ([#15194](https://github.com/rapidsai/cudf/pull/15194)) [@davidwendt](https://github.com/davidwendt) -- Generalize GHA selectors for pure Python testing ([#15191](https://github.com/rapidsai/cudf/pull/15191)) [@bdice](https://github.com/bdice) -- Improvements for `__cuda_array_interface__` tests ([#15188](https://github.com/rapidsai/cudf/pull/15188)) [@bdice](https://github.com/bdice) -- Allow to_pandas to return pandas.ArrowDtype ([#15182](https://github.com/rapidsai/cudf/pull/15182)) [@mroeschke](https://github.com/mroeschke) -- Ignore `byte_range` in `read_json` when the size is not smaller than the input data ([#15180](https://github.com/rapidsai/cudf/pull/15180)) [@vuule](https://github.com/vuule) -- Expose new stable_sort and finish stream_compaction in pylibcudf ([#15175](https://github.com/rapidsai/cudf/pull/15175)) [@wence-](https://github.com/wence-) -- [ci] update matrix filters for dask-cudf builds ([#15174](https://github.com/rapidsai/cudf/pull/15174)) [@jameslamb](https://github.com/jameslamb) -- Change make_strings_children to return uvector ([#15171](https://github.com/rapidsai/cudf/pull/15171)) [@davidwendt](https://github.com/davidwendt) -- Don't override to_pandas for Datelike columns ([#15167](https://github.com/rapidsai/cudf/pull/15167)) [@mroeschke](https://github.com/mroeschke) -- Drop python-snappy from dependencies. ([#15161](https://github.com/rapidsai/cudf/pull/15161)) [@bdice](https://github.com/bdice) -- Add microkernels for fixed-width and fixed-width dictionary in Parquet decode ([#15159](https://github.com/rapidsai/cudf/pull/15159)) [@abellina](https://github.com/abellina) -- Make HostColumnVector.DataType accessor methods public ([#15157](https://github.com/rapidsai/cudf/pull/15157)) [@jbrennan333](https://github.com/jbrennan333) -- Java bindings for left outer distinct join ([#15154](https://github.com/rapidsai/cudf/pull/15154)) [@jlowe](https://github.com/jlowe) -- Forward-merge branch-24.02 to branch-24.04 ([#15153](https://github.com/rapidsai/cudf/pull/15153)) [@bdice](https://github.com/bdice) -- Enable pandas pytests for `cudf.pandas` ([#15147](https://github.com/rapidsai/cudf/pull/15147)) [@galipremsagar](https://github.com/galipremsagar) -- Add java option to keep quotes for JSON reads ([#15146](https://github.com/rapidsai/cudf/pull/15146)) [@revans2](https://github.com/revans2) -- Change cross-pandas-version testing in `cudf` ([#15145](https://github.com/rapidsai/cudf/pull/15145)) [@galipremsagar](https://github.com/galipremsagar) -- Use `hostdevice_vector` in `kernel_error` to avoid the pageable copy ([#15140](https://github.com/rapidsai/cudf/pull/15140)) [@vuule](https://github.com/vuule) -- Clean up Columns.astype & cudf.dtype ([#15125](https://github.com/rapidsai/cudf/pull/15125)) [@mroeschke](https://github.com/mroeschke) -- Simplify some to_pandas implementations ([#15123](https://github.com/rapidsai/cudf/pull/15123)) [@mroeschke](https://github.com/mroeschke) -- Java: Add leak tracking for Scalar instances ([#15121](https://github.com/rapidsai/cudf/pull/15121)) [@jlowe](https://github.com/jlowe) -- Remove calls to strings_column_view::offsets_begin() ([#15112](https://github.com/rapidsai/cudf/pull/15112)) [@davidwendt](https://github.com/davidwendt) -- Add support for Python 3.11, require NumPy 1.23+ ([#15111](https://github.com/rapidsai/cudf/pull/15111)) [@jameslamb](https://github.com/jameslamb) -- Compile-time ipow computation with array lookup ([#15110](https://github.com/rapidsai/cudf/pull/15110)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) -- Upgrade to `arrow-14.0.2` ([#15108](https://github.com/rapidsai/cudf/pull/15108)) [@galipremsagar](https://github.com/galipremsagar) -- Dynamically set version in RAPIDS doc builds ([#15101](https://github.com/rapidsai/cudf/pull/15101)) [@jakirkham](https://github.com/jakirkham) -- Add support for `pandas-2.2` in `cudf` ([#15100](https://github.com/rapidsai/cudf/pull/15100)) [@galipremsagar](https://github.com/galipremsagar) -- Update devcontainers to CUDA Toolkit 12.2 ([#15099](https://github.com/rapidsai/cudf/pull/15099)) [@trxcllnt](https://github.com/trxcllnt) -- Fix `datetime` binop pytest failures in pandas-2.2 ([#15090](https://github.com/rapidsai/cudf/pull/15090)) [@galipremsagar](https://github.com/galipremsagar) -- Validate types in pylibcudf Column/Table constructors ([#15088](https://github.com/rapidsai/cudf/pull/15088)) [@wence-](https://github.com/wence-) -- xfail test_join_ordering_pandas_compat for pandas 2.2 ([#15080](https://github.com/rapidsai/cudf/pull/15080)) [@mroeschke](https://github.com/mroeschke) -- Add general purpose host memory allocator reference to cuIO with a demo of pooled-pinned allocation. ([#15079](https://github.com/rapidsai/cudf/pull/15079)) [@nvdbaranec](https://github.com/nvdbaranec) -- Adjust test_binops for pandas 2.2 ([#15078](https://github.com/rapidsai/cudf/pull/15078)) [@mroeschke](https://github.com/mroeschke) -- Remove offsets_begin() call from nvtext::generate_ngrams ([#15077](https://github.com/rapidsai/cudf/pull/15077)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::detail::has_nonempty_null_rows ([#15076](https://github.com/rapidsai/cudf/pull/15076)) [@davidwendt](https://github.com/davidwendt) -- Deprecate cudf::hashing::spark_murmurhash3_x86_32 ([#15074](https://github.com/rapidsai/cudf/pull/15074)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf::test::to_host to handle both offset types for strings columns ([#15073](https://github.com/rapidsai/cudf/pull/15073)) [@davidwendt](https://github.com/davidwendt) -- Add condition for test_groupby_nulls_basic in pandas 2.2 ([#15072](https://github.com/rapidsai/cudf/pull/15072)) [@mroeschke](https://github.com/mroeschke) -- xfail tests in test_udf_masked_ops due to pandas 2.2 bug ([#15071](https://github.com/rapidsai/cudf/pull/15071)) [@mroeschke](https://github.com/mroeschke) -- target branch-24.04 for GitHub Actions workflows ([#15069](https://github.com/rapidsai/cudf/pull/15069)) [@jameslamb](https://github.com/jameslamb) -- Implement stable version of `cudf::sort` ([#15066](https://github.com/rapidsai/cudf/pull/15066)) [@wence-](https://github.com/wence-) -- Fix ORC and JSON tests failures for pandas 2.2 ([#15062](https://github.com/rapidsai/cudf/pull/15062)) [@mroeschke](https://github.com/mroeschke) -- Adjust test_joining for pandas 2.2 ([#15060](https://github.com/rapidsai/cudf/pull/15060)) [@mroeschke](https://github.com/mroeschke) -- Align MultiIndex.get_indexder with pandas 2.2 change ([#15059](https://github.com/rapidsai/cudf/pull/15059)) [@mroeschke](https://github.com/mroeschke) -- Fix test_resample index dtype checking for pandas 2.2 ([#15058](https://github.com/rapidsai/cudf/pull/15058)) [@mroeschke](https://github.com/mroeschke) -- Split out strings/replace.cu and rework its gtests ([#15054](https://github.com/rapidsai/cudf/pull/15054)) [@davidwendt](https://github.com/davidwendt) -- Avoid incompatible value type setting in test_rolling for pandas 2.2 ([#15050](https://github.com/rapidsai/cudf/pull/15050)) [@mroeschke](https://github.com/mroeschke) -- Change chained replace inplace test to COW test for pandas 2.2 ([#15049](https://github.com/rapidsai/cudf/pull/15049)) [@mroeschke](https://github.com/mroeschke) -- Deprecate datelike isin casting strings to dates to match pandas 2.2 ([#15046](https://github.com/rapidsai/cudf/pull/15046)) [@mroeschke](https://github.com/mroeschke) -- Avoid chained indexing in test_indexing for pandas 2.2 ([#15045](https://github.com/rapidsai/cudf/pull/15045)) [@mroeschke](https://github.com/mroeschke) -- Avoid pandas 2.2 `DeprecationWarning` in test_hdf ([#15044](https://github.com/rapidsai/cudf/pull/15044)) [@mroeschke](https://github.com/mroeschke) -- Use appropriate make_offsets_child_column for building lists columns ([#15043](https://github.com/rapidsai/cudf/pull/15043)) [@davidwendt](https://github.com/davidwendt) -- Factor out position-offsets logic from strings split_helper utility ([#15040](https://github.com/rapidsai/cudf/pull/15040)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-24.02 to branch-24.04 ([#15039](https://github.com/rapidsai/cudf/pull/15039)) [@bdice](https://github.com/bdice) -- Clean up nvtx macros ([#15038](https://github.com/rapidsai/cudf/pull/15038)) [@PointKernel](https://github.com/PointKernel) -- Add xfailures for test_applymap for pandas 2.2 ([#15034](https://github.com/rapidsai/cudf/pull/15034)) [@mroeschke](https://github.com/mroeschke) -- Expose libcudf filter expression in read_parquet ([#15028](https://github.com/rapidsai/cudf/pull/15028)) [@wence-](https://github.com/wence-) -- Adjust tests in test_dataframe.py for pandas 2.2 ([#15023](https://github.com/rapidsai/cudf/pull/15023)) [@mroeschke](https://github.com/mroeschke) -- Adjust test_datetime_infer_format for pandas 2.2 ([#15021](https://github.com/rapidsai/cudf/pull/15021)) [@mroeschke](https://github.com/mroeschke) -- Performance optimizations for parquet sub-rowgroup reader. ([#15020](https://github.com/rapidsai/cudf/pull/15020)) [@nvdbaranec](https://github.com/nvdbaranec) -- JNI bindings for distinct_hash_join ([#15019](https://github.com/rapidsai/cudf/pull/15019)) [@jlowe](https://github.com/jlowe) -- Change copy_if_safe to call thrust instead of the overload function ([#15018](https://github.com/rapidsai/cudf/pull/15018)) [@davidwendt](https://github.com/davidwendt) -- Improve performance of copy_if_else for long strings ([#15017](https://github.com/rapidsai/cudf/pull/15017)) [@davidwendt](https://github.com/davidwendt) -- Fix is_string_dtype test for pandas 2.2 ([#15012](https://github.com/rapidsai/cudf/pull/15012)) [@mroeschke](https://github.com/mroeschke) -- Rework cudf::strings::detail::copy_range for offsetalator ([#15010](https://github.com/rapidsai/cudf/pull/15010)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::get_json_object() ([#15009](https://github.com/rapidsai/cudf/pull/15009)) [@davidwendt](https://github.com/davidwendt) -- Align integral types in ORC to specs ([#15008](https://github.com/rapidsai/cudf/pull/15008)) [@vuule](https://github.com/vuule) -- Clean up detail sequence header inclusion ([#15007](https://github.com/rapidsai/cudf/pull/15007)) [@PointKernel](https://github.com/PointKernel) -- Add groupby.apply(include_groups=) to match pandas 2.2 deprecation ([#15006](https://github.com/rapidsai/cudf/pull/15006)) [@mroeschke](https://github.com/mroeschke) -- Use offsetalator in cudf::interleave_columns() ([#15004](https://github.com/rapidsai/cudf/pull/15004)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::row_bit_count() ([#15003](https://github.com/rapidsai/cudf/pull/15003)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::strings::wrap() ([#15002](https://github.com/rapidsai/cudf/pull/15002)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::strings::reverse ([#15001](https://github.com/rapidsai/cudf/pull/15001)) [@davidwendt](https://github.com/davidwendt) -- Deprecate groupby fillna ([#15000](https://github.com/rapidsai/cudf/pull/15000)) [@mroeschke](https://github.com/mroeschke) -- Ensure to_* IO methods respect pandas 2.2 keyword only deprecation ([#14999](https://github.com/rapidsai/cudf/pull/14999)) [@mroeschke](https://github.com/mroeschke) -- Remove unneeded calls to create_chars_child_column utility ([#14997](https://github.com/rapidsai/cudf/pull/14997)) [@davidwendt](https://github.com/davidwendt) -- Add environment-agnostic scripts for running ctests and pytests ([#14992](https://github.com/rapidsai/cudf/pull/14992)) [@trxcllnt](https://github.com/trxcllnt) -- Filter all `DeprecationWarning`'s by `ArrowTable.to_pandas()` ([#14989](https://github.com/rapidsai/cudf/pull/14989)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate replace with categorical columns ([#14988](https://github.com/rapidsai/cudf/pull/14988)) [@mroeschke](https://github.com/mroeschke) -- Deprecate delim_whitespace in read_csv for pandas 2.2 ([#14986](https://github.com/rapidsai/cudf/pull/14986)) [@mroeschke](https://github.com/mroeschke) -- Deprecate parameters similar to pandas 2.2 ([#14984](https://github.com/rapidsai/cudf/pull/14984)) [@mroeschke](https://github.com/mroeschke) -- Ensure that `ctest` is called with `--no-tests=error`. ([#14983](https://github.com/rapidsai/cudf/pull/14983)) [@bdice](https://github.com/bdice) -- Deprecate non-integer `periods` in `date_range` and `interval_range` ([#14976](https://github.com/rapidsai/cudf/pull/14976)) [@galipremsagar](https://github.com/galipremsagar) -- Update ops-bot.yaml ([#14974](https://github.com/rapidsai/cudf/pull/14974)) [@AyodeAwe](https://github.com/AyodeAwe) -- Use page statistics in Parquet reader ([#14973](https://github.com/rapidsai/cudf/pull/14973)) [@etseidl](https://github.com/etseidl) -- Use fused types for overloaded function signatures ([#14969](https://github.com/rapidsai/cudf/pull/14969)) [@vyasr](https://github.com/vyasr) -- Deprecate certain frequency strings ([#14967](https://github.com/rapidsai/cudf/pull/14967)) [@galipremsagar](https://github.com/galipremsagar) -- Update copyrights for 24.04. ([#14964](https://github.com/rapidsai/cudf/pull/14964)) [@bdice](https://github.com/bdice) -- Add missing atomic operators, refactor atomic operators, move atomic operators to detail namespace. ([#14962](https://github.com/rapidsai/cudf/pull/14962)) [@bdice](https://github.com/bdice) -- Introduce `GetJsonObjectOptions` in `getJSONObject` Java API ([#14956](https://github.com/rapidsai/cudf/pull/14956)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- JNI JSON read with DataSource and infered schema, along with basic java nested Schema JSON reads ([#14954](https://github.com/rapidsai/cudf/pull/14954)) [@revans2](https://github.com/revans2) -- Make codecov only informational (always pass). ([#14952](https://github.com/rapidsai/cudf/pull/14952)) [@bdice](https://github.com/bdice) -- Replace legacy cudf and dask_cudf imports as (d)gd ([#14944](https://github.com/rapidsai/cudf/pull/14944)) [@mroeschke](https://github.com/mroeschke) -- Replace _is_datetime64tz/interval_dtype with isinstance ([#14943](https://github.com/rapidsai/cudf/pull/14943)) [@mroeschke](https://github.com/mroeschke) -- Update tests for pandas 2. ([#14941](https://github.com/rapidsai/cudf/pull/14941)) [@bdice](https://github.com/bdice) -- Use more public pandas APIs ([#14929](https://github.com/rapidsai/cudf/pull/14929)) [@mroeschke](https://github.com/mroeschke) -- Replace local copyright check with pre-commit-hooks verify-copyright ([#14917](https://github.com/rapidsai/cudf/pull/14917)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Add `pandas-2.x` support in `cudf` ([#14916](https://github.com/rapidsai/cudf/pull/14916)) [@galipremsagar](https://github.com/galipremsagar) -- Use offsetalator in nvtext::byte_pair_encoding ([#14888](https://github.com/rapidsai/cudf/pull/14888)) [@davidwendt](https://github.com/davidwendt) -- De-DOS line-endings ([#14880](https://github.com/rapidsai/cudf/pull/14880)) [@wence-](https://github.com/wence-) -- Add detail `cuco_allocator` ([#14877](https://github.com/rapidsai/cudf/pull/14877)) [@PointKernel](https://github.com/PointKernel) -- Move all core types to using enum class in Cython ([#14876](https://github.com/rapidsai/cudf/pull/14876)) [@vyasr](https://github.com/vyasr) -- Read `cudf.__version__` in Sphinx build ([#14872](https://github.com/rapidsai/cudf/pull/14872)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Use int64 offset types for accessing code-points in nvtext::normalize ([#14868](https://github.com/rapidsai/cudf/pull/14868)) [@davidwendt](https://github.com/davidwendt) -- Read version from VERSION file in CMake ([#14867](https://github.com/rapidsai/cudf/pull/14867)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Update conda-cpp-post-build-checks to branch-24.04. ([#14854](https://github.com/rapidsai/cudf/pull/14854)) [@bdice](https://github.com/bdice) -- Update cudf for compatibility with the latest cuco ([#14849](https://github.com/rapidsai/cudf/pull/14849)) [@PointKernel](https://github.com/PointKernel) -- Remove deprecated strings functions ([#14848](https://github.com/rapidsai/cudf/pull/14848)) [@davidwendt](https://github.com/davidwendt) -- Fix CI workflows for pandas-tests and add test summary. ([#14847](https://github.com/rapidsai/cudf/pull/14847)) [@bdice](https://github.com/bdice) -- Use offsetalator in cudf::strings::copy_slice ([#14844](https://github.com/rapidsai/cudf/pull/14844)) [@davidwendt](https://github.com/davidwendt) -- Fix V2 Parquet page alignment for use with zStandard compression ([#14841](https://github.com/rapidsai/cudf/pull/14841)) [@etseidl](https://github.com/etseidl) -- Fix calls to deprecated strings factory API in examples. ([#14838](https://github.com/rapidsai/cudf/pull/14838)) [@bdice](https://github.com/bdice) -- Update pre-commit hooks ([#14837](https://github.com/rapidsai/cudf/pull/14837)) [@bdice](https://github.com/bdice) -- Use `rapids_cuda_set_runtime` to determine cuda runtime usage by target ([#14833](https://github.com/rapidsai/cudf/pull/14833)) [@vyasr](https://github.com/vyasr) -- Remove get_mem_info functions from custom memory resources ([#14832](https://github.com/rapidsai/cudf/pull/14832)) [@harrism](https://github.com/harrism) -- Fix debug build by splitting row_operator_tests_utilities.cu ([#14826](https://github.com/rapidsai/cudf/pull/14826)) [@davidwendt](https://github.com/davidwendt) -- Remove -DNVBench_ENABLE_CUPTI=OFF. ([#14820](https://github.com/rapidsai/cudf/pull/14820)) [@bdice](https://github.com/bdice) -- Use cuco::static_set in the hash-based groupby ([#14813](https://github.com/rapidsai/cudf/pull/14813)) [@PointKernel](https://github.com/PointKernel) -- Branch 24.04 merge branch 24.02 ([#14809](https://github.com/rapidsai/cudf/pull/14809)) [@vyasr](https://github.com/vyasr) -- Branch 24.04 merge branch 24.02 ([#14806](https://github.com/rapidsai/cudf/pull/14806)) [@vyasr](https://github.com/vyasr) -- Introduce basic "cudf" backend for Dask Expressions ([#14805](https://github.com/rapidsai/cudf/pull/14805)) [@rjzamora](https://github.com/rjzamora) -- Remove `build_struct|list_column` ([#14786](https://github.com/rapidsai/cudf/pull/14786)) [@mroeschke](https://github.com/mroeschke) -- Use offsetalator in nvtext tokenize functions ([#14783](https://github.com/rapidsai/cudf/pull/14783)) [@davidwendt](https://github.com/davidwendt) -- Reduce execution time of Python ORC tests ([#14776](https://github.com/rapidsai/cudf/pull/14776)) [@vuule](https://github.com/vuule) -- Use offsetalator in cudf::strings::split functions ([#14757](https://github.com/rapidsai/cudf/pull/14757)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::strings::findall ([#14745](https://github.com/rapidsai/cudf/pull/14745)) [@davidwendt](https://github.com/davidwendt) -- Use offsetalator in cudf::strings::url_decode ([#14744](https://github.com/rapidsai/cudf/pull/14744)) [@davidwendt](https://github.com/davidwendt) -- Use get_offset_value utility in strings shift function ([#14743](https://github.com/rapidsai/cudf/pull/14743)) [@davidwendt](https://github.com/davidwendt) -- Use as_column instead of full ([#14698](https://github.com/rapidsai/cudf/pull/14698)) [@mroeschke](https://github.com/mroeschke) -- List all notable breaking changes ([#13535](https://github.com/rapidsai/cudf/pull/13535)) [@galipremsagar](https://github.com/galipremsagar) +- Use `conda env create --yes` instead of `--force` ([#15403](https://github.com/NVIDIA/cudf/pull/15403)) [@bdice](https://github.com/bdice) +- Restructure pylibcudf/arrow interop facilities ([#15325](https://github.com/NVIDIA/cudf/pull/15325)) [@vyasr](https://github.com/vyasr) +- Change exceptions thrown by copying APIs ([#15319](https://github.com/NVIDIA/cudf/pull/15319)) [@vyasr](https://github.com/vyasr) +- Enable branch testing for `cudf.pandas` ([#15316](https://github.com/NVIDIA/cudf/pull/15316)) [@galipremsagar](https://github.com/galipremsagar) +- Replace black with ruff-format ([#15312](https://github.com/NVIDIA/cudf/pull/15312)) [@mroeschke](https://github.com/mroeschke) +- This fixes an NPE when trying to read empty JSON data by adding a new API for missing information ([#15307](https://github.com/NVIDIA/cudf/pull/15307)) [@revans2](https://github.com/revans2) +- Address poor performance of Parquet string decoding ([#15304](https://github.com/NVIDIA/cudf/pull/15304)) [@etseidl](https://github.com/etseidl) +- Update script input name ([#15301](https://github.com/NVIDIA/cudf/pull/15301)) [@AyodeAwe](https://github.com/AyodeAwe) +- Make test_read_parquet_partitioned_filtered data deterministic ([#15296](https://github.com/NVIDIA/cudf/pull/15296)) [@mroeschke](https://github.com/mroeschke) +- Add timeout for `cudf.pandas` pandas tests ([#15284](https://github.com/NVIDIA/cudf/pull/15284)) [@galipremsagar](https://github.com/galipremsagar) +- Add upper bound to prevent usage of NumPy 2 ([#15283](https://github.com/NVIDIA/cudf/pull/15283)) [@bdice](https://github.com/bdice) +- Fix cudf::test::to_host return of host_vector ([#15263](https://github.com/NVIDIA/cudf/pull/15263)) [@davidwendt](https://github.com/davidwendt) +- Implement grouped product scan ([#15254](https://github.com/NVIDIA/cudf/pull/15254)) [@wence-](https://github.com/wence-) +- Add CUDA 12.4 to supported PTX versions ([#15247](https://github.com/NVIDIA/cudf/pull/15247)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Implement DataFrame|Series.squeeze ([#15244](https://github.com/NVIDIA/cudf/pull/15244)) [@mroeschke](https://github.com/mroeschke) +- Roll back ipow changes due to register pressure. ([#15242](https://github.com/NVIDIA/cudf/pull/15242)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Remove create_chars_child_column utility ([#15241](https://github.com/NVIDIA/cudf/pull/15241)) [@davidwendt](https://github.com/davidwendt) +- Update dlpack to version 0.8 ([#15237](https://github.com/NVIDIA/cudf/pull/15237)) [@dantegd](https://github.com/dantegd) +- Improve performance in JSON reader when `mixed_types_as_string` option is enabled ([#15236](https://github.com/NVIDIA/cudf/pull/15236)) [@shrshi](https://github.com/shrshi) +- Remove row conversion code from libcudf ([#15234](https://github.com/NVIDIA/cudf/pull/15234)) [@ttnghia](https://github.com/ttnghia) +- Use variable substitution for RAPIDS version in Doxyfile ([#15231](https://github.com/NVIDIA/cudf/pull/15231)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Add ListColumns.to_pandas(arrow_type=) ([#15228](https://github.com/NVIDIA/cudf/pull/15228)) [@mroeschke](https://github.com/mroeschke) +- Treat dask-cudf CI artifacts as pure wheels ([#15223](https://github.com/NVIDIA/cudf/pull/15223)) [@bdice](https://github.com/bdice) +- Clean up usage of __CUDA_ARCH__ and other macros. ([#15218](https://github.com/NVIDIA/cudf/pull/15218)) [@bdice](https://github.com/bdice) +- DOC: use constants in performance-comparisons.ipynb ([#15215](https://github.com/NVIDIA/cudf/pull/15215)) [@raybellwaves](https://github.com/raybellwaves) +- Rewrite conversion in terms of column ([#15213](https://github.com/NVIDIA/cudf/pull/15213)) [@vyasr](https://github.com/vyasr) +- Switch `pytest-xdist` algo to `worksteal` ([#15207](https://github.com/NVIDIA/cudf/pull/15207)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate strings_column_view::offsets_begin() ([#15205](https://github.com/NVIDIA/cudf/pull/15205)) [@davidwendt](https://github.com/davidwendt) +- Add `get_upstream_resource` method to `stream_checking_resource_adaptor` ([#15203](https://github.com/NVIDIA/cudf/pull/15203)) [@miscco](https://github.com/miscco) +- Tune up row size estimation in the data generator ([#15202](https://github.com/NVIDIA/cudf/pull/15202)) [@vuule](https://github.com/vuule) +- Fix `offset` value for generating test data in `parquet_chunked_reader_test.cu` ([#15200](https://github.com/NVIDIA/cudf/pull/15200)) [@ttnghia](https://github.com/ttnghia) +- Change strings_column_view::char_size to return int64 ([#15197](https://github.com/NVIDIA/cudf/pull/15197)) [@davidwendt](https://github.com/davidwendt) +- Fix includes for row_operators.cuh ([#15194](https://github.com/NVIDIA/cudf/pull/15194)) [@davidwendt](https://github.com/davidwendt) +- Generalize GHA selectors for pure Python testing ([#15191](https://github.com/NVIDIA/cudf/pull/15191)) [@bdice](https://github.com/bdice) +- Improvements for `__cuda_array_interface__` tests ([#15188](https://github.com/NVIDIA/cudf/pull/15188)) [@bdice](https://github.com/bdice) +- Allow to_pandas to return pandas.ArrowDtype ([#15182](https://github.com/NVIDIA/cudf/pull/15182)) [@mroeschke](https://github.com/mroeschke) +- Ignore `byte_range` in `read_json` when the size is not smaller than the input data ([#15180](https://github.com/NVIDIA/cudf/pull/15180)) [@vuule](https://github.com/vuule) +- Expose new stable_sort and finish stream_compaction in pylibcudf ([#15175](https://github.com/NVIDIA/cudf/pull/15175)) [@wence-](https://github.com/wence-) +- [ci] update matrix filters for dask-cudf builds ([#15174](https://github.com/NVIDIA/cudf/pull/15174)) [@jameslamb](https://github.com/jameslamb) +- Change make_strings_children to return uvector ([#15171](https://github.com/NVIDIA/cudf/pull/15171)) [@davidwendt](https://github.com/davidwendt) +- Don't override to_pandas for Datelike columns ([#15167](https://github.com/NVIDIA/cudf/pull/15167)) [@mroeschke](https://github.com/mroeschke) +- Drop python-snappy from dependencies. ([#15161](https://github.com/NVIDIA/cudf/pull/15161)) [@bdice](https://github.com/bdice) +- Add microkernels for fixed-width and fixed-width dictionary in Parquet decode ([#15159](https://github.com/NVIDIA/cudf/pull/15159)) [@abellina](https://github.com/abellina) +- Make HostColumnVector.DataType accessor methods public ([#15157](https://github.com/NVIDIA/cudf/pull/15157)) [@jbrennan333](https://github.com/jbrennan333) +- Java bindings for left outer distinct join ([#15154](https://github.com/NVIDIA/cudf/pull/15154)) [@jlowe](https://github.com/jlowe) +- Forward-merge branch-24.02 to branch-24.04 ([#15153](https://github.com/NVIDIA/cudf/pull/15153)) [@bdice](https://github.com/bdice) +- Enable pandas pytests for `cudf.pandas` ([#15147](https://github.com/NVIDIA/cudf/pull/15147)) [@galipremsagar](https://github.com/galipremsagar) +- Add java option to keep quotes for JSON reads ([#15146](https://github.com/NVIDIA/cudf/pull/15146)) [@revans2](https://github.com/revans2) +- Change cross-pandas-version testing in `cudf` ([#15145](https://github.com/NVIDIA/cudf/pull/15145)) [@galipremsagar](https://github.com/galipremsagar) +- Use `hostdevice_vector` in `kernel_error` to avoid the pageable copy ([#15140](https://github.com/NVIDIA/cudf/pull/15140)) [@vuule](https://github.com/vuule) +- Clean up Columns.astype & cudf.dtype ([#15125](https://github.com/NVIDIA/cudf/pull/15125)) [@mroeschke](https://github.com/mroeschke) +- Simplify some to_pandas implementations ([#15123](https://github.com/NVIDIA/cudf/pull/15123)) [@mroeschke](https://github.com/mroeschke) +- Java: Add leak tracking for Scalar instances ([#15121](https://github.com/NVIDIA/cudf/pull/15121)) [@jlowe](https://github.com/jlowe) +- Remove calls to strings_column_view::offsets_begin() ([#15112](https://github.com/NVIDIA/cudf/pull/15112)) [@davidwendt](https://github.com/davidwendt) +- Add support for Python 3.11, require NumPy 1.23+ ([#15111](https://github.com/NVIDIA/cudf/pull/15111)) [@jameslamb](https://github.com/jameslamb) +- Compile-time ipow computation with array lookup ([#15110](https://github.com/NVIDIA/cudf/pull/15110)) [@pmattione-nvidia](https://github.com/pmattione-nvidia) +- Upgrade to `arrow-14.0.2` ([#15108](https://github.com/NVIDIA/cudf/pull/15108)) [@galipremsagar](https://github.com/galipremsagar) +- Dynamically set version in RAPIDS doc builds ([#15101](https://github.com/NVIDIA/cudf/pull/15101)) [@jakirkham](https://github.com/jakirkham) +- Add support for `pandas-2.2` in `cudf` ([#15100](https://github.com/NVIDIA/cudf/pull/15100)) [@galipremsagar](https://github.com/galipremsagar) +- Update devcontainers to CUDA Toolkit 12.2 ([#15099](https://github.com/NVIDIA/cudf/pull/15099)) [@trxcllnt](https://github.com/trxcllnt) +- Fix `datetime` binop pytest failures in pandas-2.2 ([#15090](https://github.com/NVIDIA/cudf/pull/15090)) [@galipremsagar](https://github.com/galipremsagar) +- Validate types in pylibcudf Column/Table constructors ([#15088](https://github.com/NVIDIA/cudf/pull/15088)) [@wence-](https://github.com/wence-) +- xfail test_join_ordering_pandas_compat for pandas 2.2 ([#15080](https://github.com/NVIDIA/cudf/pull/15080)) [@mroeschke](https://github.com/mroeschke) +- Add general purpose host memory allocator reference to cuIO with a demo of pooled-pinned allocation. ([#15079](https://github.com/NVIDIA/cudf/pull/15079)) [@nvdbaranec](https://github.com/nvdbaranec) +- Adjust test_binops for pandas 2.2 ([#15078](https://github.com/NVIDIA/cudf/pull/15078)) [@mroeschke](https://github.com/mroeschke) +- Remove offsets_begin() call from nvtext::generate_ngrams ([#15077](https://github.com/NVIDIA/cudf/pull/15077)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::detail::has_nonempty_null_rows ([#15076](https://github.com/NVIDIA/cudf/pull/15076)) [@davidwendt](https://github.com/davidwendt) +- Deprecate cudf::hashing::spark_murmurhash3_x86_32 ([#15074](https://github.com/NVIDIA/cudf/pull/15074)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf::test::to_host to handle both offset types for strings columns ([#15073](https://github.com/NVIDIA/cudf/pull/15073)) [@davidwendt](https://github.com/davidwendt) +- Add condition for test_groupby_nulls_basic in pandas 2.2 ([#15072](https://github.com/NVIDIA/cudf/pull/15072)) [@mroeschke](https://github.com/mroeschke) +- xfail tests in test_udf_masked_ops due to pandas 2.2 bug ([#15071](https://github.com/NVIDIA/cudf/pull/15071)) [@mroeschke](https://github.com/mroeschke) +- target branch-24.04 for GitHub Actions workflows ([#15069](https://github.com/NVIDIA/cudf/pull/15069)) [@jameslamb](https://github.com/jameslamb) +- Implement stable version of `cudf::sort` ([#15066](https://github.com/NVIDIA/cudf/pull/15066)) [@wence-](https://github.com/wence-) +- Fix ORC and JSON tests failures for pandas 2.2 ([#15062](https://github.com/NVIDIA/cudf/pull/15062)) [@mroeschke](https://github.com/mroeschke) +- Adjust test_joining for pandas 2.2 ([#15060](https://github.com/NVIDIA/cudf/pull/15060)) [@mroeschke](https://github.com/mroeschke) +- Align MultiIndex.get_indexder with pandas 2.2 change ([#15059](https://github.com/NVIDIA/cudf/pull/15059)) [@mroeschke](https://github.com/mroeschke) +- Fix test_resample index dtype checking for pandas 2.2 ([#15058](https://github.com/NVIDIA/cudf/pull/15058)) [@mroeschke](https://github.com/mroeschke) +- Split out strings/replace.cu and rework its gtests ([#15054](https://github.com/NVIDIA/cudf/pull/15054)) [@davidwendt](https://github.com/davidwendt) +- Avoid incompatible value type setting in test_rolling for pandas 2.2 ([#15050](https://github.com/NVIDIA/cudf/pull/15050)) [@mroeschke](https://github.com/mroeschke) +- Change chained replace inplace test to COW test for pandas 2.2 ([#15049](https://github.com/NVIDIA/cudf/pull/15049)) [@mroeschke](https://github.com/mroeschke) +- Deprecate datelike isin casting strings to dates to match pandas 2.2 ([#15046](https://github.com/NVIDIA/cudf/pull/15046)) [@mroeschke](https://github.com/mroeschke) +- Avoid chained indexing in test_indexing for pandas 2.2 ([#15045](https://github.com/NVIDIA/cudf/pull/15045)) [@mroeschke](https://github.com/mroeschke) +- Avoid pandas 2.2 `DeprecationWarning` in test_hdf ([#15044](https://github.com/NVIDIA/cudf/pull/15044)) [@mroeschke](https://github.com/mroeschke) +- Use appropriate make_offsets_child_column for building lists columns ([#15043](https://github.com/NVIDIA/cudf/pull/15043)) [@davidwendt](https://github.com/davidwendt) +- Factor out position-offsets logic from strings split_helper utility ([#15040](https://github.com/NVIDIA/cudf/pull/15040)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-24.02 to branch-24.04 ([#15039](https://github.com/NVIDIA/cudf/pull/15039)) [@bdice](https://github.com/bdice) +- Clean up nvtx macros ([#15038](https://github.com/NVIDIA/cudf/pull/15038)) [@PointKernel](https://github.com/PointKernel) +- Add xfailures for test_applymap for pandas 2.2 ([#15034](https://github.com/NVIDIA/cudf/pull/15034)) [@mroeschke](https://github.com/mroeschke) +- Expose libcudf filter expression in read_parquet ([#15028](https://github.com/NVIDIA/cudf/pull/15028)) [@wence-](https://github.com/wence-) +- Adjust tests in test_dataframe.py for pandas 2.2 ([#15023](https://github.com/NVIDIA/cudf/pull/15023)) [@mroeschke](https://github.com/mroeschke) +- Adjust test_datetime_infer_format for pandas 2.2 ([#15021](https://github.com/NVIDIA/cudf/pull/15021)) [@mroeschke](https://github.com/mroeschke) +- Performance optimizations for parquet sub-rowgroup reader. ([#15020](https://github.com/NVIDIA/cudf/pull/15020)) [@nvdbaranec](https://github.com/nvdbaranec) +- JNI bindings for distinct_hash_join ([#15019](https://github.com/NVIDIA/cudf/pull/15019)) [@jlowe](https://github.com/jlowe) +- Change copy_if_safe to call thrust instead of the overload function ([#15018](https://github.com/NVIDIA/cudf/pull/15018)) [@davidwendt](https://github.com/davidwendt) +- Improve performance of copy_if_else for long strings ([#15017](https://github.com/NVIDIA/cudf/pull/15017)) [@davidwendt](https://github.com/davidwendt) +- Fix is_string_dtype test for pandas 2.2 ([#15012](https://github.com/NVIDIA/cudf/pull/15012)) [@mroeschke](https://github.com/mroeschke) +- Rework cudf::strings::detail::copy_range for offsetalator ([#15010](https://github.com/NVIDIA/cudf/pull/15010)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::get_json_object() ([#15009](https://github.com/NVIDIA/cudf/pull/15009)) [@davidwendt](https://github.com/davidwendt) +- Align integral types in ORC to specs ([#15008](https://github.com/NVIDIA/cudf/pull/15008)) [@vuule](https://github.com/vuule) +- Clean up detail sequence header inclusion ([#15007](https://github.com/NVIDIA/cudf/pull/15007)) [@PointKernel](https://github.com/PointKernel) +- Add groupby.apply(include_groups=) to match pandas 2.2 deprecation ([#15006](https://github.com/NVIDIA/cudf/pull/15006)) [@mroeschke](https://github.com/mroeschke) +- Use offsetalator in cudf::interleave_columns() ([#15004](https://github.com/NVIDIA/cudf/pull/15004)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::row_bit_count() ([#15003](https://github.com/NVIDIA/cudf/pull/15003)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::strings::wrap() ([#15002](https://github.com/NVIDIA/cudf/pull/15002)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::strings::reverse ([#15001](https://github.com/NVIDIA/cudf/pull/15001)) [@davidwendt](https://github.com/davidwendt) +- Deprecate groupby fillna ([#15000](https://github.com/NVIDIA/cudf/pull/15000)) [@mroeschke](https://github.com/mroeschke) +- Ensure to_* IO methods respect pandas 2.2 keyword only deprecation ([#14999](https://github.com/NVIDIA/cudf/pull/14999)) [@mroeschke](https://github.com/mroeschke) +- Remove unneeded calls to create_chars_child_column utility ([#14997](https://github.com/NVIDIA/cudf/pull/14997)) [@davidwendt](https://github.com/davidwendt) +- Add environment-agnostic scripts for running ctests and pytests ([#14992](https://github.com/NVIDIA/cudf/pull/14992)) [@trxcllnt](https://github.com/trxcllnt) +- Filter all `DeprecationWarning`'s by `ArrowTable.to_pandas()` ([#14989](https://github.com/NVIDIA/cudf/pull/14989)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate replace with categorical columns ([#14988](https://github.com/NVIDIA/cudf/pull/14988)) [@mroeschke](https://github.com/mroeschke) +- Deprecate delim_whitespace in read_csv for pandas 2.2 ([#14986](https://github.com/NVIDIA/cudf/pull/14986)) [@mroeschke](https://github.com/mroeschke) +- Deprecate parameters similar to pandas 2.2 ([#14984](https://github.com/NVIDIA/cudf/pull/14984)) [@mroeschke](https://github.com/mroeschke) +- Ensure that `ctest` is called with `--no-tests=error`. ([#14983](https://github.com/NVIDIA/cudf/pull/14983)) [@bdice](https://github.com/bdice) +- Deprecate non-integer `periods` in `date_range` and `interval_range` ([#14976](https://github.com/NVIDIA/cudf/pull/14976)) [@galipremsagar](https://github.com/galipremsagar) +- Update ops-bot.yaml ([#14974](https://github.com/NVIDIA/cudf/pull/14974)) [@AyodeAwe](https://github.com/AyodeAwe) +- Use page statistics in Parquet reader ([#14973](https://github.com/NVIDIA/cudf/pull/14973)) [@etseidl](https://github.com/etseidl) +- Use fused types for overloaded function signatures ([#14969](https://github.com/NVIDIA/cudf/pull/14969)) [@vyasr](https://github.com/vyasr) +- Deprecate certain frequency strings ([#14967](https://github.com/NVIDIA/cudf/pull/14967)) [@galipremsagar](https://github.com/galipremsagar) +- Update copyrights for 24.04. ([#14964](https://github.com/NVIDIA/cudf/pull/14964)) [@bdice](https://github.com/bdice) +- Add missing atomic operators, refactor atomic operators, move atomic operators to detail namespace. ([#14962](https://github.com/NVIDIA/cudf/pull/14962)) [@bdice](https://github.com/bdice) +- Introduce `GetJsonObjectOptions` in `getJSONObject` Java API ([#14956](https://github.com/NVIDIA/cudf/pull/14956)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- JNI JSON read with DataSource and infered schema, along with basic java nested Schema JSON reads ([#14954](https://github.com/NVIDIA/cudf/pull/14954)) [@revans2](https://github.com/revans2) +- Make codecov only informational (always pass). ([#14952](https://github.com/NVIDIA/cudf/pull/14952)) [@bdice](https://github.com/bdice) +- Replace legacy cudf and dask_cudf imports as (d)gd ([#14944](https://github.com/NVIDIA/cudf/pull/14944)) [@mroeschke](https://github.com/mroeschke) +- Replace _is_datetime64tz/interval_dtype with isinstance ([#14943](https://github.com/NVIDIA/cudf/pull/14943)) [@mroeschke](https://github.com/mroeschke) +- Update tests for pandas 2. ([#14941](https://github.com/NVIDIA/cudf/pull/14941)) [@bdice](https://github.com/bdice) +- Use more public pandas APIs ([#14929](https://github.com/NVIDIA/cudf/pull/14929)) [@mroeschke](https://github.com/mroeschke) +- Replace local copyright check with pre-commit-hooks verify-copyright ([#14917](https://github.com/NVIDIA/cudf/pull/14917)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Add `pandas-2.x` support in `cudf` ([#14916](https://github.com/NVIDIA/cudf/pull/14916)) [@galipremsagar](https://github.com/galipremsagar) +- Use offsetalator in nvtext::byte_pair_encoding ([#14888](https://github.com/NVIDIA/cudf/pull/14888)) [@davidwendt](https://github.com/davidwendt) +- De-DOS line-endings ([#14880](https://github.com/NVIDIA/cudf/pull/14880)) [@wence-](https://github.com/wence-) +- Add detail `cuco_allocator` ([#14877](https://github.com/NVIDIA/cudf/pull/14877)) [@PointKernel](https://github.com/PointKernel) +- Move all core types to using enum class in Cython ([#14876](https://github.com/NVIDIA/cudf/pull/14876)) [@vyasr](https://github.com/vyasr) +- Read `cudf.__version__` in Sphinx build ([#14872](https://github.com/NVIDIA/cudf/pull/14872)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Use int64 offset types for accessing code-points in nvtext::normalize ([#14868](https://github.com/NVIDIA/cudf/pull/14868)) [@davidwendt](https://github.com/davidwendt) +- Read version from VERSION file in CMake ([#14867](https://github.com/NVIDIA/cudf/pull/14867)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Update conda-cpp-post-build-checks to branch-24.04. ([#14854](https://github.com/NVIDIA/cudf/pull/14854)) [@bdice](https://github.com/bdice) +- Update cudf for compatibility with the latest cuco ([#14849](https://github.com/NVIDIA/cudf/pull/14849)) [@PointKernel](https://github.com/PointKernel) +- Remove deprecated strings functions ([#14848](https://github.com/NVIDIA/cudf/pull/14848)) [@davidwendt](https://github.com/davidwendt) +- Fix CI workflows for pandas-tests and add test summary. ([#14847](https://github.com/NVIDIA/cudf/pull/14847)) [@bdice](https://github.com/bdice) +- Use offsetalator in cudf::strings::copy_slice ([#14844](https://github.com/NVIDIA/cudf/pull/14844)) [@davidwendt](https://github.com/davidwendt) +- Fix V2 Parquet page alignment for use with zStandard compression ([#14841](https://github.com/NVIDIA/cudf/pull/14841)) [@etseidl](https://github.com/etseidl) +- Fix calls to deprecated strings factory API in examples. ([#14838](https://github.com/NVIDIA/cudf/pull/14838)) [@bdice](https://github.com/bdice) +- Update pre-commit hooks ([#14837](https://github.com/NVIDIA/cudf/pull/14837)) [@bdice](https://github.com/bdice) +- Use `rapids_cuda_set_runtime` to determine cuda runtime usage by target ([#14833](https://github.com/NVIDIA/cudf/pull/14833)) [@vyasr](https://github.com/vyasr) +- Remove get_mem_info functions from custom memory resources ([#14832](https://github.com/NVIDIA/cudf/pull/14832)) [@harrism](https://github.com/harrism) +- Fix debug build by splitting row_operator_tests_utilities.cu ([#14826](https://github.com/NVIDIA/cudf/pull/14826)) [@davidwendt](https://github.com/davidwendt) +- Remove -DNVBench_ENABLE_CUPTI=OFF. ([#14820](https://github.com/NVIDIA/cudf/pull/14820)) [@bdice](https://github.com/bdice) +- Use cuco::static_set in the hash-based groupby ([#14813](https://github.com/NVIDIA/cudf/pull/14813)) [@PointKernel](https://github.com/PointKernel) +- Branch 24.04 merge branch 24.02 ([#14809](https://github.com/NVIDIA/cudf/pull/14809)) [@vyasr](https://github.com/vyasr) +- Branch 24.04 merge branch 24.02 ([#14806](https://github.com/NVIDIA/cudf/pull/14806)) [@vyasr](https://github.com/vyasr) +- Introduce basic "cudf" backend for Dask Expressions ([#14805](https://github.com/NVIDIA/cudf/pull/14805)) [@rjzamora](https://github.com/rjzamora) +- Remove `build_struct|list_column` ([#14786](https://github.com/NVIDIA/cudf/pull/14786)) [@mroeschke](https://github.com/mroeschke) +- Use offsetalator in nvtext tokenize functions ([#14783](https://github.com/NVIDIA/cudf/pull/14783)) [@davidwendt](https://github.com/davidwendt) +- Reduce execution time of Python ORC tests ([#14776](https://github.com/NVIDIA/cudf/pull/14776)) [@vuule](https://github.com/vuule) +- Use offsetalator in cudf::strings::split functions ([#14757](https://github.com/NVIDIA/cudf/pull/14757)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::strings::findall ([#14745](https://github.com/NVIDIA/cudf/pull/14745)) [@davidwendt](https://github.com/davidwendt) +- Use offsetalator in cudf::strings::url_decode ([#14744](https://github.com/NVIDIA/cudf/pull/14744)) [@davidwendt](https://github.com/davidwendt) +- Use get_offset_value utility in strings shift function ([#14743](https://github.com/NVIDIA/cudf/pull/14743)) [@davidwendt](https://github.com/davidwendt) +- Use as_column instead of full ([#14698](https://github.com/NVIDIA/cudf/pull/14698)) [@mroeschke](https://github.com/mroeschke) +- List all notable breaking changes ([#13535](https://github.com/NVIDIA/cudf/pull/13535)) [@galipremsagar](https://github.com/galipremsagar) # cuDF 24.02.00 (12 Feb 2024) ## 🚨 Breaking Changes -- Remove **kwargs from astype ([#14765](https://github.com/rapidsai/cudf/pull/14765)) [@mroeschke](https://github.com/mroeschke) -- Remove mimesis as a testing dependency ([#14723](https://github.com/rapidsai/cudf/pull/14723)) [@mroeschke](https://github.com/mroeschke) -- Update to Dask's `shuffle_method` kwarg ([#14708](https://github.com/rapidsai/cudf/pull/14708)) [@pentschev](https://github.com/pentschev) -- Drop Pascal GPU support. ([#14630](https://github.com/rapidsai/cudf/pull/14630)) [@bdice](https://github.com/bdice) -- Update to CCCL 2.2.0. ([#14576](https://github.com/rapidsai/cudf/pull/14576)) [@bdice](https://github.com/bdice) -- Expunge as_frame conversions in Column algorithms ([#14491](https://github.com/rapidsai/cudf/pull/14491)) [@wence-](https://github.com/wence-) -- Deprecate cudf::make_strings_column accepting typed offsets ([#14461](https://github.com/rapidsai/cudf/pull/14461)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated nvtext::load_merge_pairs_file ([#14460](https://github.com/rapidsai/cudf/pull/14460)) [@davidwendt](https://github.com/davidwendt) -- Include writer code and writerVersion in ORC files ([#14458](https://github.com/rapidsai/cudf/pull/14458)) [@vuule](https://github.com/vuule) -- Remove null mask for zero nulls in json readers ([#14451](https://github.com/rapidsai/cudf/pull/14451)) [@karthikeyann](https://github.com/karthikeyann) -- REF: Remove **kwargs from to_pandas, raise if nullable is not implemented ([#14438](https://github.com/rapidsai/cudf/pull/14438)) [@mroeschke](https://github.com/mroeschke) -- Consolidate 1D pandas object handling in as_column ([#14394](https://github.com/rapidsai/cudf/pull/14394)) [@mroeschke](https://github.com/mroeschke) -- Move chars column to parent data buffer in strings column ([#14202](https://github.com/rapidsai/cudf/pull/14202)) [@karthikeyann](https://github.com/karthikeyann) -- Switch to scikit-build-core ([#13531](https://github.com/rapidsai/cudf/pull/13531)) [@vyasr](https://github.com/vyasr) +- Remove **kwargs from astype ([#14765](https://github.com/NVIDIA/cudf/pull/14765)) [@mroeschke](https://github.com/mroeschke) +- Remove mimesis as a testing dependency ([#14723](https://github.com/NVIDIA/cudf/pull/14723)) [@mroeschke](https://github.com/mroeschke) +- Update to Dask's `shuffle_method` kwarg ([#14708](https://github.com/NVIDIA/cudf/pull/14708)) [@pentschev](https://github.com/pentschev) +- Drop Pascal GPU support. ([#14630](https://github.com/NVIDIA/cudf/pull/14630)) [@bdice](https://github.com/bdice) +- Update to CCCL 2.2.0. ([#14576](https://github.com/NVIDIA/cudf/pull/14576)) [@bdice](https://github.com/bdice) +- Expunge as_frame conversions in Column algorithms ([#14491](https://github.com/NVIDIA/cudf/pull/14491)) [@wence-](https://github.com/wence-) +- Deprecate cudf::make_strings_column accepting typed offsets ([#14461](https://github.com/NVIDIA/cudf/pull/14461)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated nvtext::load_merge_pairs_file ([#14460](https://github.com/NVIDIA/cudf/pull/14460)) [@davidwendt](https://github.com/davidwendt) +- Include writer code and writerVersion in ORC files ([#14458](https://github.com/NVIDIA/cudf/pull/14458)) [@vuule](https://github.com/vuule) +- Remove null mask for zero nulls in json readers ([#14451](https://github.com/NVIDIA/cudf/pull/14451)) [@karthikeyann](https://github.com/karthikeyann) +- REF: Remove **kwargs from to_pandas, raise if nullable is not implemented ([#14438](https://github.com/NVIDIA/cudf/pull/14438)) [@mroeschke](https://github.com/mroeschke) +- Consolidate 1D pandas object handling in as_column ([#14394](https://github.com/NVIDIA/cudf/pull/14394)) [@mroeschke](https://github.com/mroeschke) +- Move chars column to parent data buffer in strings column ([#14202](https://github.com/NVIDIA/cudf/pull/14202)) [@karthikeyann](https://github.com/karthikeyann) +- Switch to scikit-build-core ([#13531](https://github.com/NVIDIA/cudf/pull/13531)) [@vyasr](https://github.com/vyasr) ## 🐛 Bug Fixes -- Exclude tests from builds ([#14981](https://github.com/rapidsai/cudf/pull/14981)) [@vyasr](https://github.com/vyasr) -- Fix the bounce buffer size in ORC writer ([#14947](https://github.com/rapidsai/cudf/pull/14947)) [@vuule](https://github.com/vuule) -- Revert sum/product aggregation to always produce `int64_t` type ([#14907](https://github.com/rapidsai/cudf/pull/14907)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Fixed an issue with output chunking computation stemming from input chunking. ([#14889](https://github.com/rapidsai/cudf/pull/14889)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix total_byte_size in Parquet row group metadata ([#14802](https://github.com/rapidsai/cudf/pull/14802)) [@etseidl](https://github.com/etseidl) -- Fix index difference to follow the pandas format ([#14789](https://github.com/rapidsai/cudf/pull/14789)) [@amiralimi](https://github.com/amiralimi) -- Fix shared-workflows repo name ([#14784](https://github.com/rapidsai/cudf/pull/14784)) [@raydouglass](https://github.com/raydouglass) -- Remove unparseable attributes from all nodes ([#14780](https://github.com/rapidsai/cudf/pull/14780)) [@vyasr](https://github.com/vyasr) -- Refactor and add validation to IntervalIndex.__init__ ([#14778](https://github.com/rapidsai/cudf/pull/14778)) [@mroeschke](https://github.com/mroeschke) -- Work around incompatibilities between V2 page header handling and zStandard compression in Parquet writer ([#14772](https://github.com/rapidsai/cudf/pull/14772)) [@etseidl](https://github.com/etseidl) -- Fix calls to deprecated strings factory API ([#14771](https://github.com/rapidsai/cudf/pull/14771)) [@davidwendt](https://github.com/davidwendt) -- Fix ptx file discovery in editable installs ([#14767](https://github.com/rapidsai/cudf/pull/14767)) [@vyasr](https://github.com/vyasr) -- Revise ``shuffle`` deprecation to align with dask/dask ([#14762](https://github.com/rapidsai/cudf/pull/14762)) [@rjzamora](https://github.com/rjzamora) -- Enable intermediate proxies to be picklable ([#14752](https://github.com/rapidsai/cudf/pull/14752)) [@shwina](https://github.com/shwina) -- Add CUDF_TEST_PROGRAM_MAIN macro to tests lacking it ([#14751](https://github.com/rapidsai/cudf/pull/14751)) [@etseidl](https://github.com/etseidl) -- Fix CMake args ([#14746](https://github.com/rapidsai/cudf/pull/14746)) [@vyasr](https://github.com/vyasr) -- Fix logic bug introduced in #14730 ([#14742](https://github.com/rapidsai/cudf/pull/14742)) [@wence-](https://github.com/wence-) -- [Java] Choose The Correct RoundingMode For Checking Decimal OutOfBounds ([#14731](https://github.com/rapidsai/cudf/pull/14731)) [@razajafri](https://github.com/razajafri) -- Fix ``Groupby.get_group`` ([#14728](https://github.com/rapidsai/cudf/pull/14728)) [@rjzamora](https://github.com/rjzamora) -- Ensure that all CUDA kernels in cudf have hidden visibility. ([#14726](https://github.com/rapidsai/cudf/pull/14726)) [@robertmaynard](https://github.com/robertmaynard) -- Split cuda versions for notebook testing ([#14722](https://github.com/rapidsai/cudf/pull/14722)) [@raydouglass](https://github.com/raydouglass) -- Fix to_numeric not preserving Series index and name ([#14718](https://github.com/rapidsai/cudf/pull/14718)) [@mroeschke](https://github.com/mroeschke) -- Update dask-cudf wheel name ([#14713](https://github.com/rapidsai/cudf/pull/14713)) [@raydouglass](https://github.com/raydouglass) -- Fix strings::contains matching end of string target ([#14711](https://github.com/rapidsai/cudf/pull/14711)) [@davidwendt](https://github.com/davidwendt) -- Update to Dask's `shuffle_method` kwarg ([#14708](https://github.com/rapidsai/cudf/pull/14708)) [@pentschev](https://github.com/pentschev) -- Write file-level statistics when writing ORC files with zero rows ([#14707](https://github.com/rapidsai/cudf/pull/14707)) [@vuule](https://github.com/vuule) -- Potential fix for peformance regression in #14415 ([#14706](https://github.com/rapidsai/cudf/pull/14706)) [@etseidl](https://github.com/etseidl) -- Ensure DataFrame column types are preserved during serialization ([#14705](https://github.com/rapidsai/cudf/pull/14705)) [@mroeschke](https://github.com/mroeschke) -- Skip numba test that fails on ARM ([#14702](https://github.com/rapidsai/cudf/pull/14702)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Allow Z in datetime string parsing in non pandas compat mode ([#14701](https://github.com/rapidsai/cudf/pull/14701)) [@mroeschke](https://github.com/mroeschke) -- Fix nan_as_null not being respected when passing arrow object ([#14688](https://github.com/rapidsai/cudf/pull/14688)) [@mroeschke](https://github.com/mroeschke) -- Fix constructing Series/Index from arrow array and dtype ([#14686](https://github.com/rapidsai/cudf/pull/14686)) [@mroeschke](https://github.com/mroeschke) -- Fix Aggregation Type Promotion: Ensure Unsigned Input Types Result in Unsigned Output for Sum and Multiply ([#14679](https://github.com/rapidsai/cudf/pull/14679)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Add BaseOffset as a final proxy type to pass instancechecks for offsets against `BaseOffset` ([#14678](https://github.com/rapidsai/cudf/pull/14678)) [@shwina](https://github.com/shwina) -- Add row conversion code from spark-rapids-jni ([#14664](https://github.com/rapidsai/cudf/pull/14664)) [@ttnghia](https://github.com/ttnghia) -- Unconditionally export the CCCL path ([#14656](https://github.com/rapidsai/cudf/pull/14656)) [@vyasr](https://github.com/vyasr) -- Ensure libcudf searches for our patched version of CCCL first ([#14655](https://github.com/rapidsai/cudf/pull/14655)) [@robertmaynard](https://github.com/robertmaynard) -- Constrain CUDA in notebook testing to prevent CUDA 12.1 usage until we have pynvjitlink ([#14648](https://github.com/rapidsai/cudf/pull/14648)) [@vyasr](https://github.com/vyasr) -- Fix invalid memory access in Parquet reader ([#14637](https://github.com/rapidsai/cudf/pull/14637)) [@etseidl](https://github.com/etseidl) -- Use column_empty over as_column([]) ([#14632](https://github.com/rapidsai/cudf/pull/14632)) [@mroeschke](https://github.com/mroeschke) -- Add (implicit) handling for torch tensors in is_scalar ([#14623](https://github.com/rapidsai/cudf/pull/14623)) [@wence-](https://github.com/wence-) -- Fix astype/fillna not maintaining column subclass and types ([#14615](https://github.com/rapidsai/cudf/pull/14615)) [@mroeschke](https://github.com/mroeschke) -- Remove non-empty nulls in cudf::get_json_object ([#14609](https://github.com/rapidsai/cudf/pull/14609)) [@davidwendt](https://github.com/davidwendt) -- Remove `cuda::proclaim_return_type` from nested lambda ([#14607](https://github.com/rapidsai/cudf/pull/14607)) [@ttnghia](https://github.com/ttnghia) -- Fix DataFrame.reindex when column reindexing to MultiIndex/RangeIndex ([#14605](https://github.com/rapidsai/cudf/pull/14605)) [@mroeschke](https://github.com/mroeschke) -- Address potential race conditions in Parquet reader ([#14602](https://github.com/rapidsai/cudf/pull/14602)) [@etseidl](https://github.com/etseidl) -- Fix DataFrame.reindex removing column name ([#14601](https://github.com/rapidsai/cudf/pull/14601)) [@mroeschke](https://github.com/mroeschke) -- Remove unsanitized input test data from copy gtests ([#14600](https://github.com/rapidsai/cudf/pull/14600)) [@davidwendt](https://github.com/davidwendt) -- Fix race detected in Parquet writer ([#14598](https://github.com/rapidsai/cudf/pull/14598)) [@etseidl](https://github.com/etseidl) -- Correct invalid or missing return types ([#14587](https://github.com/rapidsai/cudf/pull/14587)) [@robertmaynard](https://github.com/robertmaynard) -- Fix unsanitized nulls from strings segmented-reduce ([#14586](https://github.com/rapidsai/cudf/pull/14586)) [@davidwendt](https://github.com/davidwendt) -- Upgrade to nvCOMP 3.0.5 ([#14581](https://github.com/rapidsai/cudf/pull/14581)) [@davidwendt](https://github.com/davidwendt) -- Fix unsanitized nulls produced by `cudf::clamp` APIs ([#14580](https://github.com/rapidsai/cudf/pull/14580)) [@davidwendt](https://github.com/davidwendt) -- Fix unsanitized nulls produced by libcudf dictionary decode ([#14578](https://github.com/rapidsai/cudf/pull/14578)) [@davidwendt](https://github.com/davidwendt) -- Fixes a symbol group lookup table issue ([#14561](https://github.com/rapidsai/cudf/pull/14561)) [@elstehle](https://github.com/elstehle) -- Drop llvm16 from cuda118-conda devcontainer image ([#14526](https://github.com/rapidsai/cudf/pull/14526)) [@charlesbluca](https://github.com/charlesbluca) -- REF: Make DataFrame.from_pandas process by column ([#14483](https://github.com/rapidsai/cudf/pull/14483)) [@mroeschke](https://github.com/mroeschke) -- Improve memory footprint of isin by using contains ([#14478](https://github.com/rapidsai/cudf/pull/14478)) [@wence-](https://github.com/wence-) -- Move creation of env.yaml outside the current directory ([#14476](https://github.com/rapidsai/cudf/pull/14476)) [@davidwendt](https://github.com/davidwendt) -- Enable `pd.Timestamp` objects to be picklable when `cudf.pandas` is active ([#14474](https://github.com/rapidsai/cudf/pull/14474)) [@shwina](https://github.com/shwina) -- Correct dtype of count aggregations on empty dataframes ([#14473](https://github.com/rapidsai/cudf/pull/14473)) [@wence-](https://github.com/wence-) -- Avoid DataFrame conversion in `MultiIndex.from_pandas` ([#14470](https://github.com/rapidsai/cudf/pull/14470)) [@mroeschke](https://github.com/mroeschke) -- JSON writer: avoid default stream use in `string_scalar` constructors ([#14444](https://github.com/rapidsai/cudf/pull/14444)) [@vuule](https://github.com/vuule) -- Fix default stream use in the CSV reader ([#14443](https://github.com/rapidsai/cudf/pull/14443)) [@vuule](https://github.com/vuule) -- Preserve DataFrame(columns=).columns dtype during empty-like construction ([#14381](https://github.com/rapidsai/cudf/pull/14381)) [@mroeschke](https://github.com/mroeschke) -- Defer PTX file load to runtime ([#13690](https://github.com/rapidsai/cudf/pull/13690)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Exclude tests from builds ([#14981](https://github.com/NVIDIA/cudf/pull/14981)) [@vyasr](https://github.com/vyasr) +- Fix the bounce buffer size in ORC writer ([#14947](https://github.com/NVIDIA/cudf/pull/14947)) [@vuule](https://github.com/vuule) +- Revert sum/product aggregation to always produce `int64_t` type ([#14907](https://github.com/NVIDIA/cudf/pull/14907)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Fixed an issue with output chunking computation stemming from input chunking. ([#14889](https://github.com/NVIDIA/cudf/pull/14889)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix total_byte_size in Parquet row group metadata ([#14802](https://github.com/NVIDIA/cudf/pull/14802)) [@etseidl](https://github.com/etseidl) +- Fix index difference to follow the pandas format ([#14789](https://github.com/NVIDIA/cudf/pull/14789)) [@amiralimi](https://github.com/amiralimi) +- Fix shared-workflows repo name ([#14784](https://github.com/NVIDIA/cudf/pull/14784)) [@raydouglass](https://github.com/raydouglass) +- Remove unparseable attributes from all nodes ([#14780](https://github.com/NVIDIA/cudf/pull/14780)) [@vyasr](https://github.com/vyasr) +- Refactor and add validation to IntervalIndex.__init__ ([#14778](https://github.com/NVIDIA/cudf/pull/14778)) [@mroeschke](https://github.com/mroeschke) +- Work around incompatibilities between V2 page header handling and zStandard compression in Parquet writer ([#14772](https://github.com/NVIDIA/cudf/pull/14772)) [@etseidl](https://github.com/etseidl) +- Fix calls to deprecated strings factory API ([#14771](https://github.com/NVIDIA/cudf/pull/14771)) [@davidwendt](https://github.com/davidwendt) +- Fix ptx file discovery in editable installs ([#14767](https://github.com/NVIDIA/cudf/pull/14767)) [@vyasr](https://github.com/vyasr) +- Revise ``shuffle`` deprecation to align with dask/dask ([#14762](https://github.com/NVIDIA/cudf/pull/14762)) [@rjzamora](https://github.com/rjzamora) +- Enable intermediate proxies to be picklable ([#14752](https://github.com/NVIDIA/cudf/pull/14752)) [@shwina](https://github.com/shwina) +- Add CUDF_TEST_PROGRAM_MAIN macro to tests lacking it ([#14751](https://github.com/NVIDIA/cudf/pull/14751)) [@etseidl](https://github.com/etseidl) +- Fix CMake args ([#14746](https://github.com/NVIDIA/cudf/pull/14746)) [@vyasr](https://github.com/vyasr) +- Fix logic bug introduced in #14730 ([#14742](https://github.com/NVIDIA/cudf/pull/14742)) [@wence-](https://github.com/wence-) +- [Java] Choose The Correct RoundingMode For Checking Decimal OutOfBounds ([#14731](https://github.com/NVIDIA/cudf/pull/14731)) [@razajafri](https://github.com/razajafri) +- Fix ``Groupby.get_group`` ([#14728](https://github.com/NVIDIA/cudf/pull/14728)) [@rjzamora](https://github.com/rjzamora) +- Ensure that all CUDA kernels in cudf have hidden visibility. ([#14726](https://github.com/NVIDIA/cudf/pull/14726)) [@robertmaynard](https://github.com/robertmaynard) +- Split cuda versions for notebook testing ([#14722](https://github.com/NVIDIA/cudf/pull/14722)) [@raydouglass](https://github.com/raydouglass) +- Fix to_numeric not preserving Series index and name ([#14718](https://github.com/NVIDIA/cudf/pull/14718)) [@mroeschke](https://github.com/mroeschke) +- Update dask-cudf wheel name ([#14713](https://github.com/NVIDIA/cudf/pull/14713)) [@raydouglass](https://github.com/raydouglass) +- Fix strings::contains matching end of string target ([#14711](https://github.com/NVIDIA/cudf/pull/14711)) [@davidwendt](https://github.com/davidwendt) +- Update to Dask's `shuffle_method` kwarg ([#14708](https://github.com/NVIDIA/cudf/pull/14708)) [@pentschev](https://github.com/pentschev) +- Write file-level statistics when writing ORC files with zero rows ([#14707](https://github.com/NVIDIA/cudf/pull/14707)) [@vuule](https://github.com/vuule) +- Potential fix for peformance regression in #14415 ([#14706](https://github.com/NVIDIA/cudf/pull/14706)) [@etseidl](https://github.com/etseidl) +- Ensure DataFrame column types are preserved during serialization ([#14705](https://github.com/NVIDIA/cudf/pull/14705)) [@mroeschke](https://github.com/mroeschke) +- Skip numba test that fails on ARM ([#14702](https://github.com/NVIDIA/cudf/pull/14702)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Allow Z in datetime string parsing in non pandas compat mode ([#14701](https://github.com/NVIDIA/cudf/pull/14701)) [@mroeschke](https://github.com/mroeschke) +- Fix nan_as_null not being respected when passing arrow object ([#14688](https://github.com/NVIDIA/cudf/pull/14688)) [@mroeschke](https://github.com/mroeschke) +- Fix constructing Series/Index from arrow array and dtype ([#14686](https://github.com/NVIDIA/cudf/pull/14686)) [@mroeschke](https://github.com/mroeschke) +- Fix Aggregation Type Promotion: Ensure Unsigned Input Types Result in Unsigned Output for Sum and Multiply ([#14679](https://github.com/NVIDIA/cudf/pull/14679)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Add BaseOffset as a final proxy type to pass instancechecks for offsets against `BaseOffset` ([#14678](https://github.com/NVIDIA/cudf/pull/14678)) [@shwina](https://github.com/shwina) +- Add row conversion code from spark-rapids-jni ([#14664](https://github.com/NVIDIA/cudf/pull/14664)) [@ttnghia](https://github.com/ttnghia) +- Unconditionally export the CCCL path ([#14656](https://github.com/NVIDIA/cudf/pull/14656)) [@vyasr](https://github.com/vyasr) +- Ensure libcudf searches for our patched version of CCCL first ([#14655](https://github.com/NVIDIA/cudf/pull/14655)) [@robertmaynard](https://github.com/robertmaynard) +- Constrain CUDA in notebook testing to prevent CUDA 12.1 usage until we have pynvjitlink ([#14648](https://github.com/NVIDIA/cudf/pull/14648)) [@vyasr](https://github.com/vyasr) +- Fix invalid memory access in Parquet reader ([#14637](https://github.com/NVIDIA/cudf/pull/14637)) [@etseidl](https://github.com/etseidl) +- Use column_empty over as_column([]) ([#14632](https://github.com/NVIDIA/cudf/pull/14632)) [@mroeschke](https://github.com/mroeschke) +- Add (implicit) handling for torch tensors in is_scalar ([#14623](https://github.com/NVIDIA/cudf/pull/14623)) [@wence-](https://github.com/wence-) +- Fix astype/fillna not maintaining column subclass and types ([#14615](https://github.com/NVIDIA/cudf/pull/14615)) [@mroeschke](https://github.com/mroeschke) +- Remove non-empty nulls in cudf::get_json_object ([#14609](https://github.com/NVIDIA/cudf/pull/14609)) [@davidwendt](https://github.com/davidwendt) +- Remove `cuda::proclaim_return_type` from nested lambda ([#14607](https://github.com/NVIDIA/cudf/pull/14607)) [@ttnghia](https://github.com/ttnghia) +- Fix DataFrame.reindex when column reindexing to MultiIndex/RangeIndex ([#14605](https://github.com/NVIDIA/cudf/pull/14605)) [@mroeschke](https://github.com/mroeschke) +- Address potential race conditions in Parquet reader ([#14602](https://github.com/NVIDIA/cudf/pull/14602)) [@etseidl](https://github.com/etseidl) +- Fix DataFrame.reindex removing column name ([#14601](https://github.com/NVIDIA/cudf/pull/14601)) [@mroeschke](https://github.com/mroeschke) +- Remove unsanitized input test data from copy gtests ([#14600](https://github.com/NVIDIA/cudf/pull/14600)) [@davidwendt](https://github.com/davidwendt) +- Fix race detected in Parquet writer ([#14598](https://github.com/NVIDIA/cudf/pull/14598)) [@etseidl](https://github.com/etseidl) +- Correct invalid or missing return types ([#14587](https://github.com/NVIDIA/cudf/pull/14587)) [@robertmaynard](https://github.com/robertmaynard) +- Fix unsanitized nulls from strings segmented-reduce ([#14586](https://github.com/NVIDIA/cudf/pull/14586)) [@davidwendt](https://github.com/davidwendt) +- Upgrade to nvCOMP 3.0.5 ([#14581](https://github.com/NVIDIA/cudf/pull/14581)) [@davidwendt](https://github.com/davidwendt) +- Fix unsanitized nulls produced by `cudf::clamp` APIs ([#14580](https://github.com/NVIDIA/cudf/pull/14580)) [@davidwendt](https://github.com/davidwendt) +- Fix unsanitized nulls produced by libcudf dictionary decode ([#14578](https://github.com/NVIDIA/cudf/pull/14578)) [@davidwendt](https://github.com/davidwendt) +- Fixes a symbol group lookup table issue ([#14561](https://github.com/NVIDIA/cudf/pull/14561)) [@elstehle](https://github.com/elstehle) +- Drop llvm16 from cuda118-conda devcontainer image ([#14526](https://github.com/NVIDIA/cudf/pull/14526)) [@charlesbluca](https://github.com/charlesbluca) +- REF: Make DataFrame.from_pandas process by column ([#14483](https://github.com/NVIDIA/cudf/pull/14483)) [@mroeschke](https://github.com/mroeschke) +- Improve memory footprint of isin by using contains ([#14478](https://github.com/NVIDIA/cudf/pull/14478)) [@wence-](https://github.com/wence-) +- Move creation of env.yaml outside the current directory ([#14476](https://github.com/NVIDIA/cudf/pull/14476)) [@davidwendt](https://github.com/davidwendt) +- Enable `pd.Timestamp` objects to be picklable when `cudf.pandas` is active ([#14474](https://github.com/NVIDIA/cudf/pull/14474)) [@shwina](https://github.com/shwina) +- Correct dtype of count aggregations on empty dataframes ([#14473](https://github.com/NVIDIA/cudf/pull/14473)) [@wence-](https://github.com/wence-) +- Avoid DataFrame conversion in `MultiIndex.from_pandas` ([#14470](https://github.com/NVIDIA/cudf/pull/14470)) [@mroeschke](https://github.com/mroeschke) +- JSON writer: avoid default stream use in `string_scalar` constructors ([#14444](https://github.com/NVIDIA/cudf/pull/14444)) [@vuule](https://github.com/vuule) +- Fix default stream use in the CSV reader ([#14443](https://github.com/NVIDIA/cudf/pull/14443)) [@vuule](https://github.com/vuule) +- Preserve DataFrame(columns=).columns dtype during empty-like construction ([#14381](https://github.com/NVIDIA/cudf/pull/14381)) [@mroeschke](https://github.com/mroeschke) +- Defer PTX file load to runtime ([#13690](https://github.com/NVIDIA/cudf/pull/13690)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 📖 Documentation -- Disable parallel build ([#14796](https://github.com/rapidsai/cudf/pull/14796)) [@vyasr](https://github.com/vyasr) -- Add pylibcudf to the docs ([#14791](https://github.com/rapidsai/cudf/pull/14791)) [@vyasr](https://github.com/vyasr) -- Describe unpickling expectations when cudf.pandas is enabled ([#14693](https://github.com/rapidsai/cudf/pull/14693)) [@shwina](https://github.com/shwina) -- Update CONTRIBUTING for pyproject-only builds ([#14653](https://github.com/rapidsai/cudf/pull/14653)) [@vyasr](https://github.com/vyasr) -- More doxygen fixes ([#14639](https://github.com/rapidsai/cudf/pull/14639)) [@vyasr](https://github.com/vyasr) -- Enable doxygen XML generation and fix issues ([#14477](https://github.com/rapidsai/cudf/pull/14477)) [@vyasr](https://github.com/vyasr) -- Some doxygen improvements ([#14469](https://github.com/rapidsai/cudf/pull/14469)) [@vyasr](https://github.com/vyasr) -- Remove warning in dask-cudf docs ([#14454](https://github.com/rapidsai/cudf/pull/14454)) [@wence-](https://github.com/wence-) -- Update README links with redirects. ([#14378](https://github.com/rapidsai/cudf/pull/14378)) [@bdice](https://github.com/bdice) -- Add pip install instructions to README ([#13677](https://github.com/rapidsai/cudf/pull/13677)) [@shwina](https://github.com/shwina) +- Disable parallel build ([#14796](https://github.com/NVIDIA/cudf/pull/14796)) [@vyasr](https://github.com/vyasr) +- Add pylibcudf to the docs ([#14791](https://github.com/NVIDIA/cudf/pull/14791)) [@vyasr](https://github.com/vyasr) +- Describe unpickling expectations when cudf.pandas is enabled ([#14693](https://github.com/NVIDIA/cudf/pull/14693)) [@shwina](https://github.com/shwina) +- Update CONTRIBUTING for pyproject-only builds ([#14653](https://github.com/NVIDIA/cudf/pull/14653)) [@vyasr](https://github.com/vyasr) +- More doxygen fixes ([#14639](https://github.com/NVIDIA/cudf/pull/14639)) [@vyasr](https://github.com/vyasr) +- Enable doxygen XML generation and fix issues ([#14477](https://github.com/NVIDIA/cudf/pull/14477)) [@vyasr](https://github.com/vyasr) +- Some doxygen improvements ([#14469](https://github.com/NVIDIA/cudf/pull/14469)) [@vyasr](https://github.com/vyasr) +- Remove warning in dask-cudf docs ([#14454](https://github.com/NVIDIA/cudf/pull/14454)) [@wence-](https://github.com/wence-) +- Update README links with redirects. ([#14378](https://github.com/NVIDIA/cudf/pull/14378)) [@bdice](https://github.com/bdice) +- Add pip install instructions to README ([#13677](https://github.com/NVIDIA/cudf/pull/13677)) [@shwina](https://github.com/shwina) ## 🚀 New Features -- Add ci check for external kernels ([#14768](https://github.com/rapidsai/cudf/pull/14768)) [@robertmaynard](https://github.com/robertmaynard) -- JSON single quote normalization API ([#14729](https://github.com/rapidsai/cudf/pull/14729)) [@shrshi](https://github.com/shrshi) -- Write cuDF version in Parquet "created_by" metadata field ([#14721](https://github.com/rapidsai/cudf/pull/14721)) [@etseidl](https://github.com/etseidl) -- Implement remaining copying APIs in pylibcudf along with required helper functions ([#14640](https://github.com/rapidsai/cudf/pull/14640)) [@vyasr](https://github.com/vyasr) -- Don't constrain `numba<0.58` ([#14616](https://github.com/rapidsai/cudf/pull/14616)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add DELTA_LENGTH_BYTE_ARRAY encoder and decoder for Parquet ([#14590](https://github.com/rapidsai/cudf/pull/14590)) [@etseidl](https://github.com/etseidl) -- JSON - Parse mixed types as string in JSON reader ([#14572](https://github.com/rapidsai/cudf/pull/14572)) [@karthikeyann](https://github.com/karthikeyann) -- JSON quote normalization ([#14545](https://github.com/rapidsai/cudf/pull/14545)) [@shrshi](https://github.com/shrshi) -- Make DefaultHostMemoryAllocator settable ([#14523](https://github.com/rapidsai/cudf/pull/14523)) [@gerashegalov](https://github.com/gerashegalov) -- Implement more copying APIs in pylibcudf ([#14508](https://github.com/rapidsai/cudf/pull/14508)) [@vyasr](https://github.com/vyasr) -- Include writer code and writerVersion in ORC files ([#14458](https://github.com/rapidsai/cudf/pull/14458)) [@vuule](https://github.com/vuule) -- Parquet sub-rowgroup reading. ([#14360](https://github.com/rapidsai/cudf/pull/14360)) [@nvdbaranec](https://github.com/nvdbaranec) -- Move chars column to parent data buffer in strings column ([#14202](https://github.com/rapidsai/cudf/pull/14202)) [@karthikeyann](https://github.com/karthikeyann) -- PARQUET-2261 Size Statistics ([#14000](https://github.com/rapidsai/cudf/pull/14000)) [@etseidl](https://github.com/etseidl) -- Improve GroupBy JIT error handling ([#13854](https://github.com/rapidsai/cudf/pull/13854)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Generate unified Python/C++ docs ([#13846](https://github.com/rapidsai/cudf/pull/13846)) [@vyasr](https://github.com/vyasr) -- Expand JIT groupby test suite ([#13813](https://github.com/rapidsai/cudf/pull/13813)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add ci check for external kernels ([#14768](https://github.com/NVIDIA/cudf/pull/14768)) [@robertmaynard](https://github.com/robertmaynard) +- JSON single quote normalization API ([#14729](https://github.com/NVIDIA/cudf/pull/14729)) [@shrshi](https://github.com/shrshi) +- Write cuDF version in Parquet "created_by" metadata field ([#14721](https://github.com/NVIDIA/cudf/pull/14721)) [@etseidl](https://github.com/etseidl) +- Implement remaining copying APIs in pylibcudf along with required helper functions ([#14640](https://github.com/NVIDIA/cudf/pull/14640)) [@vyasr](https://github.com/vyasr) +- Don't constrain `numba<0.58` ([#14616](https://github.com/NVIDIA/cudf/pull/14616)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add DELTA_LENGTH_BYTE_ARRAY encoder and decoder for Parquet ([#14590](https://github.com/NVIDIA/cudf/pull/14590)) [@etseidl](https://github.com/etseidl) +- JSON - Parse mixed types as string in JSON reader ([#14572](https://github.com/NVIDIA/cudf/pull/14572)) [@karthikeyann](https://github.com/karthikeyann) +- JSON quote normalization ([#14545](https://github.com/NVIDIA/cudf/pull/14545)) [@shrshi](https://github.com/shrshi) +- Make DefaultHostMemoryAllocator settable ([#14523](https://github.com/NVIDIA/cudf/pull/14523)) [@gerashegalov](https://github.com/gerashegalov) +- Implement more copying APIs in pylibcudf ([#14508](https://github.com/NVIDIA/cudf/pull/14508)) [@vyasr](https://github.com/vyasr) +- Include writer code and writerVersion in ORC files ([#14458](https://github.com/NVIDIA/cudf/pull/14458)) [@vuule](https://github.com/vuule) +- Parquet sub-rowgroup reading. ([#14360](https://github.com/NVIDIA/cudf/pull/14360)) [@nvdbaranec](https://github.com/nvdbaranec) +- Move chars column to parent data buffer in strings column ([#14202](https://github.com/NVIDIA/cudf/pull/14202)) [@karthikeyann](https://github.com/karthikeyann) +- PARQUET-2261 Size Statistics ([#14000](https://github.com/NVIDIA/cudf/pull/14000)) [@etseidl](https://github.com/etseidl) +- Improve GroupBy JIT error handling ([#13854](https://github.com/NVIDIA/cudf/pull/13854)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Generate unified Python/C++ docs ([#13846](https://github.com/NVIDIA/cudf/pull/13846)) [@vyasr](https://github.com/vyasr) +- Expand JIT groupby test suite ([#13813](https://github.com/NVIDIA/cudf/pull/13813)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🛠️ Improvements -- Pin `pytest<8` ([#14920](https://github.com/rapidsai/cudf/pull/14920)) [@galipremsagar](https://github.com/galipremsagar) -- Move cudf::char_utf8 definition from detail to public header ([#14779](https://github.com/rapidsai/cudf/pull/14779)) [@davidwendt](https://github.com/davidwendt) -- Clean up `TimedeltaIndex.__init__` constructor ([#14775](https://github.com/rapidsai/cudf/pull/14775)) [@mroeschke](https://github.com/mroeschke) -- Clean up `DatetimeIndex.__init__` constructor ([#14774](https://github.com/rapidsai/cudf/pull/14774)) [@mroeschke](https://github.com/mroeschke) -- Some `frame.py` typing, move seldom used methods in `frame.py` ([#14766](https://github.com/rapidsai/cudf/pull/14766)) [@mroeschke](https://github.com/mroeschke) -- Remove **kwargs from astype ([#14765](https://github.com/rapidsai/cudf/pull/14765)) [@mroeschke](https://github.com/mroeschke) -- fix benchmarks compatibility with newer pytest-cases ([#14764](https://github.com/rapidsai/cudf/pull/14764)) [@jameslamb](https://github.com/jameslamb) -- Add `pynvjitlink` as a dependency ([#14763](https://github.com/rapidsai/cudf/pull/14763)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Resolve degenerate performance in `create_structs_data` ([#14761](https://github.com/rapidsai/cudf/pull/14761)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Simplify ColumnAccessor methods; avoid unnecessary validations ([#14758](https://github.com/rapidsai/cudf/pull/14758)) [@mroeschke](https://github.com/mroeschke) -- Pin pytest-cases<3.8.2 ([#14756](https://github.com/rapidsai/cudf/pull/14756)) [@mroeschke](https://github.com/mroeschke) -- Use _from_data instead of _from_columns for initialzing Frame ([#14755](https://github.com/rapidsai/cudf/pull/14755)) [@mroeschke](https://github.com/mroeschke) -- Consolidate cudf object handling in as_column ([#14754](https://github.com/rapidsai/cudf/pull/14754)) [@mroeschke](https://github.com/mroeschke) -- Reduce execution time of Parquet C++ tests ([#14750](https://github.com/rapidsai/cudf/pull/14750)) [@vuule](https://github.com/vuule) -- Implement to_datetime(..., utc=True) ([#14749](https://github.com/rapidsai/cudf/pull/14749)) [@mroeschke](https://github.com/mroeschke) -- Remove usages of rapids-env-update ([#14748](https://github.com/rapidsai/cudf/pull/14748)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) -- Provide explicit pool size and avoid RMM detail APIs ([#14741](https://github.com/rapidsai/cudf/pull/14741)) [@harrism](https://github.com/harrism) -- Implement `cudf.MultiIndex.from_arrays` ([#14740](https://github.com/rapidsai/cudf/pull/14740)) [@mroeschke](https://github.com/mroeschke) -- Remove unused/single use methods ([#14739](https://github.com/rapidsai/cudf/pull/14739)) [@mroeschke](https://github.com/mroeschke) -- refactor CUDA versions in dependencies.yaml ([#14733](https://github.com/rapidsai/cudf/pull/14733)) [@jameslamb](https://github.com/jameslamb) -- Remove unneeded methods in Column ([#14730](https://github.com/rapidsai/cudf/pull/14730)) [@mroeschke](https://github.com/mroeschke) -- Clean up base column methods ([#14725](https://github.com/rapidsai/cudf/pull/14725)) [@mroeschke](https://github.com/mroeschke) -- Ensure column.fillna signatures are consistent ([#14724](https://github.com/rapidsai/cudf/pull/14724)) [@mroeschke](https://github.com/mroeschke) -- Remove mimesis as a testing dependency ([#14723](https://github.com/rapidsai/cudf/pull/14723)) [@mroeschke](https://github.com/mroeschke) -- Replace as_numerical with as_numerical_column/codes ([#14719](https://github.com/rapidsai/cudf/pull/14719)) [@mroeschke](https://github.com/mroeschke) -- Use offsetalator in gather_chars ([#14700](https://github.com/rapidsai/cudf/pull/14700)) [@davidwendt](https://github.com/davidwendt) -- Use make_strings_children for fill() specialization logic ([#14697](https://github.com/rapidsai/cudf/pull/14697)) [@davidwendt](https://github.com/davidwendt) -- Change `io::detail::orc` namespace into `io::orc::detail` ([#14696](https://github.com/rapidsai/cudf/pull/14696)) [@ttnghia](https://github.com/ttnghia) -- Fix call to deprecated factory function ([#14695](https://github.com/rapidsai/cudf/pull/14695)) [@davidwendt](https://github.com/davidwendt) -- Use as_column instead of arange for range like inputs ([#14689](https://github.com/rapidsai/cudf/pull/14689)) [@mroeschke](https://github.com/mroeschke) -- Reorganize ORC reader into multiple files and perform some small fixes to cuIO code ([#14665](https://github.com/rapidsai/cudf/pull/14665)) [@ttnghia](https://github.com/ttnghia) -- Split parquet test into multiple files ([#14663](https://github.com/rapidsai/cudf/pull/14663)) [@etseidl](https://github.com/etseidl) -- Custom error messages for IO with nonexistent files ([#14662](https://github.com/rapidsai/cudf/pull/14662)) [@vuule](https://github.com/vuule) -- Explicitly pass .dtype into is_foo_dtype functions ([#14657](https://github.com/rapidsai/cudf/pull/14657)) [@mroeschke](https://github.com/mroeschke) -- Basic validation in reader benchmarks ([#14647](https://github.com/rapidsai/cudf/pull/14647)) [@vuule](https://github.com/vuule) -- Update dependencies.yaml to support CUDA 12.*. ([#14644](https://github.com/rapidsai/cudf/pull/14644)) [@bdice](https://github.com/bdice) -- Consolidate memoryview handling in as_column ([#14643](https://github.com/rapidsai/cudf/pull/14643)) [@mroeschke](https://github.com/mroeschke) -- Convert `FieldType` to scoped enum ([#14642](https://github.com/rapidsai/cudf/pull/14642)) [@vuule](https://github.com/vuule) -- Use instance over is_foo_dtype ([#14641](https://github.com/rapidsai/cudf/pull/14641)) [@mroeschke](https://github.com/mroeschke) -- Use isinstance over is_foo_dtype internally ([#14638](https://github.com/rapidsai/cudf/pull/14638)) [@mroeschke](https://github.com/mroeschke) -- Remove unnecessary **kwargs in function signatures ([#14635](https://github.com/rapidsai/cudf/pull/14635)) [@mroeschke](https://github.com/mroeschke) -- Drop nvbench patch for nvml. ([#14631](https://github.com/rapidsai/cudf/pull/14631)) [@bdice](https://github.com/bdice) -- Drop Pascal GPU support. ([#14630](https://github.com/rapidsai/cudf/pull/14630)) [@bdice](https://github.com/bdice) -- Add cpp/doxygen/xml to .gitignore ([#14613](https://github.com/rapidsai/cudf/pull/14613)) [@davidwendt](https://github.com/davidwendt) -- Create strings-specific make_offsets_child_column for multiple offset types ([#14612](https://github.com/rapidsai/cudf/pull/14612)) [@davidwendt](https://github.com/davidwendt) -- Use the offsetalator in cudf::concatenate for strings ([#14611](https://github.com/rapidsai/cudf/pull/14611)) [@davidwendt](https://github.com/davidwendt) -- Make Parquet ColumnIndex null_counts optional ([#14596](https://github.com/rapidsai/cudf/pull/14596)) [@etseidl](https://github.com/etseidl) -- Support `freq` in DatetimeIndex ([#14593](https://github.com/rapidsai/cudf/pull/14593)) [@shwina](https://github.com/shwina) -- Remove legacy benchmarks for cuDF-python ([#14591](https://github.com/rapidsai/cudf/pull/14591)) [@osidekyle](https://github.com/osidekyle) -- Remove WORKSPACE env var from cudf_test temp_directory class ([#14588](https://github.com/rapidsai/cudf/pull/14588)) [@davidwendt](https://github.com/davidwendt) -- Use exceptions instead of return values to handle errors in `CompactProtocolReader` ([#14582](https://github.com/rapidsai/cudf/pull/14582)) [@vuule](https://github.com/vuule) -- Use cuda::proclaim_return_type on device lambdas. ([#14577](https://github.com/rapidsai/cudf/pull/14577)) [@bdice](https://github.com/bdice) -- Update to CCCL 2.2.0. ([#14576](https://github.com/rapidsai/cudf/pull/14576)) [@bdice](https://github.com/bdice) -- Update dependencies.yaml to new pip index ([#14575](https://github.com/rapidsai/cudf/pull/14575)) [@vyasr](https://github.com/vyasr) -- Simplify Python CMake ([#14565](https://github.com/rapidsai/cudf/pull/14565)) [@vyasr](https://github.com/vyasr) -- Java expose parquet pass_read_limit ([#14564](https://github.com/rapidsai/cudf/pull/14564)) [@revans2](https://github.com/revans2) -- Add column sanitization checks in `CUDF_TEST_EXPECT_COLUMN_*` macros ([#14559](https://github.com/rapidsai/cudf/pull/14559)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Use cudf_test temp_directory class for nvtext::subword_tokenize gbenchmark ([#14558](https://github.com/rapidsai/cudf/pull/14558)) [@davidwendt](https://github.com/davidwendt) -- Fix return type of prefix increment overloads ([#14544](https://github.com/rapidsai/cudf/pull/14544)) [@vuule](https://github.com/vuule) -- Make bpe_merge_pairs_impl member private ([#14543](https://github.com/rapidsai/cudf/pull/14543)) [@davidwendt](https://github.com/davidwendt) -- Small clean up in `io::statistics` ([#14542](https://github.com/rapidsai/cudf/pull/14542)) [@vuule](https://github.com/vuule) -- Change json gtest environment variable to compile-time definition ([#14541](https://github.com/rapidsai/cudf/pull/14541)) [@davidwendt](https://github.com/davidwendt) -- Remove extra total chars size calculation from cudf::concatenate ([#14540](https://github.com/rapidsai/cudf/pull/14540)) [@davidwendt](https://github.com/davidwendt) -- Refactor IndexedFrame.hash_values to use cudf::hashing functions, add xxhash64 to cudf Python. ([#14538](https://github.com/rapidsai/cudf/pull/14538)) [@bdice](https://github.com/bdice) -- Move non-templated inline function definitions from table_view.hpp to table_view.cpp ([#14535](https://github.com/rapidsai/cudf/pull/14535)) [@davidwendt](https://github.com/davidwendt) -- Add JNI for strings::code_points ([#14533](https://github.com/rapidsai/cudf/pull/14533)) [@thirtiseven](https://github.com/thirtiseven) -- Add a test for issue 12773 ([#14529](https://github.com/rapidsai/cudf/pull/14529)) [@vyasr](https://github.com/vyasr) -- Split libarrow build dependencies. ([#14506](https://github.com/rapidsai/cudf/pull/14506)) [@bdice](https://github.com/bdice) -- Implement `IndexedFrame.duplicated` with `distinct_indices` + `scatter` ([#14493](https://github.com/rapidsai/cudf/pull/14493)) [@wence-](https://github.com/wence-) -- Expunge as_frame conversions in Column algorithms ([#14491](https://github.com/rapidsai/cudf/pull/14491)) [@wence-](https://github.com/wence-) -- Remove unsanitized null from input strings column in rank_tests.cpp ([#14475](https://github.com/rapidsai/cudf/pull/14475)) [@davidwendt](https://github.com/davidwendt) -- Refactor Parquet kernel_error ([#14464](https://github.com/rapidsai/cudf/pull/14464)) [@etseidl](https://github.com/etseidl) -- Deprecate cudf::make_strings_column accepting typed offsets ([#14461](https://github.com/rapidsai/cudf/pull/14461)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated nvtext::load_merge_pairs_file ([#14460](https://github.com/rapidsai/cudf/pull/14460)) [@davidwendt](https://github.com/davidwendt) -- Introduce Comprehensive Pathological Unit Tests for Issue #14409 ([#14459](https://github.com/rapidsai/cudf/pull/14459)) [@aocsa](https://github.com/aocsa) -- Expose stream parameter in public nvtext APIs ([#14456](https://github.com/rapidsai/cudf/pull/14456)) [@davidwendt](https://github.com/davidwendt) -- Include encode type in the error message when unsupported Parquet encoding is detected ([#14453](https://github.com/rapidsai/cudf/pull/14453)) [@ZelboK](https://github.com/ZelboK) -- Remove null mask for zero nulls in json readers ([#14451](https://github.com/rapidsai/cudf/pull/14451)) [@karthikeyann](https://github.com/karthikeyann) -- Refactor cudf.Series.__init__ ([#14450](https://github.com/rapidsai/cudf/pull/14450)) [@mroeschke](https://github.com/mroeschke) -- Remove the use of `volatile` in Parquet ([#14448](https://github.com/rapidsai/cudf/pull/14448)) [@vuule](https://github.com/vuule) -- REF: Remove **kwargs from to_pandas, raise if nullable is not implemented ([#14438](https://github.com/rapidsai/cudf/pull/14438)) [@mroeschke](https://github.com/mroeschke) -- Testing stream pool implementation ([#14437](https://github.com/rapidsai/cudf/pull/14437)) [@shrshi](https://github.com/shrshi) -- Match pandas join ordering obligations in pandas-compatible mode ([#14428](https://github.com/rapidsai/cudf/pull/14428)) [@wence-](https://github.com/wence-) -- Forward-merge branch-23.12 to branch-24.02 ([#14426](https://github.com/rapidsai/cudf/pull/14426)) [@bdice](https://github.com/bdice) -- Use isinstance(..., cudf.IntervalDtype) instead of is_interval_dtype ([#14424](https://github.com/rapidsai/cudf/pull/14424)) [@mroeschke](https://github.com/mroeschke) -- Use isinstance(..., cudf.CategoricalDtype) instead of is_categorical_dtype ([#14423](https://github.com/rapidsai/cudf/pull/14423)) [@mroeschke](https://github.com/mroeschke) -- Forward-merge branch-23.12 to branch-24.02 ([#14422](https://github.com/rapidsai/cudf/pull/14422)) [@bdice](https://github.com/bdice) -- REF: Remove instances of pd.core ([#14421](https://github.com/rapidsai/cudf/pull/14421)) [@mroeschke](https://github.com/mroeschke) -- Expose streams in public filling APIs for label_bins ([#14401](https://github.com/rapidsai/cudf/pull/14401)) [@ZelboK](https://github.com/ZelboK) -- Consolidate 1D pandas object handling in as_column ([#14394](https://github.com/rapidsai/cudf/pull/14394)) [@mroeschke](https://github.com/mroeschke) -- Limit DELTA_BINARY_PACKED encoder to the same number of bits as the physical type being encoded ([#14392](https://github.com/rapidsai/cudf/pull/14392)) [@etseidl](https://github.com/etseidl) -- Add SHA-1 and SHA-2 hash functions. ([#14391](https://github.com/rapidsai/cudf/pull/14391)) [@bdice](https://github.com/bdice) -- Expose streams in Parquet reader and writer APIs ([#14359](https://github.com/rapidsai/cudf/pull/14359)) [@shrshi](https://github.com/shrshi) -- Update to fmt 10.1.1 and spdlog 1.12.0. ([#14355](https://github.com/rapidsai/cudf/pull/14355)) [@bdice](https://github.com/bdice) -- Replace default stream for scalars and column factories usages (because of defaulted arguments) ([#14354](https://github.com/rapidsai/cudf/pull/14354)) [@karthikeyann](https://github.com/karthikeyann) -- Expose streams in ORC reader and writer APIs ([#14350](https://github.com/rapidsai/cudf/pull/14350)) [@shrshi](https://github.com/shrshi) -- Convert compression and io to string axis type in IO benchmarks ([#14347](https://github.com/rapidsai/cudf/pull/14347)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Add cuDF devcontainers ([#14015](https://github.com/rapidsai/cudf/pull/14015)) [@trxcllnt](https://github.com/trxcllnt) -- Refactoring of Buffers (last step towards unifying COW and Spilling) ([#13801](https://github.com/rapidsai/cudf/pull/13801)) [@madsbk](https://github.com/madsbk) -- Switch to scikit-build-core ([#13531](https://github.com/rapidsai/cudf/pull/13531)) [@vyasr](https://github.com/vyasr) -- Simplify null count checking in column equality comparator ([#13312](https://github.com/rapidsai/cudf/pull/13312)) [@vyasr](https://github.com/vyasr) +- Pin `pytest<8` ([#14920](https://github.com/NVIDIA/cudf/pull/14920)) [@galipremsagar](https://github.com/galipremsagar) +- Move cudf::char_utf8 definition from detail to public header ([#14779](https://github.com/NVIDIA/cudf/pull/14779)) [@davidwendt](https://github.com/davidwendt) +- Clean up `TimedeltaIndex.__init__` constructor ([#14775](https://github.com/NVIDIA/cudf/pull/14775)) [@mroeschke](https://github.com/mroeschke) +- Clean up `DatetimeIndex.__init__` constructor ([#14774](https://github.com/NVIDIA/cudf/pull/14774)) [@mroeschke](https://github.com/mroeschke) +- Some `frame.py` typing, move seldom used methods in `frame.py` ([#14766](https://github.com/NVIDIA/cudf/pull/14766)) [@mroeschke](https://github.com/mroeschke) +- Remove **kwargs from astype ([#14765](https://github.com/NVIDIA/cudf/pull/14765)) [@mroeschke](https://github.com/mroeschke) +- fix benchmarks compatibility with newer pytest-cases ([#14764](https://github.com/NVIDIA/cudf/pull/14764)) [@jameslamb](https://github.com/jameslamb) +- Add `pynvjitlink` as a dependency ([#14763](https://github.com/NVIDIA/cudf/pull/14763)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Resolve degenerate performance in `create_structs_data` ([#14761](https://github.com/NVIDIA/cudf/pull/14761)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Simplify ColumnAccessor methods; avoid unnecessary validations ([#14758](https://github.com/NVIDIA/cudf/pull/14758)) [@mroeschke](https://github.com/mroeschke) +- Pin pytest-cases<3.8.2 ([#14756](https://github.com/NVIDIA/cudf/pull/14756)) [@mroeschke](https://github.com/mroeschke) +- Use _from_data instead of _from_columns for initialzing Frame ([#14755](https://github.com/NVIDIA/cudf/pull/14755)) [@mroeschke](https://github.com/mroeschke) +- Consolidate cudf object handling in as_column ([#14754](https://github.com/NVIDIA/cudf/pull/14754)) [@mroeschke](https://github.com/mroeschke) +- Reduce execution time of Parquet C++ tests ([#14750](https://github.com/NVIDIA/cudf/pull/14750)) [@vuule](https://github.com/vuule) +- Implement to_datetime(..., utc=True) ([#14749](https://github.com/NVIDIA/cudf/pull/14749)) [@mroeschke](https://github.com/mroeschke) +- Remove usages of rapids-env-update ([#14748](https://github.com/NVIDIA/cudf/pull/14748)) [@KyleFromNVIDIA](https://github.com/KyleFromNVIDIA) +- Provide explicit pool size and avoid RMM detail APIs ([#14741](https://github.com/NVIDIA/cudf/pull/14741)) [@harrism](https://github.com/harrism) +- Implement `cudf.MultiIndex.from_arrays` ([#14740](https://github.com/NVIDIA/cudf/pull/14740)) [@mroeschke](https://github.com/mroeschke) +- Remove unused/single use methods ([#14739](https://github.com/NVIDIA/cudf/pull/14739)) [@mroeschke](https://github.com/mroeschke) +- refactor CUDA versions in dependencies.yaml ([#14733](https://github.com/NVIDIA/cudf/pull/14733)) [@jameslamb](https://github.com/jameslamb) +- Remove unneeded methods in Column ([#14730](https://github.com/NVIDIA/cudf/pull/14730)) [@mroeschke](https://github.com/mroeschke) +- Clean up base column methods ([#14725](https://github.com/NVIDIA/cudf/pull/14725)) [@mroeschke](https://github.com/mroeschke) +- Ensure column.fillna signatures are consistent ([#14724](https://github.com/NVIDIA/cudf/pull/14724)) [@mroeschke](https://github.com/mroeschke) +- Remove mimesis as a testing dependency ([#14723](https://github.com/NVIDIA/cudf/pull/14723)) [@mroeschke](https://github.com/mroeschke) +- Replace as_numerical with as_numerical_column/codes ([#14719](https://github.com/NVIDIA/cudf/pull/14719)) [@mroeschke](https://github.com/mroeschke) +- Use offsetalator in gather_chars ([#14700](https://github.com/NVIDIA/cudf/pull/14700)) [@davidwendt](https://github.com/davidwendt) +- Use make_strings_children for fill() specialization logic ([#14697](https://github.com/NVIDIA/cudf/pull/14697)) [@davidwendt](https://github.com/davidwendt) +- Change `io::detail::orc` namespace into `io::orc::detail` ([#14696](https://github.com/NVIDIA/cudf/pull/14696)) [@ttnghia](https://github.com/ttnghia) +- Fix call to deprecated factory function ([#14695](https://github.com/NVIDIA/cudf/pull/14695)) [@davidwendt](https://github.com/davidwendt) +- Use as_column instead of arange for range like inputs ([#14689](https://github.com/NVIDIA/cudf/pull/14689)) [@mroeschke](https://github.com/mroeschke) +- Reorganize ORC reader into multiple files and perform some small fixes to cuIO code ([#14665](https://github.com/NVIDIA/cudf/pull/14665)) [@ttnghia](https://github.com/ttnghia) +- Split parquet test into multiple files ([#14663](https://github.com/NVIDIA/cudf/pull/14663)) [@etseidl](https://github.com/etseidl) +- Custom error messages for IO with nonexistent files ([#14662](https://github.com/NVIDIA/cudf/pull/14662)) [@vuule](https://github.com/vuule) +- Explicitly pass .dtype into is_foo_dtype functions ([#14657](https://github.com/NVIDIA/cudf/pull/14657)) [@mroeschke](https://github.com/mroeschke) +- Basic validation in reader benchmarks ([#14647](https://github.com/NVIDIA/cudf/pull/14647)) [@vuule](https://github.com/vuule) +- Update dependencies.yaml to support CUDA 12.*. ([#14644](https://github.com/NVIDIA/cudf/pull/14644)) [@bdice](https://github.com/bdice) +- Consolidate memoryview handling in as_column ([#14643](https://github.com/NVIDIA/cudf/pull/14643)) [@mroeschke](https://github.com/mroeschke) +- Convert `FieldType` to scoped enum ([#14642](https://github.com/NVIDIA/cudf/pull/14642)) [@vuule](https://github.com/vuule) +- Use instance over is_foo_dtype ([#14641](https://github.com/NVIDIA/cudf/pull/14641)) [@mroeschke](https://github.com/mroeschke) +- Use isinstance over is_foo_dtype internally ([#14638](https://github.com/NVIDIA/cudf/pull/14638)) [@mroeschke](https://github.com/mroeschke) +- Remove unnecessary **kwargs in function signatures ([#14635](https://github.com/NVIDIA/cudf/pull/14635)) [@mroeschke](https://github.com/mroeschke) +- Drop nvbench patch for nvml. ([#14631](https://github.com/NVIDIA/cudf/pull/14631)) [@bdice](https://github.com/bdice) +- Drop Pascal GPU support. ([#14630](https://github.com/NVIDIA/cudf/pull/14630)) [@bdice](https://github.com/bdice) +- Add cpp/doxygen/xml to .gitignore ([#14613](https://github.com/NVIDIA/cudf/pull/14613)) [@davidwendt](https://github.com/davidwendt) +- Create strings-specific make_offsets_child_column for multiple offset types ([#14612](https://github.com/NVIDIA/cudf/pull/14612)) [@davidwendt](https://github.com/davidwendt) +- Use the offsetalator in cudf::concatenate for strings ([#14611](https://github.com/NVIDIA/cudf/pull/14611)) [@davidwendt](https://github.com/davidwendt) +- Make Parquet ColumnIndex null_counts optional ([#14596](https://github.com/NVIDIA/cudf/pull/14596)) [@etseidl](https://github.com/etseidl) +- Support `freq` in DatetimeIndex ([#14593](https://github.com/NVIDIA/cudf/pull/14593)) [@shwina](https://github.com/shwina) +- Remove legacy benchmarks for cuDF-python ([#14591](https://github.com/NVIDIA/cudf/pull/14591)) [@osidekyle](https://github.com/osidekyle) +- Remove WORKSPACE env var from cudf_test temp_directory class ([#14588](https://github.com/NVIDIA/cudf/pull/14588)) [@davidwendt](https://github.com/davidwendt) +- Use exceptions instead of return values to handle errors in `CompactProtocolReader` ([#14582](https://github.com/NVIDIA/cudf/pull/14582)) [@vuule](https://github.com/vuule) +- Use cuda::proclaim_return_type on device lambdas. ([#14577](https://github.com/NVIDIA/cudf/pull/14577)) [@bdice](https://github.com/bdice) +- Update to CCCL 2.2.0. ([#14576](https://github.com/NVIDIA/cudf/pull/14576)) [@bdice](https://github.com/bdice) +- Update dependencies.yaml to new pip index ([#14575](https://github.com/NVIDIA/cudf/pull/14575)) [@vyasr](https://github.com/vyasr) +- Simplify Python CMake ([#14565](https://github.com/NVIDIA/cudf/pull/14565)) [@vyasr](https://github.com/vyasr) +- Java expose parquet pass_read_limit ([#14564](https://github.com/NVIDIA/cudf/pull/14564)) [@revans2](https://github.com/revans2) +- Add column sanitization checks in `CUDF_TEST_EXPECT_COLUMN_*` macros ([#14559](https://github.com/NVIDIA/cudf/pull/14559)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Use cudf_test temp_directory class for nvtext::subword_tokenize gbenchmark ([#14558](https://github.com/NVIDIA/cudf/pull/14558)) [@davidwendt](https://github.com/davidwendt) +- Fix return type of prefix increment overloads ([#14544](https://github.com/NVIDIA/cudf/pull/14544)) [@vuule](https://github.com/vuule) +- Make bpe_merge_pairs_impl member private ([#14543](https://github.com/NVIDIA/cudf/pull/14543)) [@davidwendt](https://github.com/davidwendt) +- Small clean up in `io::statistics` ([#14542](https://github.com/NVIDIA/cudf/pull/14542)) [@vuule](https://github.com/vuule) +- Change json gtest environment variable to compile-time definition ([#14541](https://github.com/NVIDIA/cudf/pull/14541)) [@davidwendt](https://github.com/davidwendt) +- Remove extra total chars size calculation from cudf::concatenate ([#14540](https://github.com/NVIDIA/cudf/pull/14540)) [@davidwendt](https://github.com/davidwendt) +- Refactor IndexedFrame.hash_values to use cudf::hashing functions, add xxhash64 to cudf Python. ([#14538](https://github.com/NVIDIA/cudf/pull/14538)) [@bdice](https://github.com/bdice) +- Move non-templated inline function definitions from table_view.hpp to table_view.cpp ([#14535](https://github.com/NVIDIA/cudf/pull/14535)) [@davidwendt](https://github.com/davidwendt) +- Add JNI for strings::code_points ([#14533](https://github.com/NVIDIA/cudf/pull/14533)) [@thirtiseven](https://github.com/thirtiseven) +- Add a test for issue 12773 ([#14529](https://github.com/NVIDIA/cudf/pull/14529)) [@vyasr](https://github.com/vyasr) +- Split libarrow build dependencies. ([#14506](https://github.com/NVIDIA/cudf/pull/14506)) [@bdice](https://github.com/bdice) +- Implement `IndexedFrame.duplicated` with `distinct_indices` + `scatter` ([#14493](https://github.com/NVIDIA/cudf/pull/14493)) [@wence-](https://github.com/wence-) +- Expunge as_frame conversions in Column algorithms ([#14491](https://github.com/NVIDIA/cudf/pull/14491)) [@wence-](https://github.com/wence-) +- Remove unsanitized null from input strings column in rank_tests.cpp ([#14475](https://github.com/NVIDIA/cudf/pull/14475)) [@davidwendt](https://github.com/davidwendt) +- Refactor Parquet kernel_error ([#14464](https://github.com/NVIDIA/cudf/pull/14464)) [@etseidl](https://github.com/etseidl) +- Deprecate cudf::make_strings_column accepting typed offsets ([#14461](https://github.com/NVIDIA/cudf/pull/14461)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated nvtext::load_merge_pairs_file ([#14460](https://github.com/NVIDIA/cudf/pull/14460)) [@davidwendt](https://github.com/davidwendt) +- Introduce Comprehensive Pathological Unit Tests for Issue #14409 ([#14459](https://github.com/NVIDIA/cudf/pull/14459)) [@aocsa](https://github.com/aocsa) +- Expose stream parameter in public nvtext APIs ([#14456](https://github.com/NVIDIA/cudf/pull/14456)) [@davidwendt](https://github.com/davidwendt) +- Include encode type in the error message when unsupported Parquet encoding is detected ([#14453](https://github.com/NVIDIA/cudf/pull/14453)) [@ZelboK](https://github.com/ZelboK) +- Remove null mask for zero nulls in json readers ([#14451](https://github.com/NVIDIA/cudf/pull/14451)) [@karthikeyann](https://github.com/karthikeyann) +- Refactor cudf.Series.__init__ ([#14450](https://github.com/NVIDIA/cudf/pull/14450)) [@mroeschke](https://github.com/mroeschke) +- Remove the use of `volatile` in Parquet ([#14448](https://github.com/NVIDIA/cudf/pull/14448)) [@vuule](https://github.com/vuule) +- REF: Remove **kwargs from to_pandas, raise if nullable is not implemented ([#14438](https://github.com/NVIDIA/cudf/pull/14438)) [@mroeschke](https://github.com/mroeschke) +- Testing stream pool implementation ([#14437](https://github.com/NVIDIA/cudf/pull/14437)) [@shrshi](https://github.com/shrshi) +- Match pandas join ordering obligations in pandas-compatible mode ([#14428](https://github.com/NVIDIA/cudf/pull/14428)) [@wence-](https://github.com/wence-) +- Forward-merge branch-23.12 to branch-24.02 ([#14426](https://github.com/NVIDIA/cudf/pull/14426)) [@bdice](https://github.com/bdice) +- Use isinstance(..., cudf.IntervalDtype) instead of is_interval_dtype ([#14424](https://github.com/NVIDIA/cudf/pull/14424)) [@mroeschke](https://github.com/mroeschke) +- Use isinstance(..., cudf.CategoricalDtype) instead of is_categorical_dtype ([#14423](https://github.com/NVIDIA/cudf/pull/14423)) [@mroeschke](https://github.com/mroeschke) +- Forward-merge branch-23.12 to branch-24.02 ([#14422](https://github.com/NVIDIA/cudf/pull/14422)) [@bdice](https://github.com/bdice) +- REF: Remove instances of pd.core ([#14421](https://github.com/NVIDIA/cudf/pull/14421)) [@mroeschke](https://github.com/mroeschke) +- Expose streams in public filling APIs for label_bins ([#14401](https://github.com/NVIDIA/cudf/pull/14401)) [@ZelboK](https://github.com/ZelboK) +- Consolidate 1D pandas object handling in as_column ([#14394](https://github.com/NVIDIA/cudf/pull/14394)) [@mroeschke](https://github.com/mroeschke) +- Limit DELTA_BINARY_PACKED encoder to the same number of bits as the physical type being encoded ([#14392](https://github.com/NVIDIA/cudf/pull/14392)) [@etseidl](https://github.com/etseidl) +- Add SHA-1 and SHA-2 hash functions. ([#14391](https://github.com/NVIDIA/cudf/pull/14391)) [@bdice](https://github.com/bdice) +- Expose streams in Parquet reader and writer APIs ([#14359](https://github.com/NVIDIA/cudf/pull/14359)) [@shrshi](https://github.com/shrshi) +- Update to fmt 10.1.1 and spdlog 1.12.0. ([#14355](https://github.com/NVIDIA/cudf/pull/14355)) [@bdice](https://github.com/bdice) +- Replace default stream for scalars and column factories usages (because of defaulted arguments) ([#14354](https://github.com/NVIDIA/cudf/pull/14354)) [@karthikeyann](https://github.com/karthikeyann) +- Expose streams in ORC reader and writer APIs ([#14350](https://github.com/NVIDIA/cudf/pull/14350)) [@shrshi](https://github.com/shrshi) +- Convert compression and io to string axis type in IO benchmarks ([#14347](https://github.com/NVIDIA/cudf/pull/14347)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Add cuDF devcontainers ([#14015](https://github.com/NVIDIA/cudf/pull/14015)) [@trxcllnt](https://github.com/trxcllnt) +- Refactoring of Buffers (last step towards unifying COW and Spilling) ([#13801](https://github.com/NVIDIA/cudf/pull/13801)) [@madsbk](https://github.com/madsbk) +- Switch to scikit-build-core ([#13531](https://github.com/NVIDIA/cudf/pull/13531)) [@vyasr](https://github.com/vyasr) +- Simplify null count checking in column equality comparator ([#13312](https://github.com/NVIDIA/cudf/pull/13312)) [@vyasr](https://github.com/vyasr) # cuDF 23.12.00 (6 Dec 2023) ## 🚨 Breaking Changes -- Raise error in `reindex` when `index` is not unique ([#14400](https://github.com/rapidsai/cudf/pull/14400)) [@galipremsagar](https://github.com/galipremsagar) -- Expose stream parameter to get_json_object API ([#14297](https://github.com/rapidsai/cudf/pull/14297)) [@davidwendt](https://github.com/davidwendt) -- Refactor cudf_kafka to use skbuild ([#14292](https://github.com/rapidsai/cudf/pull/14292)) [@jdye64](https://github.com/jdye64) -- Expose stream parameter in public strings convert APIs ([#14255](https://github.com/rapidsai/cudf/pull/14255)) [@davidwendt](https://github.com/davidwendt) -- Upgrade to nvCOMP 3.0.4 ([#13815](https://github.com/rapidsai/cudf/pull/13815)) [@vuule](https://github.com/vuule) +- Raise error in `reindex` when `index` is not unique ([#14400](https://github.com/NVIDIA/cudf/pull/14400)) [@galipremsagar](https://github.com/galipremsagar) +- Expose stream parameter to get_json_object API ([#14297](https://github.com/NVIDIA/cudf/pull/14297)) [@davidwendt](https://github.com/davidwendt) +- Refactor cudf_kafka to use skbuild ([#14292](https://github.com/NVIDIA/cudf/pull/14292)) [@jdye64](https://github.com/jdye64) +- Expose stream parameter in public strings convert APIs ([#14255](https://github.com/NVIDIA/cudf/pull/14255)) [@davidwendt](https://github.com/davidwendt) +- Upgrade to nvCOMP 3.0.4 ([#13815](https://github.com/NVIDIA/cudf/pull/13815)) [@vuule](https://github.com/vuule) ## 🐛 Bug Fixes -- Update actions/labeler to v4 ([#14562](https://github.com/rapidsai/cudf/pull/14562)) [@raydouglass](https://github.com/raydouglass) -- Fix data corruption when skipping rows ([#14557](https://github.com/rapidsai/cudf/pull/14557)) [@etseidl](https://github.com/etseidl) -- Fix function name typo in `cudf.pandas` profiler ([#14514](https://github.com/rapidsai/cudf/pull/14514)) [@galipremsagar](https://github.com/galipremsagar) -- Fix intermediate type checking in expression parsing ([#14445](https://github.com/rapidsai/cudf/pull/14445)) [@vyasr](https://github.com/vyasr) -- Forward merge `branch-23.10` into `branch-23.12` ([#14435](https://github.com/rapidsai/cudf/pull/14435)) [@raydouglass](https://github.com/raydouglass) -- Remove needs: wheel-build-cudf. ([#14427](https://github.com/rapidsai/cudf/pull/14427)) [@bdice](https://github.com/bdice) -- Fix dask dependency in custreamz ([#14420](https://github.com/rapidsai/cudf/pull/14420)) [@vyasr](https://github.com/vyasr) -- Ensure nvbench initializes nvml context when built statically ([#14411](https://github.com/rapidsai/cudf/pull/14411)) [@robertmaynard](https://github.com/robertmaynard) -- Support java AST String literal with desired encoding ([#14402](https://github.com/rapidsai/cudf/pull/14402)) [@winningsix](https://github.com/winningsix) -- Raise error in `reindex` when `index` is not unique ([#14400](https://github.com/rapidsai/cudf/pull/14400)) [@galipremsagar](https://github.com/galipremsagar) -- Always build nvbench statically so we don't need to package it ([#14399](https://github.com/rapidsai/cudf/pull/14399)) [@robertmaynard](https://github.com/robertmaynard) -- Fix token-count logic in nvtext::tokenize_with_vocabulary ([#14393](https://github.com/rapidsai/cudf/pull/14393)) [@davidwendt](https://github.com/davidwendt) -- Fix as_column(pd.Timestamp/Timedelta, length=) not respecting length ([#14390](https://github.com/rapidsai/cudf/pull/14390)) [@mroeschke](https://github.com/mroeschke) -- cudf.pandas: cuDF subpath checking in module `__getattr__` ([#14388](https://github.com/rapidsai/cudf/pull/14388)) [@shwina](https://github.com/shwina) -- Fix and disable encoding for nanosecond statistics in ORC writer ([#14367](https://github.com/rapidsai/cudf/pull/14367)) [@vuule](https://github.com/vuule) -- Add the new manylinux builds to the build job ([#14351](https://github.com/rapidsai/cudf/pull/14351)) [@vyasr](https://github.com/vyasr) -- cudf jit parser now supports .pragma instructions with quotes ([#14348](https://github.com/rapidsai/cudf/pull/14348)) [@robertmaynard](https://github.com/robertmaynard) -- Fix overflow check in `cudf::merge` ([#14345](https://github.com/rapidsai/cudf/pull/14345)) [@divyegala](https://github.com/divyegala) -- Add cramjam ([#14344](https://github.com/rapidsai/cudf/pull/14344)) [@vyasr](https://github.com/vyasr) -- Enable `dask_cudf/io` pytests in CI ([#14338](https://github.com/rapidsai/cudf/pull/14338)) [@galipremsagar](https://github.com/galipremsagar) -- Temporarily avoid the current build of pydata-sphinx-theme ([#14332](https://github.com/rapidsai/cudf/pull/14332)) [@vyasr](https://github.com/vyasr) -- Fix host buffer access from device function in the Parquet reader ([#14328](https://github.com/rapidsai/cudf/pull/14328)) [@vuule](https://github.com/vuule) -- Run IO tests for Dask-cuDF ([#14327](https://github.com/rapidsai/cudf/pull/14327)) [@rjzamora](https://github.com/rjzamora) -- Fix logical type issues in the Parquet writer ([#14322](https://github.com/rapidsai/cudf/pull/14322)) [@vuule](https://github.com/vuule) -- Remove aws-sdk-pinning and revert to arrow 12.0.1 ([#14319](https://github.com/rapidsai/cudf/pull/14319)) [@vyasr](https://github.com/vyasr) -- test is_valid before reading column data ([#14318](https://github.com/rapidsai/cudf/pull/14318)) [@etseidl](https://github.com/etseidl) -- Fix gtest validity setting for TextTokenizeTest.Vocabulary ([#14312](https://github.com/rapidsai/cudf/pull/14312)) [@davidwendt](https://github.com/davidwendt) -- Fixes stack context for json lines format that recovers from invalid JSON lines ([#14309](https://github.com/rapidsai/cudf/pull/14309)) [@elstehle](https://github.com/elstehle) -- Downgrade to Arrow 12.0.0 for aws-sdk-cpp and fix cudf_kafka builds for new CI containers ([#14296](https://github.com/rapidsai/cudf/pull/14296)) [@vyasr](https://github.com/vyasr) -- fixing thread index overflow issue ([#14290](https://github.com/rapidsai/cudf/pull/14290)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix memset error in nvtext::edit_distance_matrix ([#14283](https://github.com/rapidsai/cudf/pull/14283)) [@davidwendt](https://github.com/davidwendt) -- Changes JSON reader's recovery option's behaviour to ignore all characters after a valid JSON record ([#14279](https://github.com/rapidsai/cudf/pull/14279)) [@elstehle](https://github.com/elstehle) -- Handle empty string correctly in Parquet statistics ([#14257](https://github.com/rapidsai/cudf/pull/14257)) [@etseidl](https://github.com/etseidl) -- Fixes behaviour for incomplete lines when `recover_with_nulls` is enabled ([#14252](https://github.com/rapidsai/cudf/pull/14252)) [@elstehle](https://github.com/elstehle) -- cudf::detail::pinned_allocator doesn't throw from `deallocate` ([#14251](https://github.com/rapidsai/cudf/pull/14251)) [@robertmaynard](https://github.com/robertmaynard) -- Fix strings replace for adjacent, identical multi-byte UTF-8 character targets ([#14235](https://github.com/rapidsai/cudf/pull/14235)) [@davidwendt](https://github.com/davidwendt) -- Fix the precision when converting a decimal128 column to an arrow array ([#14230](https://github.com/rapidsai/cudf/pull/14230)) [@jihoonson](https://github.com/jihoonson) -- Fixing parquet list of struct interpretation ([#13715](https://github.com/rapidsai/cudf/pull/13715)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Update actions/labeler to v4 ([#14562](https://github.com/NVIDIA/cudf/pull/14562)) [@raydouglass](https://github.com/raydouglass) +- Fix data corruption when skipping rows ([#14557](https://github.com/NVIDIA/cudf/pull/14557)) [@etseidl](https://github.com/etseidl) +- Fix function name typo in `cudf.pandas` profiler ([#14514](https://github.com/NVIDIA/cudf/pull/14514)) [@galipremsagar](https://github.com/galipremsagar) +- Fix intermediate type checking in expression parsing ([#14445](https://github.com/NVIDIA/cudf/pull/14445)) [@vyasr](https://github.com/vyasr) +- Forward merge `branch-23.10` into `branch-23.12` ([#14435](https://github.com/NVIDIA/cudf/pull/14435)) [@raydouglass](https://github.com/raydouglass) +- Remove needs: wheel-build-cudf. ([#14427](https://github.com/NVIDIA/cudf/pull/14427)) [@bdice](https://github.com/bdice) +- Fix dask dependency in custreamz ([#14420](https://github.com/NVIDIA/cudf/pull/14420)) [@vyasr](https://github.com/vyasr) +- Ensure nvbench initializes nvml context when built statically ([#14411](https://github.com/NVIDIA/cudf/pull/14411)) [@robertmaynard](https://github.com/robertmaynard) +- Support java AST String literal with desired encoding ([#14402](https://github.com/NVIDIA/cudf/pull/14402)) [@winningsix](https://github.com/winningsix) +- Raise error in `reindex` when `index` is not unique ([#14400](https://github.com/NVIDIA/cudf/pull/14400)) [@galipremsagar](https://github.com/galipremsagar) +- Always build nvbench statically so we don't need to package it ([#14399](https://github.com/NVIDIA/cudf/pull/14399)) [@robertmaynard](https://github.com/robertmaynard) +- Fix token-count logic in nvtext::tokenize_with_vocabulary ([#14393](https://github.com/NVIDIA/cudf/pull/14393)) [@davidwendt](https://github.com/davidwendt) +- Fix as_column(pd.Timestamp/Timedelta, length=) not respecting length ([#14390](https://github.com/NVIDIA/cudf/pull/14390)) [@mroeschke](https://github.com/mroeschke) +- cudf.pandas: cuDF subpath checking in module `__getattr__` ([#14388](https://github.com/NVIDIA/cudf/pull/14388)) [@shwina](https://github.com/shwina) +- Fix and disable encoding for nanosecond statistics in ORC writer ([#14367](https://github.com/NVIDIA/cudf/pull/14367)) [@vuule](https://github.com/vuule) +- Add the new manylinux builds to the build job ([#14351](https://github.com/NVIDIA/cudf/pull/14351)) [@vyasr](https://github.com/vyasr) +- cudf jit parser now supports .pragma instructions with quotes ([#14348](https://github.com/NVIDIA/cudf/pull/14348)) [@robertmaynard](https://github.com/robertmaynard) +- Fix overflow check in `cudf::merge` ([#14345](https://github.com/NVIDIA/cudf/pull/14345)) [@divyegala](https://github.com/divyegala) +- Add cramjam ([#14344](https://github.com/NVIDIA/cudf/pull/14344)) [@vyasr](https://github.com/vyasr) +- Enable `dask_cudf/io` pytests in CI ([#14338](https://github.com/NVIDIA/cudf/pull/14338)) [@galipremsagar](https://github.com/galipremsagar) +- Temporarily avoid the current build of pydata-sphinx-theme ([#14332](https://github.com/NVIDIA/cudf/pull/14332)) [@vyasr](https://github.com/vyasr) +- Fix host buffer access from device function in the Parquet reader ([#14328](https://github.com/NVIDIA/cudf/pull/14328)) [@vuule](https://github.com/vuule) +- Run IO tests for Dask-cuDF ([#14327](https://github.com/NVIDIA/cudf/pull/14327)) [@rjzamora](https://github.com/rjzamora) +- Fix logical type issues in the Parquet writer ([#14322](https://github.com/NVIDIA/cudf/pull/14322)) [@vuule](https://github.com/vuule) +- Remove aws-sdk-pinning and revert to arrow 12.0.1 ([#14319](https://github.com/NVIDIA/cudf/pull/14319)) [@vyasr](https://github.com/vyasr) +- test is_valid before reading column data ([#14318](https://github.com/NVIDIA/cudf/pull/14318)) [@etseidl](https://github.com/etseidl) +- Fix gtest validity setting for TextTokenizeTest.Vocabulary ([#14312](https://github.com/NVIDIA/cudf/pull/14312)) [@davidwendt](https://github.com/davidwendt) +- Fixes stack context for json lines format that recovers from invalid JSON lines ([#14309](https://github.com/NVIDIA/cudf/pull/14309)) [@elstehle](https://github.com/elstehle) +- Downgrade to Arrow 12.0.0 for aws-sdk-cpp and fix cudf_kafka builds for new CI containers ([#14296](https://github.com/NVIDIA/cudf/pull/14296)) [@vyasr](https://github.com/vyasr) +- fixing thread index overflow issue ([#14290](https://github.com/NVIDIA/cudf/pull/14290)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix memset error in nvtext::edit_distance_matrix ([#14283](https://github.com/NVIDIA/cudf/pull/14283)) [@davidwendt](https://github.com/davidwendt) +- Changes JSON reader's recovery option's behaviour to ignore all characters after a valid JSON record ([#14279](https://github.com/NVIDIA/cudf/pull/14279)) [@elstehle](https://github.com/elstehle) +- Handle empty string correctly in Parquet statistics ([#14257](https://github.com/NVIDIA/cudf/pull/14257)) [@etseidl](https://github.com/etseidl) +- Fixes behaviour for incomplete lines when `recover_with_nulls` is enabled ([#14252](https://github.com/NVIDIA/cudf/pull/14252)) [@elstehle](https://github.com/elstehle) +- cudf::detail::pinned_allocator doesn't throw from `deallocate` ([#14251](https://github.com/NVIDIA/cudf/pull/14251)) [@robertmaynard](https://github.com/robertmaynard) +- Fix strings replace for adjacent, identical multi-byte UTF-8 character targets ([#14235](https://github.com/NVIDIA/cudf/pull/14235)) [@davidwendt](https://github.com/davidwendt) +- Fix the precision when converting a decimal128 column to an arrow array ([#14230](https://github.com/NVIDIA/cudf/pull/14230)) [@jihoonson](https://github.com/jihoonson) +- Fixing parquet list of struct interpretation ([#13715](https://github.com/NVIDIA/cudf/pull/13715)) [@hyperbolic2346](https://github.com/hyperbolic2346) ## 📖 Documentation -- Fix io reference in docs. ([#14452](https://github.com/rapidsai/cudf/pull/14452)) [@bdice](https://github.com/bdice) -- Update README ([#14374](https://github.com/rapidsai/cudf/pull/14374)) [@shwina](https://github.com/shwina) -- Example code for blog on new row comparators ([#13795](https://github.com/rapidsai/cudf/pull/13795)) [@divyegala](https://github.com/divyegala) +- Fix io reference in docs. ([#14452](https://github.com/NVIDIA/cudf/pull/14452)) [@bdice](https://github.com/bdice) +- Update README ([#14374](https://github.com/NVIDIA/cudf/pull/14374)) [@shwina](https://github.com/shwina) +- Example code for blog on new row comparators ([#13795](https://github.com/NVIDIA/cudf/pull/13795)) [@divyegala](https://github.com/divyegala) ## 🚀 New Features -- Expose streams in public unary APIs ([#14342](https://github.com/rapidsai/cudf/pull/14342)) [@vyasr](https://github.com/vyasr) -- Add python tests for Parquet DELTA_BINARY_PACKED encoder ([#14316](https://github.com/rapidsai/cudf/pull/14316)) [@etseidl](https://github.com/etseidl) -- Update rapids-cmake functions to non-deprecated signatures ([#14265](https://github.com/rapidsai/cudf/pull/14265)) [@robertmaynard](https://github.com/robertmaynard) -- Expose streams in public null mask APIs ([#14263](https://github.com/rapidsai/cudf/pull/14263)) [@vyasr](https://github.com/vyasr) -- Expose streams in binaryop APIs ([#14187](https://github.com/rapidsai/cudf/pull/14187)) [@vyasr](https://github.com/vyasr) -- Add pylibcudf.Scalar that interoperates with Arrow scalars ([#14133](https://github.com/rapidsai/cudf/pull/14133)) [@vyasr](https://github.com/vyasr) -- Add decoder for DELTA_BYTE_ARRAY to Parquet reader ([#14101](https://github.com/rapidsai/cudf/pull/14101)) [@etseidl](https://github.com/etseidl) -- Add DELTA_BINARY_PACKED encoder for Parquet writer ([#14100](https://github.com/rapidsai/cudf/pull/14100)) [@etseidl](https://github.com/etseidl) -- Add BytePairEncoder class to cuDF ([#13891](https://github.com/rapidsai/cudf/pull/13891)) [@davidwendt](https://github.com/davidwendt) -- Upgrade to nvCOMP 3.0.4 ([#13815](https://github.com/rapidsai/cudf/pull/13815)) [@vuule](https://github.com/vuule) -- Use `pynvjitlink` for CUDA 12+ MVC ([#13650](https://github.com/rapidsai/cudf/pull/13650)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Expose streams in public unary APIs ([#14342](https://github.com/NVIDIA/cudf/pull/14342)) [@vyasr](https://github.com/vyasr) +- Add python tests for Parquet DELTA_BINARY_PACKED encoder ([#14316](https://github.com/NVIDIA/cudf/pull/14316)) [@etseidl](https://github.com/etseidl) +- Update rapids-cmake functions to non-deprecated signatures ([#14265](https://github.com/NVIDIA/cudf/pull/14265)) [@robertmaynard](https://github.com/robertmaynard) +- Expose streams in public null mask APIs ([#14263](https://github.com/NVIDIA/cudf/pull/14263)) [@vyasr](https://github.com/vyasr) +- Expose streams in binaryop APIs ([#14187](https://github.com/NVIDIA/cudf/pull/14187)) [@vyasr](https://github.com/vyasr) +- Add pylibcudf.Scalar that interoperates with Arrow scalars ([#14133](https://github.com/NVIDIA/cudf/pull/14133)) [@vyasr](https://github.com/vyasr) +- Add decoder for DELTA_BYTE_ARRAY to Parquet reader ([#14101](https://github.com/NVIDIA/cudf/pull/14101)) [@etseidl](https://github.com/etseidl) +- Add DELTA_BINARY_PACKED encoder for Parquet writer ([#14100](https://github.com/NVIDIA/cudf/pull/14100)) [@etseidl](https://github.com/etseidl) +- Add BytePairEncoder class to cuDF ([#13891](https://github.com/NVIDIA/cudf/pull/13891)) [@davidwendt](https://github.com/davidwendt) +- Upgrade to nvCOMP 3.0.4 ([#13815](https://github.com/NVIDIA/cudf/pull/13815)) [@vuule](https://github.com/vuule) +- Use `pynvjitlink` for CUDA 12+ MVC ([#13650](https://github.com/NVIDIA/cudf/pull/13650)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🛠️ Improvements -- Build concurrency for nightly and merge triggers ([#14441](https://github.com/rapidsai/cudf/pull/14441)) [@bdice](https://github.com/bdice) -- Cleanup remaining usages of dask dependencies ([#14407](https://github.com/rapidsai/cudf/pull/14407)) [@galipremsagar](https://github.com/galipremsagar) -- Update to Arrow 14.0.1. ([#14387](https://github.com/rapidsai/cudf/pull/14387)) [@bdice](https://github.com/bdice) -- Remove Cython libcpp wrappers ([#14382](https://github.com/rapidsai/cudf/pull/14382)) [@vyasr](https://github.com/vyasr) -- Forward-merge branch-23.10 to branch-23.12 ([#14372](https://github.com/rapidsai/cudf/pull/14372)) [@bdice](https://github.com/bdice) -- Upgrade to arrow 14 ([#14371](https://github.com/rapidsai/cudf/pull/14371)) [@galipremsagar](https://github.com/galipremsagar) -- Fix a pytest typo in `test_kurt_skew_error` ([#14368](https://github.com/rapidsai/cudf/pull/14368)) [@galipremsagar](https://github.com/galipremsagar) -- Use new rapids-dask-dependency metapackage for managing dask versions ([#14364](https://github.com/rapidsai/cudf/pull/14364)) [@vyasr](https://github.com/vyasr) -- Change `nullable()` to `has_nulls()` in `cudf::detail::gather` ([#14363](https://github.com/rapidsai/cudf/pull/14363)) [@divyegala](https://github.com/divyegala) -- Split up scan_inclusive.cu to improve its compile time ([#14358](https://github.com/rapidsai/cudf/pull/14358)) [@davidwendt](https://github.com/davidwendt) -- Implement user_datasource_wrapper is_empty() and is_device_read_preferred(). ([#14357](https://github.com/rapidsai/cudf/pull/14357)) [@tpn](https://github.com/tpn) -- Added streams to CSV reader and writer api ([#14340](https://github.com/rapidsai/cudf/pull/14340)) [@shrshi](https://github.com/shrshi) -- Upgrade wheels to use arrow 13 ([#14339](https://github.com/rapidsai/cudf/pull/14339)) [@vyasr](https://github.com/vyasr) -- Rework nvtext::byte_pair_encoding API ([#14337](https://github.com/rapidsai/cudf/pull/14337)) [@davidwendt](https://github.com/davidwendt) -- Improve performance of nvtext::tokenize_with_vocabulary for long strings ([#14336](https://github.com/rapidsai/cudf/pull/14336)) [@davidwendt](https://github.com/davidwendt) -- Upgrade `arrow` to `13` ([#14330](https://github.com/rapidsai/cudf/pull/14330)) [@galipremsagar](https://github.com/galipremsagar) -- Expose stream parameter in public nvtext replace APIs ([#14329](https://github.com/rapidsai/cudf/pull/14329)) [@davidwendt](https://github.com/davidwendt) -- Drop `pyorc` dependency and use `pandas`/`pyarrow` instead ([#14323](https://github.com/rapidsai/cudf/pull/14323)) [@galipremsagar](https://github.com/galipremsagar) -- Avoid `pyarrow.fs` import for local storage ([#14321](https://github.com/rapidsai/cudf/pull/14321)) [@rjzamora](https://github.com/rjzamora) -- Unpin `dask` and `distributed` for `23.12` development ([#14320](https://github.com/rapidsai/cudf/pull/14320)) [@galipremsagar](https://github.com/galipremsagar) -- Expose stream parameter in public nvtext tokenize APIs ([#14317](https://github.com/rapidsai/cudf/pull/14317)) [@davidwendt](https://github.com/davidwendt) -- Added streams to JSON reader and writer api ([#14313](https://github.com/rapidsai/cudf/pull/14313)) [@shrshi](https://github.com/shrshi) -- Minor improvements in `source_info` ([#14308](https://github.com/rapidsai/cudf/pull/14308)) [@vuule](https://github.com/vuule) -- Forward-merge branch-23.10 to branch-23.12 ([#14307](https://github.com/rapidsai/cudf/pull/14307)) [@bdice](https://github.com/bdice) -- Add stream parameter to Set Operations (Public List APIs) ([#14305](https://github.com/rapidsai/cudf/pull/14305)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Expose stream parameter to get_json_object API ([#14297](https://github.com/rapidsai/cudf/pull/14297)) [@davidwendt](https://github.com/davidwendt) -- Sort dictionary data alphabetically in the ORC writer ([#14295](https://github.com/rapidsai/cudf/pull/14295)) [@vuule](https://github.com/vuule) -- Expose stream parameter in public strings filter APIs ([#14293](https://github.com/rapidsai/cudf/pull/14293)) [@davidwendt](https://github.com/davidwendt) -- Refactor cudf_kafka to use skbuild ([#14292](https://github.com/rapidsai/cudf/pull/14292)) [@jdye64](https://github.com/jdye64) -- Update `shared-action-workflows` references ([#14289](https://github.com/rapidsai/cudf/pull/14289)) [@AyodeAwe](https://github.com/AyodeAwe) -- Register ``partd`` encode dispatch in ``dask_cudf`` ([#14287](https://github.com/rapidsai/cudf/pull/14287)) [@rjzamora](https://github.com/rjzamora) -- Update versioning strategy ([#14285](https://github.com/rapidsai/cudf/pull/14285)) [@vyasr](https://github.com/vyasr) -- Move and rename byte-pair-encoding source files ([#14284](https://github.com/rapidsai/cudf/pull/14284)) [@davidwendt](https://github.com/davidwendt) -- Expose stream parameter in public strings combine APIs ([#14281](https://github.com/rapidsai/cudf/pull/14281)) [@davidwendt](https://github.com/davidwendt) -- Expose stream parameter in public strings contains APIs ([#14280](https://github.com/rapidsai/cudf/pull/14280)) [@davidwendt](https://github.com/davidwendt) -- Add stream parameter to List Sort and Filter APIs ([#14272](https://github.com/rapidsai/cudf/pull/14272)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Use branch-23.12 workflows. ([#14271](https://github.com/rapidsai/cudf/pull/14271)) [@bdice](https://github.com/bdice) -- Refactor LogicalType for Parquet ([#14264](https://github.com/rapidsai/cudf/pull/14264)) [@etseidl](https://github.com/etseidl) -- Centralize chunked reading code in the parquet reader to reader_impl_chunking.cu ([#14262](https://github.com/rapidsai/cudf/pull/14262)) [@nvdbaranec](https://github.com/nvdbaranec) -- Expose stream parameter in public strings replace APIs ([#14261](https://github.com/rapidsai/cudf/pull/14261)) [@davidwendt](https://github.com/davidwendt) -- Expose stream parameter in public strings APIs ([#14260](https://github.com/rapidsai/cudf/pull/14260)) [@davidwendt](https://github.com/davidwendt) -- Cleanup of namespaces in parquet code. ([#14259](https://github.com/rapidsai/cudf/pull/14259)) [@nvdbaranec](https://github.com/nvdbaranec) -- Make parquet schema index type consistent ([#14256](https://github.com/rapidsai/cudf/pull/14256)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Expose stream parameter in public strings convert APIs ([#14255](https://github.com/rapidsai/cudf/pull/14255)) [@davidwendt](https://github.com/davidwendt) -- Add in java bindings for DataSource ([#14254](https://github.com/rapidsai/cudf/pull/14254)) [@revans2](https://github.com/revans2) -- Reimplement `cudf::merge` for nested types without using comparators ([#14250](https://github.com/rapidsai/cudf/pull/14250)) [@divyegala](https://github.com/divyegala) -- Add stream parameter to List Manipulation and Operations APIs ([#14248](https://github.com/rapidsai/cudf/pull/14248)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Expose stream parameter in public strings split/partition APIs ([#14247](https://github.com/rapidsai/cudf/pull/14247)) [@davidwendt](https://github.com/davidwendt) -- Improve `contains_column` by invoking `contains_table` ([#14238](https://github.com/rapidsai/cudf/pull/14238)) [@PointKernel](https://github.com/PointKernel) -- Detect and report errors in Parquet header parsing ([#14237](https://github.com/rapidsai/cudf/pull/14237)) [@etseidl](https://github.com/etseidl) -- Normalizing offsets iterator ([#14234](https://github.com/rapidsai/cudf/pull/14234)) [@davidwendt](https://github.com/davidwendt) -- Forward merge `23.10` into `23.12` ([#14231](https://github.com/rapidsai/cudf/pull/14231)) [@galipremsagar](https://github.com/galipremsagar) -- Return error if BOOL8 column-type is used with integers-to-hex ([#14208](https://github.com/rapidsai/cudf/pull/14208)) [@davidwendt](https://github.com/davidwendt) -- Enable indexalator for device code ([#14206](https://github.com/rapidsai/cudf/pull/14206)) [@davidwendt](https://github.com/davidwendt) -- Marginally reduce memory footprint of joins ([#14197](https://github.com/rapidsai/cudf/pull/14197)) [@wence-](https://github.com/wence-) -- Add nvtx annotations to spilling-based data movement ([#14196](https://github.com/rapidsai/cudf/pull/14196)) [@wence-](https://github.com/wence-) -- Optimize ORC writer for decimal columns ([#14190](https://github.com/rapidsai/cudf/pull/14190)) [@vuule](https://github.com/vuule) -- Remove the use of volatile in ORC ([#14175](https://github.com/rapidsai/cudf/pull/14175)) [@vuule](https://github.com/vuule) -- Add `bytes_per_second` to distinct_count of stream_compaction nvbench. ([#14172](https://github.com/rapidsai/cudf/pull/14172)) [@Blonck](https://github.com/Blonck) -- Add `bytes_per_second` to transpose benchmark ([#14170](https://github.com/rapidsai/cudf/pull/14170)) [@Blonck](https://github.com/Blonck) -- cuDF: Build CUDA 12.0 ARM conda packages. ([#14112](https://github.com/rapidsai/cudf/pull/14112)) [@bdice](https://github.com/bdice) -- Add `bytes_per_second` to shift benchmark ([#13950](https://github.com/rapidsai/cudf/pull/13950)) [@Blonck](https://github.com/Blonck) -- Extract `debug_utilities.hpp/cu` from `column_utilities.hpp/cu` ([#13720](https://github.com/rapidsai/cudf/pull/13720)) [@ttnghia](https://github.com/ttnghia) +- Build concurrency for nightly and merge triggers ([#14441](https://github.com/NVIDIA/cudf/pull/14441)) [@bdice](https://github.com/bdice) +- Cleanup remaining usages of dask dependencies ([#14407](https://github.com/NVIDIA/cudf/pull/14407)) [@galipremsagar](https://github.com/galipremsagar) +- Update to Arrow 14.0.1. ([#14387](https://github.com/NVIDIA/cudf/pull/14387)) [@bdice](https://github.com/bdice) +- Remove Cython libcpp wrappers ([#14382](https://github.com/NVIDIA/cudf/pull/14382)) [@vyasr](https://github.com/vyasr) +- Forward-merge branch-23.10 to branch-23.12 ([#14372](https://github.com/NVIDIA/cudf/pull/14372)) [@bdice](https://github.com/bdice) +- Upgrade to arrow 14 ([#14371](https://github.com/NVIDIA/cudf/pull/14371)) [@galipremsagar](https://github.com/galipremsagar) +- Fix a pytest typo in `test_kurt_skew_error` ([#14368](https://github.com/NVIDIA/cudf/pull/14368)) [@galipremsagar](https://github.com/galipremsagar) +- Use new rapids-dask-dependency metapackage for managing dask versions ([#14364](https://github.com/NVIDIA/cudf/pull/14364)) [@vyasr](https://github.com/vyasr) +- Change `nullable()` to `has_nulls()` in `cudf::detail::gather` ([#14363](https://github.com/NVIDIA/cudf/pull/14363)) [@divyegala](https://github.com/divyegala) +- Split up scan_inclusive.cu to improve its compile time ([#14358](https://github.com/NVIDIA/cudf/pull/14358)) [@davidwendt](https://github.com/davidwendt) +- Implement user_datasource_wrapper is_empty() and is_device_read_preferred(). ([#14357](https://github.com/NVIDIA/cudf/pull/14357)) [@tpn](https://github.com/tpn) +- Added streams to CSV reader and writer api ([#14340](https://github.com/NVIDIA/cudf/pull/14340)) [@shrshi](https://github.com/shrshi) +- Upgrade wheels to use arrow 13 ([#14339](https://github.com/NVIDIA/cudf/pull/14339)) [@vyasr](https://github.com/vyasr) +- Rework nvtext::byte_pair_encoding API ([#14337](https://github.com/NVIDIA/cudf/pull/14337)) [@davidwendt](https://github.com/davidwendt) +- Improve performance of nvtext::tokenize_with_vocabulary for long strings ([#14336](https://github.com/NVIDIA/cudf/pull/14336)) [@davidwendt](https://github.com/davidwendt) +- Upgrade `arrow` to `13` ([#14330](https://github.com/NVIDIA/cudf/pull/14330)) [@galipremsagar](https://github.com/galipremsagar) +- Expose stream parameter in public nvtext replace APIs ([#14329](https://github.com/NVIDIA/cudf/pull/14329)) [@davidwendt](https://github.com/davidwendt) +- Drop `pyorc` dependency and use `pandas`/`pyarrow` instead ([#14323](https://github.com/NVIDIA/cudf/pull/14323)) [@galipremsagar](https://github.com/galipremsagar) +- Avoid `pyarrow.fs` import for local storage ([#14321](https://github.com/NVIDIA/cudf/pull/14321)) [@rjzamora](https://github.com/rjzamora) +- Unpin `dask` and `distributed` for `23.12` development ([#14320](https://github.com/NVIDIA/cudf/pull/14320)) [@galipremsagar](https://github.com/galipremsagar) +- Expose stream parameter in public nvtext tokenize APIs ([#14317](https://github.com/NVIDIA/cudf/pull/14317)) [@davidwendt](https://github.com/davidwendt) +- Added streams to JSON reader and writer api ([#14313](https://github.com/NVIDIA/cudf/pull/14313)) [@shrshi](https://github.com/shrshi) +- Minor improvements in `source_info` ([#14308](https://github.com/NVIDIA/cudf/pull/14308)) [@vuule](https://github.com/vuule) +- Forward-merge branch-23.10 to branch-23.12 ([#14307](https://github.com/NVIDIA/cudf/pull/14307)) [@bdice](https://github.com/bdice) +- Add stream parameter to Set Operations (Public List APIs) ([#14305](https://github.com/NVIDIA/cudf/pull/14305)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Expose stream parameter to get_json_object API ([#14297](https://github.com/NVIDIA/cudf/pull/14297)) [@davidwendt](https://github.com/davidwendt) +- Sort dictionary data alphabetically in the ORC writer ([#14295](https://github.com/NVIDIA/cudf/pull/14295)) [@vuule](https://github.com/vuule) +- Expose stream parameter in public strings filter APIs ([#14293](https://github.com/NVIDIA/cudf/pull/14293)) [@davidwendt](https://github.com/davidwendt) +- Refactor cudf_kafka to use skbuild ([#14292](https://github.com/NVIDIA/cudf/pull/14292)) [@jdye64](https://github.com/jdye64) +- Update `shared-action-workflows` references ([#14289](https://github.com/NVIDIA/cudf/pull/14289)) [@AyodeAwe](https://github.com/AyodeAwe) +- Register ``partd`` encode dispatch in ``dask_cudf`` ([#14287](https://github.com/NVIDIA/cudf/pull/14287)) [@rjzamora](https://github.com/rjzamora) +- Update versioning strategy ([#14285](https://github.com/NVIDIA/cudf/pull/14285)) [@vyasr](https://github.com/vyasr) +- Move and rename byte-pair-encoding source files ([#14284](https://github.com/NVIDIA/cudf/pull/14284)) [@davidwendt](https://github.com/davidwendt) +- Expose stream parameter in public strings combine APIs ([#14281](https://github.com/NVIDIA/cudf/pull/14281)) [@davidwendt](https://github.com/davidwendt) +- Expose stream parameter in public strings contains APIs ([#14280](https://github.com/NVIDIA/cudf/pull/14280)) [@davidwendt](https://github.com/davidwendt) +- Add stream parameter to List Sort and Filter APIs ([#14272](https://github.com/NVIDIA/cudf/pull/14272)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Use branch-23.12 workflows. ([#14271](https://github.com/NVIDIA/cudf/pull/14271)) [@bdice](https://github.com/bdice) +- Refactor LogicalType for Parquet ([#14264](https://github.com/NVIDIA/cudf/pull/14264)) [@etseidl](https://github.com/etseidl) +- Centralize chunked reading code in the parquet reader to reader_impl_chunking.cu ([#14262](https://github.com/NVIDIA/cudf/pull/14262)) [@nvdbaranec](https://github.com/nvdbaranec) +- Expose stream parameter in public strings replace APIs ([#14261](https://github.com/NVIDIA/cudf/pull/14261)) [@davidwendt](https://github.com/davidwendt) +- Expose stream parameter in public strings APIs ([#14260](https://github.com/NVIDIA/cudf/pull/14260)) [@davidwendt](https://github.com/davidwendt) +- Cleanup of namespaces in parquet code. ([#14259](https://github.com/NVIDIA/cudf/pull/14259)) [@nvdbaranec](https://github.com/nvdbaranec) +- Make parquet schema index type consistent ([#14256](https://github.com/NVIDIA/cudf/pull/14256)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Expose stream parameter in public strings convert APIs ([#14255](https://github.com/NVIDIA/cudf/pull/14255)) [@davidwendt](https://github.com/davidwendt) +- Add in java bindings for DataSource ([#14254](https://github.com/NVIDIA/cudf/pull/14254)) [@revans2](https://github.com/revans2) +- Reimplement `cudf::merge` for nested types without using comparators ([#14250](https://github.com/NVIDIA/cudf/pull/14250)) [@divyegala](https://github.com/divyegala) +- Add stream parameter to List Manipulation and Operations APIs ([#14248](https://github.com/NVIDIA/cudf/pull/14248)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Expose stream parameter in public strings split/partition APIs ([#14247](https://github.com/NVIDIA/cudf/pull/14247)) [@davidwendt](https://github.com/davidwendt) +- Improve `contains_column` by invoking `contains_table` ([#14238](https://github.com/NVIDIA/cudf/pull/14238)) [@PointKernel](https://github.com/PointKernel) +- Detect and report errors in Parquet header parsing ([#14237](https://github.com/NVIDIA/cudf/pull/14237)) [@etseidl](https://github.com/etseidl) +- Normalizing offsets iterator ([#14234](https://github.com/NVIDIA/cudf/pull/14234)) [@davidwendt](https://github.com/davidwendt) +- Forward merge `23.10` into `23.12` ([#14231](https://github.com/NVIDIA/cudf/pull/14231)) [@galipremsagar](https://github.com/galipremsagar) +- Return error if BOOL8 column-type is used with integers-to-hex ([#14208](https://github.com/NVIDIA/cudf/pull/14208)) [@davidwendt](https://github.com/davidwendt) +- Enable indexalator for device code ([#14206](https://github.com/NVIDIA/cudf/pull/14206)) [@davidwendt](https://github.com/davidwendt) +- Marginally reduce memory footprint of joins ([#14197](https://github.com/NVIDIA/cudf/pull/14197)) [@wence-](https://github.com/wence-) +- Add nvtx annotations to spilling-based data movement ([#14196](https://github.com/NVIDIA/cudf/pull/14196)) [@wence-](https://github.com/wence-) +- Optimize ORC writer for decimal columns ([#14190](https://github.com/NVIDIA/cudf/pull/14190)) [@vuule](https://github.com/vuule) +- Remove the use of volatile in ORC ([#14175](https://github.com/NVIDIA/cudf/pull/14175)) [@vuule](https://github.com/vuule) +- Add `bytes_per_second` to distinct_count of stream_compaction nvbench. ([#14172](https://github.com/NVIDIA/cudf/pull/14172)) [@Blonck](https://github.com/Blonck) +- Add `bytes_per_second` to transpose benchmark ([#14170](https://github.com/NVIDIA/cudf/pull/14170)) [@Blonck](https://github.com/Blonck) +- cuDF: Build CUDA 12.0 ARM conda packages. ([#14112](https://github.com/NVIDIA/cudf/pull/14112)) [@bdice](https://github.com/bdice) +- Add `bytes_per_second` to shift benchmark ([#13950](https://github.com/NVIDIA/cudf/pull/13950)) [@Blonck](https://github.com/Blonck) +- Extract `debug_utilities.hpp/cu` from `column_utilities.hpp/cu` ([#13720](https://github.com/NVIDIA/cudf/pull/13720)) [@ttnghia](https://github.com/ttnghia) # cuDF 23.10.00 (11 Oct 2023) ## 🚨 Breaking Changes -- Expose stream parameter in public nvtext ngram APIs ([#14061](https://github.com/rapidsai/cudf/pull/14061)) [@davidwendt](https://github.com/davidwendt) -- Raise `MixedTypeError` when a column of mixed-dtype is being constructed ([#14050](https://github.com/rapidsai/cudf/pull/14050)) [@galipremsagar](https://github.com/galipremsagar) -- Raise `NotImplementedError` for `MultiIndex.to_series` ([#14049](https://github.com/rapidsai/cudf/pull/14049)) [@galipremsagar](https://github.com/galipremsagar) -- Create table_input_metadata from a table_metadata ([#13920](https://github.com/rapidsai/cudf/pull/13920)) [@etseidl](https://github.com/etseidl) -- Enable RLE boolean encoding for v2 Parquet files ([#13886](https://github.com/rapidsai/cudf/pull/13886)) [@etseidl](https://github.com/etseidl) -- Change `NA` to `NaT` for `datetime` and `timedelta` types ([#13868](https://github.com/rapidsai/cudf/pull/13868)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `any`, `all` reduction behavior for `axis=None` and warn for other reductions ([#13831](https://github.com/rapidsai/cudf/pull/13831)) [@galipremsagar](https://github.com/galipremsagar) -- Add minhash support for MurmurHash3_x64_128 ([#13796](https://github.com/rapidsai/cudf/pull/13796)) [@davidwendt](https://github.com/davidwendt) -- Remove the libcudf cudf::offset_type type ([#13788](https://github.com/rapidsai/cudf/pull/13788)) [@davidwendt](https://github.com/davidwendt) -- Raise error when trying to join `datetime` and `timedelta` types with other types ([#13786](https://github.com/rapidsai/cudf/pull/13786)) [@galipremsagar](https://github.com/galipremsagar) -- Update to Cython 3.0.0 ([#13777](https://github.com/rapidsai/cudf/pull/13777)) [@vyasr](https://github.com/vyasr) -- Raise error on constructing an array from mixed type inputs ([#13768](https://github.com/rapidsai/cudf/pull/13768)) [@galipremsagar](https://github.com/galipremsagar) -- Enforce deprecations in `23.10` ([#13732](https://github.com/rapidsai/cudf/pull/13732)) [@galipremsagar](https://github.com/galipremsagar) -- Upgrade to arrow 12 ([#13728](https://github.com/rapidsai/cudf/pull/13728)) [@galipremsagar](https://github.com/galipremsagar) -- Remove Arrow dependency from the `datasource.hpp` public header ([#13698](https://github.com/rapidsai/cudf/pull/13698)) [@vuule](https://github.com/vuule) +- Expose stream parameter in public nvtext ngram APIs ([#14061](https://github.com/NVIDIA/cudf/pull/14061)) [@davidwendt](https://github.com/davidwendt) +- Raise `MixedTypeError` when a column of mixed-dtype is being constructed ([#14050](https://github.com/NVIDIA/cudf/pull/14050)) [@galipremsagar](https://github.com/galipremsagar) +- Raise `NotImplementedError` for `MultiIndex.to_series` ([#14049](https://github.com/NVIDIA/cudf/pull/14049)) [@galipremsagar](https://github.com/galipremsagar) +- Create table_input_metadata from a table_metadata ([#13920](https://github.com/NVIDIA/cudf/pull/13920)) [@etseidl](https://github.com/etseidl) +- Enable RLE boolean encoding for v2 Parquet files ([#13886](https://github.com/NVIDIA/cudf/pull/13886)) [@etseidl](https://github.com/etseidl) +- Change `NA` to `NaT` for `datetime` and `timedelta` types ([#13868](https://github.com/NVIDIA/cudf/pull/13868)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `any`, `all` reduction behavior for `axis=None` and warn for other reductions ([#13831](https://github.com/NVIDIA/cudf/pull/13831)) [@galipremsagar](https://github.com/galipremsagar) +- Add minhash support for MurmurHash3_x64_128 ([#13796](https://github.com/NVIDIA/cudf/pull/13796)) [@davidwendt](https://github.com/davidwendt) +- Remove the libcudf cudf::offset_type type ([#13788](https://github.com/NVIDIA/cudf/pull/13788)) [@davidwendt](https://github.com/davidwendt) +- Raise error when trying to join `datetime` and `timedelta` types with other types ([#13786](https://github.com/NVIDIA/cudf/pull/13786)) [@galipremsagar](https://github.com/galipremsagar) +- Update to Cython 3.0.0 ([#13777](https://github.com/NVIDIA/cudf/pull/13777)) [@vyasr](https://github.com/vyasr) +- Raise error on constructing an array from mixed type inputs ([#13768](https://github.com/NVIDIA/cudf/pull/13768)) [@galipremsagar](https://github.com/galipremsagar) +- Enforce deprecations in `23.10` ([#13732](https://github.com/NVIDIA/cudf/pull/13732)) [@galipremsagar](https://github.com/galipremsagar) +- Upgrade to arrow 12 ([#13728](https://github.com/NVIDIA/cudf/pull/13728)) [@galipremsagar](https://github.com/galipremsagar) +- Remove Arrow dependency from the `datasource.hpp` public header ([#13698](https://github.com/NVIDIA/cudf/pull/13698)) [@vuule](https://github.com/vuule) ## 🐛 Bug Fixes -- Fix inaccurate ceil/floor and inaccurate rescaling casts of fixed-point values. ([#14242](https://github.com/rapidsai/cudf/pull/14242)) [@bdice](https://github.com/bdice) -- Fix inaccuracy in decimal128 rounding. ([#14233](https://github.com/rapidsai/cudf/pull/14233)) [@bdice](https://github.com/bdice) -- Workaround for illegal instruction error in sm90 for warp instrinsics with mask ([#14201](https://github.com/rapidsai/cudf/pull/14201)) [@karthikeyann](https://github.com/karthikeyann) -- Fix pytorch related pytest ([#14198](https://github.com/rapidsai/cudf/pull/14198)) [@galipremsagar](https://github.com/galipremsagar) -- Pin to `aws-sdk-cpp<1.11` ([#14173](https://github.com/rapidsai/cudf/pull/14173)) [@pentschev](https://github.com/pentschev) -- Fix assert failure for range window functions ([#14168](https://github.com/rapidsai/cudf/pull/14168)) [@mythrocks](https://github.com/mythrocks) -- Fix Memcheck error found in JSON_TEST JsonReaderTest.ErrorStrings ([#14164](https://github.com/rapidsai/cudf/pull/14164)) [@karthikeyann](https://github.com/karthikeyann) -- Fix calls to copy_bitmask to pass stream parameter ([#14158](https://github.com/rapidsai/cudf/pull/14158)) [@davidwendt](https://github.com/davidwendt) -- Fix DataFrame from Series with different CategoricalIndexes ([#14157](https://github.com/rapidsai/cudf/pull/14157)) [@mroeschke](https://github.com/mroeschke) -- Pin to numpy<1.25 and numba<0.58 to avoid errors and deprecation warnings-as-errors. ([#14156](https://github.com/rapidsai/cudf/pull/14156)) [@bdice](https://github.com/bdice) -- Fix kernel launch error for cudf::io::orc::gpu::rowgroup_char_counts_kernel ([#14139](https://github.com/rapidsai/cudf/pull/14139)) [@davidwendt](https://github.com/davidwendt) -- Don't sort columns for DataFrame init from list of Series ([#14136](https://github.com/rapidsai/cudf/pull/14136)) [@mroeschke](https://github.com/mroeschke) -- Fix DataFrame.values with no columns but index ([#14134](https://github.com/rapidsai/cudf/pull/14134)) [@mroeschke](https://github.com/mroeschke) -- Avoid circular cimports in _lib/cpp/reduce.pxd ([#14125](https://github.com/rapidsai/cudf/pull/14125)) [@vyasr](https://github.com/vyasr) -- Add support for nested dict in `DataFrame` constructor ([#14119](https://github.com/rapidsai/cudf/pull/14119)) [@galipremsagar](https://github.com/galipremsagar) -- Restrict iterables of `DataFrame`'s as input to `DataFrame` constructor ([#14118](https://github.com/rapidsai/cudf/pull/14118)) [@galipremsagar](https://github.com/galipremsagar) -- Allow `numeric_only=True` for reduction operations on numeric types ([#14111](https://github.com/rapidsai/cudf/pull/14111)) [@galipremsagar](https://github.com/galipremsagar) -- Preserve name of the column while initializing a `DataFrame` ([#14110](https://github.com/rapidsai/cudf/pull/14110)) [@galipremsagar](https://github.com/galipremsagar) -- Correct numerous 20054-D: dynamic initialization errors found on arm+12.2 ([#14108](https://github.com/rapidsai/cudf/pull/14108)) [@robertmaynard](https://github.com/robertmaynard) -- Drop `kwargs` from `Series.count` ([#14106](https://github.com/rapidsai/cudf/pull/14106)) [@galipremsagar](https://github.com/galipremsagar) -- Fix naming issues with `Index.to_frame` and `MultiIndex.to_frame` APIs ([#14105](https://github.com/rapidsai/cudf/pull/14105)) [@galipremsagar](https://github.com/galipremsagar) -- Only use memory resources that haven't been freed ([#14103](https://github.com/rapidsai/cudf/pull/14103)) [@robertmaynard](https://github.com/robertmaynard) -- Add support for `__round__` in `Series` and `DataFrame` ([#14099](https://github.com/rapidsai/cudf/pull/14099)) [@galipremsagar](https://github.com/galipremsagar) -- Validate ignore_index type in drop_duplicates ([#14098](https://github.com/rapidsai/cudf/pull/14098)) [@mroeschke](https://github.com/mroeschke) -- Fix renaming `Series` and `Index` ([#14080](https://github.com/rapidsai/cudf/pull/14080)) [@galipremsagar](https://github.com/galipremsagar) -- Raise NotImplementedError in to_datetime if Z (or tz component) in string ([#14074](https://github.com/rapidsai/cudf/pull/14074)) [@mroeschke](https://github.com/mroeschke) -- Raise NotImplementedError for datetime strings with UTC offset ([#14070](https://github.com/rapidsai/cudf/pull/14070)) [@mroeschke](https://github.com/mroeschke) -- Update pyarrow-related dispatch logic in dask_cudf ([#14069](https://github.com/rapidsai/cudf/pull/14069)) [@rjzamora](https://github.com/rjzamora) -- Use `conda mambabuild` rather than `mamba mambabuild` ([#14067](https://github.com/rapidsai/cudf/pull/14067)) [@wence-](https://github.com/wence-) -- Raise NotImplementedError in to_datetime with dayfirst without infer_format ([#14058](https://github.com/rapidsai/cudf/pull/14058)) [@mroeschke](https://github.com/mroeschke) -- Fix various issues in `Index.intersection` ([#14054](https://github.com/rapidsai/cudf/pull/14054)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `Index.difference` to match with pandas ([#14053](https://github.com/rapidsai/cudf/pull/14053)) [@galipremsagar](https://github.com/galipremsagar) -- Fix empty string column construction ([#14052](https://github.com/rapidsai/cudf/pull/14052)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `IntervalIndex.union` to preserve type-metadata ([#14051](https://github.com/rapidsai/cudf/pull/14051)) [@galipremsagar](https://github.com/galipremsagar) -- Raise `MixedTypeError` when a column of mixed-dtype is being constructed ([#14050](https://github.com/rapidsai/cudf/pull/14050)) [@galipremsagar](https://github.com/galipremsagar) -- Raise `NotImplementedError` for `MultiIndex.to_series` ([#14049](https://github.com/rapidsai/cudf/pull/14049)) [@galipremsagar](https://github.com/galipremsagar) -- Ignore compile_commands.json ([#14048](https://github.com/rapidsai/cudf/pull/14048)) [@harrism](https://github.com/harrism) -- Raise TypeError for any non-parseable argument in to_datetime ([#14044](https://github.com/rapidsai/cudf/pull/14044)) [@mroeschke](https://github.com/mroeschke) -- Raise NotImplementedError for to_datetime with z format ([#14037](https://github.com/rapidsai/cudf/pull/14037)) [@mroeschke](https://github.com/mroeschke) -- Implement `sort_remaining` for `sort_index` ([#14033](https://github.com/rapidsai/cudf/pull/14033)) [@wence-](https://github.com/wence-) -- Raise NotImplementedError for Categoricals with timezones ([#14032](https://github.com/rapidsai/cudf/pull/14032)) [@mroeschke](https://github.com/mroeschke) -- Temporary fix Parquet metadata with empty value string being ignored from writing ([#14026](https://github.com/rapidsai/cudf/pull/14026)) [@ttnghia](https://github.com/ttnghia) -- Preserve types of scalar being returned when possible in `quantile` ([#14014](https://github.com/rapidsai/cudf/pull/14014)) [@galipremsagar](https://github.com/galipremsagar) -- Fix return type of `MultiIndex.difference` ([#14009](https://github.com/rapidsai/cudf/pull/14009)) [@galipremsagar](https://github.com/galipremsagar) -- Raise an error when timezone subtypes are encountered in `pd.IntervalDtype` ([#14006](https://github.com/rapidsai/cudf/pull/14006)) [@galipremsagar](https://github.com/galipremsagar) -- Fix map column can not be non-nullable for java ([#14003](https://github.com/rapidsai/cudf/pull/14003)) [@res-life](https://github.com/res-life) -- Fix `name` selection in `Index.difference` and `Index.intersection` ([#13986](https://github.com/rapidsai/cudf/pull/13986)) [@galipremsagar](https://github.com/galipremsagar) -- Restore column type metadata with `dropna` to fix `factorize` API ([#13980](https://github.com/rapidsai/cudf/pull/13980)) [@galipremsagar](https://github.com/galipremsagar) -- Use thread_index_type to avoid out of bounds accesses in conditional joins ([#13971](https://github.com/rapidsai/cudf/pull/13971)) [@vyasr](https://github.com/vyasr) -- Fix `MultiIndex.to_numpy` to return numpy array with tuples ([#13966](https://github.com/rapidsai/cudf/pull/13966)) [@galipremsagar](https://github.com/galipremsagar) -- Use cudf::thread_index_type in get_json_object and tdigest kernels ([#13962](https://github.com/rapidsai/cudf/pull/13962)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix an issue with `IntervalIndex.repr` when null values are present ([#13958](https://github.com/rapidsai/cudf/pull/13958)) [@galipremsagar](https://github.com/galipremsagar) -- Fix type metadata issue preservation with `Column.unique` ([#13957](https://github.com/rapidsai/cudf/pull/13957)) [@galipremsagar](https://github.com/galipremsagar) -- Handle `Interval` scalars when passed in list-like inputs to `cudf.Index` ([#13956](https://github.com/rapidsai/cudf/pull/13956)) [@galipremsagar](https://github.com/galipremsagar) -- Fix setting of categories order when `dtype` is passed to a `CategoricalColumn` ([#13955](https://github.com/rapidsai/cudf/pull/13955)) [@galipremsagar](https://github.com/galipremsagar) -- Handle `as_index` in `GroupBy.apply` ([#13951](https://github.com/rapidsai/cudf/pull/13951)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Raise error for string types in `nsmallest` and `nlargest` ([#13946](https://github.com/rapidsai/cudf/pull/13946)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `index` of `Groupby.apply` results when it is performed on empty objects ([#13944](https://github.com/rapidsai/cudf/pull/13944)) [@galipremsagar](https://github.com/galipremsagar) -- Fix integer overflow in shim `device_sum` functions ([#13943](https://github.com/rapidsai/cudf/pull/13943)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix type mismatch in groupby reduction for empty objects ([#13942](https://github.com/rapidsai/cudf/pull/13942)) [@galipremsagar](https://github.com/galipremsagar) -- Fixed processed bytes calculation in APPLY_BOOLEAN_MASK benchmark. ([#13937](https://github.com/rapidsai/cudf/pull/13937)) [@Blonck](https://github.com/Blonck) -- Fix construction of `Grouping` objects ([#13932](https://github.com/rapidsai/cudf/pull/13932)) [@galipremsagar](https://github.com/galipremsagar) -- Fix an issue with `loc` when column names is `MultiIndex` ([#13929](https://github.com/rapidsai/cudf/pull/13929)) [@galipremsagar](https://github.com/galipremsagar) -- Fix handling of typecasting in `searchsorted` ([#13925](https://github.com/rapidsai/cudf/pull/13925)) [@galipremsagar](https://github.com/galipremsagar) -- Preserve index `name` in `reindex` ([#13917](https://github.com/rapidsai/cudf/pull/13917)) [@galipremsagar](https://github.com/galipremsagar) -- Use `cudf::thread_index_type` in cuIO to prevent overflow in row indexing ([#13910](https://github.com/rapidsai/cudf/pull/13910)) [@vuule](https://github.com/vuule) -- Fix for encodings listed in the Parquet column chunk metadata ([#13907](https://github.com/rapidsai/cudf/pull/13907)) [@etseidl](https://github.com/etseidl) -- Use cudf::thread_index_type in concatenate.cu. ([#13906](https://github.com/rapidsai/cudf/pull/13906)) [@bdice](https://github.com/bdice) -- Use cudf::thread_index_type in replace.cu. ([#13905](https://github.com/rapidsai/cudf/pull/13905)) [@bdice](https://github.com/bdice) -- Add noSanitizer tag to Java reduction tests failing with sanitizer in CUDA 12 ([#13904](https://github.com/rapidsai/cudf/pull/13904)) [@jlowe](https://github.com/jlowe) -- Remove the internal use of the cudf's default stream in cuIO ([#13903](https://github.com/rapidsai/cudf/pull/13903)) [@vuule](https://github.com/vuule) -- Use cuda-nvtx-dev CUDA 12 package. ([#13901](https://github.com/rapidsai/cudf/pull/13901)) [@bdice](https://github.com/bdice) -- Use `thread_index_type` to avoid index overflow in grid-stride loops ([#13895](https://github.com/rapidsai/cudf/pull/13895)) [@PointKernel](https://github.com/PointKernel) -- Fix memory access error in cudf::shift for sliced strings ([#13894](https://github.com/rapidsai/cudf/pull/13894)) [@davidwendt](https://github.com/davidwendt) -- Raise error when trying to construct a `DataFrame` with mixed types ([#13889](https://github.com/rapidsai/cudf/pull/13889)) [@galipremsagar](https://github.com/galipremsagar) -- Return `nan` when one variable to be correlated has zero variance in JIT GroupBy Apply ([#13884](https://github.com/rapidsai/cudf/pull/13884)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Correctly detect the BOM mark in `read_csv` with compressed input ([#13881](https://github.com/rapidsai/cudf/pull/13881)) [@vuule](https://github.com/vuule) -- Check for the presence of all values in `MultiIndex.isin` ([#13879](https://github.com/rapidsai/cudf/pull/13879)) [@galipremsagar](https://github.com/galipremsagar) -- Fix nvtext::generate_character_ngrams performance regression for longer strings ([#13874](https://github.com/rapidsai/cudf/pull/13874)) [@davidwendt](https://github.com/davidwendt) -- Fix return type of `MultiIndex.levels` ([#13870](https://github.com/rapidsai/cudf/pull/13870)) [@galipremsagar](https://github.com/galipremsagar) -- Fix List's missing children metadata in JSON writer ([#13869](https://github.com/rapidsai/cudf/pull/13869)) [@karthikeyann](https://github.com/karthikeyann) -- Disable construction of Index when `freq` is set in pandas-compatibility mode ([#13857](https://github.com/rapidsai/cudf/pull/13857)) [@galipremsagar](https://github.com/galipremsagar) -- Fix an issue with fetching `NA` from a `TimedeltaColumn` ([#13853](https://github.com/rapidsai/cudf/pull/13853)) [@galipremsagar](https://github.com/galipremsagar) -- Simplify implementation of interval_range() and fix behaviour for floating `freq` ([#13844](https://github.com/rapidsai/cudf/pull/13844)) [@shwina](https://github.com/shwina) -- Fix binary operations between `Series` and `Index` ([#13842](https://github.com/rapidsai/cudf/pull/13842)) [@galipremsagar](https://github.com/galipremsagar) -- Update make_lists_column_from_scalar to use make_offsets_child_column utility ([#13841](https://github.com/rapidsai/cudf/pull/13841)) [@davidwendt](https://github.com/davidwendt) -- Fix read out of bounds in string concatenate ([#13838](https://github.com/rapidsai/cudf/pull/13838)) [@pentschev](https://github.com/pentschev) -- Raise error for more cases when `timezone-aware` data is passed to `as_column` ([#13835](https://github.com/rapidsai/cudf/pull/13835)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `any`, `all` reduction behavior for `axis=None` and warn for other reductions ([#13831](https://github.com/rapidsai/cudf/pull/13831)) [@galipremsagar](https://github.com/galipremsagar) -- Raise error when trying to construct time-zone aware timestamps ([#13830](https://github.com/rapidsai/cudf/pull/13830)) [@galipremsagar](https://github.com/galipremsagar) -- Fix cuFile I/O factories ([#13829](https://github.com/rapidsai/cudf/pull/13829)) [@vuule](https://github.com/vuule) -- DataFrame with namedtuples uses ._field as column names ([#13824](https://github.com/rapidsai/cudf/pull/13824)) [@mroeschke](https://github.com/mroeschke) -- Branch 23.10 merge 23.08 ([#13822](https://github.com/rapidsai/cudf/pull/13822)) [@vyasr](https://github.com/vyasr) -- Return a Series from JIT GroupBy apply, rather than a DataFrame ([#13820](https://github.com/rapidsai/cudf/pull/13820)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- No need to dlsym EnsureS3Finalized we can call it directly ([#13819](https://github.com/rapidsai/cudf/pull/13819)) [@robertmaynard](https://github.com/robertmaynard) -- Raise error when mixed types are being constructed ([#13816](https://github.com/rapidsai/cudf/pull/13816)) [@galipremsagar](https://github.com/galipremsagar) -- Fix unbounded sequence issue in `DataFrame` constructor ([#13811](https://github.com/rapidsai/cudf/pull/13811)) [@galipremsagar](https://github.com/galipremsagar) -- Fix Byte-Pair-Encoding usage of cuco static-map for storing merge-pairs ([#13807](https://github.com/rapidsai/cudf/pull/13807)) [@davidwendt](https://github.com/davidwendt) -- Fix for Parquet writer when requested pages per row is smaller than fragment size ([#13806](https://github.com/rapidsai/cudf/pull/13806)) [@etseidl](https://github.com/etseidl) -- Remove hangs from trying to construct un-bounded sequences ([#13799](https://github.com/rapidsai/cudf/pull/13799)) [@galipremsagar](https://github.com/galipremsagar) -- Bug/update libcudf to handle arrow12 changes ([#13794](https://github.com/rapidsai/cudf/pull/13794)) [@robertmaynard](https://github.com/robertmaynard) -- Update get_arrow to arrows 12 CMake target name of arrow::xsimd ([#13790](https://github.com/rapidsai/cudf/pull/13790)) [@robertmaynard](https://github.com/robertmaynard) -- Raise error when trying to join `datetime` and `timedelta` types with other types ([#13786](https://github.com/rapidsai/cudf/pull/13786)) [@galipremsagar](https://github.com/galipremsagar) -- Fix negative unary operation for boolean type ([#13780](https://github.com/rapidsai/cudf/pull/13780)) [@galipremsagar](https://github.com/galipremsagar) -- Fix contains(`in`) method for `Series` ([#13779](https://github.com/rapidsai/cudf/pull/13779)) [@galipremsagar](https://github.com/galipremsagar) -- Fix binary operation column ordering and missing column issues ([#13778](https://github.com/rapidsai/cudf/pull/13778)) [@galipremsagar](https://github.com/galipremsagar) -- Cast only time of day to nanos to avoid an overflow in Parquet INT96 write ([#13776](https://github.com/rapidsai/cudf/pull/13776)) [@gerashegalov](https://github.com/gerashegalov) -- Preserve names of column object in various APIs ([#13772](https://github.com/rapidsai/cudf/pull/13772)) [@galipremsagar](https://github.com/galipremsagar) -- Raise error on constructing an array from mixed type inputs ([#13768](https://github.com/rapidsai/cudf/pull/13768)) [@galipremsagar](https://github.com/galipremsagar) -- Fix construction of DataFrames from dict when columns are provided ([#13766](https://github.com/rapidsai/cudf/pull/13766)) [@wence-](https://github.com/wence-) -- Provide our own Cython declaration for make_unique ([#13746](https://github.com/rapidsai/cudf/pull/13746)) [@wence-](https://github.com/wence-) +- Fix inaccurate ceil/floor and inaccurate rescaling casts of fixed-point values. ([#14242](https://github.com/NVIDIA/cudf/pull/14242)) [@bdice](https://github.com/bdice) +- Fix inaccuracy in decimal128 rounding. ([#14233](https://github.com/NVIDIA/cudf/pull/14233)) [@bdice](https://github.com/bdice) +- Workaround for illegal instruction error in sm90 for warp instrinsics with mask ([#14201](https://github.com/NVIDIA/cudf/pull/14201)) [@karthikeyann](https://github.com/karthikeyann) +- Fix pytorch related pytest ([#14198](https://github.com/NVIDIA/cudf/pull/14198)) [@galipremsagar](https://github.com/galipremsagar) +- Pin to `aws-sdk-cpp<1.11` ([#14173](https://github.com/NVIDIA/cudf/pull/14173)) [@pentschev](https://github.com/pentschev) +- Fix assert failure for range window functions ([#14168](https://github.com/NVIDIA/cudf/pull/14168)) [@mythrocks](https://github.com/mythrocks) +- Fix Memcheck error found in JSON_TEST JsonReaderTest.ErrorStrings ([#14164](https://github.com/NVIDIA/cudf/pull/14164)) [@karthikeyann](https://github.com/karthikeyann) +- Fix calls to copy_bitmask to pass stream parameter ([#14158](https://github.com/NVIDIA/cudf/pull/14158)) [@davidwendt](https://github.com/davidwendt) +- Fix DataFrame from Series with different CategoricalIndexes ([#14157](https://github.com/NVIDIA/cudf/pull/14157)) [@mroeschke](https://github.com/mroeschke) +- Pin to numpy<1.25 and numba<0.58 to avoid errors and deprecation warnings-as-errors. ([#14156](https://github.com/NVIDIA/cudf/pull/14156)) [@bdice](https://github.com/bdice) +- Fix kernel launch error for cudf::io::orc::gpu::rowgroup_char_counts_kernel ([#14139](https://github.com/NVIDIA/cudf/pull/14139)) [@davidwendt](https://github.com/davidwendt) +- Don't sort columns for DataFrame init from list of Series ([#14136](https://github.com/NVIDIA/cudf/pull/14136)) [@mroeschke](https://github.com/mroeschke) +- Fix DataFrame.values with no columns but index ([#14134](https://github.com/NVIDIA/cudf/pull/14134)) [@mroeschke](https://github.com/mroeschke) +- Avoid circular cimports in _lib/cpp/reduce.pxd ([#14125](https://github.com/NVIDIA/cudf/pull/14125)) [@vyasr](https://github.com/vyasr) +- Add support for nested dict in `DataFrame` constructor ([#14119](https://github.com/NVIDIA/cudf/pull/14119)) [@galipremsagar](https://github.com/galipremsagar) +- Restrict iterables of `DataFrame`'s as input to `DataFrame` constructor ([#14118](https://github.com/NVIDIA/cudf/pull/14118)) [@galipremsagar](https://github.com/galipremsagar) +- Allow `numeric_only=True` for reduction operations on numeric types ([#14111](https://github.com/NVIDIA/cudf/pull/14111)) [@galipremsagar](https://github.com/galipremsagar) +- Preserve name of the column while initializing a `DataFrame` ([#14110](https://github.com/NVIDIA/cudf/pull/14110)) [@galipremsagar](https://github.com/galipremsagar) +- Correct numerous 20054-D: dynamic initialization errors found on arm+12.2 ([#14108](https://github.com/NVIDIA/cudf/pull/14108)) [@robertmaynard](https://github.com/robertmaynard) +- Drop `kwargs` from `Series.count` ([#14106](https://github.com/NVIDIA/cudf/pull/14106)) [@galipremsagar](https://github.com/galipremsagar) +- Fix naming issues with `Index.to_frame` and `MultiIndex.to_frame` APIs ([#14105](https://github.com/NVIDIA/cudf/pull/14105)) [@galipremsagar](https://github.com/galipremsagar) +- Only use memory resources that haven't been freed ([#14103](https://github.com/NVIDIA/cudf/pull/14103)) [@robertmaynard](https://github.com/robertmaynard) +- Add support for `__round__` in `Series` and `DataFrame` ([#14099](https://github.com/NVIDIA/cudf/pull/14099)) [@galipremsagar](https://github.com/galipremsagar) +- Validate ignore_index type in drop_duplicates ([#14098](https://github.com/NVIDIA/cudf/pull/14098)) [@mroeschke](https://github.com/mroeschke) +- Fix renaming `Series` and `Index` ([#14080](https://github.com/NVIDIA/cudf/pull/14080)) [@galipremsagar](https://github.com/galipremsagar) +- Raise NotImplementedError in to_datetime if Z (or tz component) in string ([#14074](https://github.com/NVIDIA/cudf/pull/14074)) [@mroeschke](https://github.com/mroeschke) +- Raise NotImplementedError for datetime strings with UTC offset ([#14070](https://github.com/NVIDIA/cudf/pull/14070)) [@mroeschke](https://github.com/mroeschke) +- Update pyarrow-related dispatch logic in dask_cudf ([#14069](https://github.com/NVIDIA/cudf/pull/14069)) [@rjzamora](https://github.com/rjzamora) +- Use `conda mambabuild` rather than `mamba mambabuild` ([#14067](https://github.com/NVIDIA/cudf/pull/14067)) [@wence-](https://github.com/wence-) +- Raise NotImplementedError in to_datetime with dayfirst without infer_format ([#14058](https://github.com/NVIDIA/cudf/pull/14058)) [@mroeschke](https://github.com/mroeschke) +- Fix various issues in `Index.intersection` ([#14054](https://github.com/NVIDIA/cudf/pull/14054)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `Index.difference` to match with pandas ([#14053](https://github.com/NVIDIA/cudf/pull/14053)) [@galipremsagar](https://github.com/galipremsagar) +- Fix empty string column construction ([#14052](https://github.com/NVIDIA/cudf/pull/14052)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `IntervalIndex.union` to preserve type-metadata ([#14051](https://github.com/NVIDIA/cudf/pull/14051)) [@galipremsagar](https://github.com/galipremsagar) +- Raise `MixedTypeError` when a column of mixed-dtype is being constructed ([#14050](https://github.com/NVIDIA/cudf/pull/14050)) [@galipremsagar](https://github.com/galipremsagar) +- Raise `NotImplementedError` for `MultiIndex.to_series` ([#14049](https://github.com/NVIDIA/cudf/pull/14049)) [@galipremsagar](https://github.com/galipremsagar) +- Ignore compile_commands.json ([#14048](https://github.com/NVIDIA/cudf/pull/14048)) [@harrism](https://github.com/harrism) +- Raise TypeError for any non-parseable argument in to_datetime ([#14044](https://github.com/NVIDIA/cudf/pull/14044)) [@mroeschke](https://github.com/mroeschke) +- Raise NotImplementedError for to_datetime with z format ([#14037](https://github.com/NVIDIA/cudf/pull/14037)) [@mroeschke](https://github.com/mroeschke) +- Implement `sort_remaining` for `sort_index` ([#14033](https://github.com/NVIDIA/cudf/pull/14033)) [@wence-](https://github.com/wence-) +- Raise NotImplementedError for Categoricals with timezones ([#14032](https://github.com/NVIDIA/cudf/pull/14032)) [@mroeschke](https://github.com/mroeschke) +- Temporary fix Parquet metadata with empty value string being ignored from writing ([#14026](https://github.com/NVIDIA/cudf/pull/14026)) [@ttnghia](https://github.com/ttnghia) +- Preserve types of scalar being returned when possible in `quantile` ([#14014](https://github.com/NVIDIA/cudf/pull/14014)) [@galipremsagar](https://github.com/galipremsagar) +- Fix return type of `MultiIndex.difference` ([#14009](https://github.com/NVIDIA/cudf/pull/14009)) [@galipremsagar](https://github.com/galipremsagar) +- Raise an error when timezone subtypes are encountered in `pd.IntervalDtype` ([#14006](https://github.com/NVIDIA/cudf/pull/14006)) [@galipremsagar](https://github.com/galipremsagar) +- Fix map column can not be non-nullable for java ([#14003](https://github.com/NVIDIA/cudf/pull/14003)) [@res-life](https://github.com/res-life) +- Fix `name` selection in `Index.difference` and `Index.intersection` ([#13986](https://github.com/NVIDIA/cudf/pull/13986)) [@galipremsagar](https://github.com/galipremsagar) +- Restore column type metadata with `dropna` to fix `factorize` API ([#13980](https://github.com/NVIDIA/cudf/pull/13980)) [@galipremsagar](https://github.com/galipremsagar) +- Use thread_index_type to avoid out of bounds accesses in conditional joins ([#13971](https://github.com/NVIDIA/cudf/pull/13971)) [@vyasr](https://github.com/vyasr) +- Fix `MultiIndex.to_numpy` to return numpy array with tuples ([#13966](https://github.com/NVIDIA/cudf/pull/13966)) [@galipremsagar](https://github.com/galipremsagar) +- Use cudf::thread_index_type in get_json_object and tdigest kernels ([#13962](https://github.com/NVIDIA/cudf/pull/13962)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix an issue with `IntervalIndex.repr` when null values are present ([#13958](https://github.com/NVIDIA/cudf/pull/13958)) [@galipremsagar](https://github.com/galipremsagar) +- Fix type metadata issue preservation with `Column.unique` ([#13957](https://github.com/NVIDIA/cudf/pull/13957)) [@galipremsagar](https://github.com/galipremsagar) +- Handle `Interval` scalars when passed in list-like inputs to `cudf.Index` ([#13956](https://github.com/NVIDIA/cudf/pull/13956)) [@galipremsagar](https://github.com/galipremsagar) +- Fix setting of categories order when `dtype` is passed to a `CategoricalColumn` ([#13955](https://github.com/NVIDIA/cudf/pull/13955)) [@galipremsagar](https://github.com/galipremsagar) +- Handle `as_index` in `GroupBy.apply` ([#13951](https://github.com/NVIDIA/cudf/pull/13951)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Raise error for string types in `nsmallest` and `nlargest` ([#13946](https://github.com/NVIDIA/cudf/pull/13946)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `index` of `Groupby.apply` results when it is performed on empty objects ([#13944](https://github.com/NVIDIA/cudf/pull/13944)) [@galipremsagar](https://github.com/galipremsagar) +- Fix integer overflow in shim `device_sum` functions ([#13943](https://github.com/NVIDIA/cudf/pull/13943)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix type mismatch in groupby reduction for empty objects ([#13942](https://github.com/NVIDIA/cudf/pull/13942)) [@galipremsagar](https://github.com/galipremsagar) +- Fixed processed bytes calculation in APPLY_BOOLEAN_MASK benchmark. ([#13937](https://github.com/NVIDIA/cudf/pull/13937)) [@Blonck](https://github.com/Blonck) +- Fix construction of `Grouping` objects ([#13932](https://github.com/NVIDIA/cudf/pull/13932)) [@galipremsagar](https://github.com/galipremsagar) +- Fix an issue with `loc` when column names is `MultiIndex` ([#13929](https://github.com/NVIDIA/cudf/pull/13929)) [@galipremsagar](https://github.com/galipremsagar) +- Fix handling of typecasting in `searchsorted` ([#13925](https://github.com/NVIDIA/cudf/pull/13925)) [@galipremsagar](https://github.com/galipremsagar) +- Preserve index `name` in `reindex` ([#13917](https://github.com/NVIDIA/cudf/pull/13917)) [@galipremsagar](https://github.com/galipremsagar) +- Use `cudf::thread_index_type` in cuIO to prevent overflow in row indexing ([#13910](https://github.com/NVIDIA/cudf/pull/13910)) [@vuule](https://github.com/vuule) +- Fix for encodings listed in the Parquet column chunk metadata ([#13907](https://github.com/NVIDIA/cudf/pull/13907)) [@etseidl](https://github.com/etseidl) +- Use cudf::thread_index_type in concatenate.cu. ([#13906](https://github.com/NVIDIA/cudf/pull/13906)) [@bdice](https://github.com/bdice) +- Use cudf::thread_index_type in replace.cu. ([#13905](https://github.com/NVIDIA/cudf/pull/13905)) [@bdice](https://github.com/bdice) +- Add noSanitizer tag to Java reduction tests failing with sanitizer in CUDA 12 ([#13904](https://github.com/NVIDIA/cudf/pull/13904)) [@jlowe](https://github.com/jlowe) +- Remove the internal use of the cudf's default stream in cuIO ([#13903](https://github.com/NVIDIA/cudf/pull/13903)) [@vuule](https://github.com/vuule) +- Use cuda-nvtx-dev CUDA 12 package. ([#13901](https://github.com/NVIDIA/cudf/pull/13901)) [@bdice](https://github.com/bdice) +- Use `thread_index_type` to avoid index overflow in grid-stride loops ([#13895](https://github.com/NVIDIA/cudf/pull/13895)) [@PointKernel](https://github.com/PointKernel) +- Fix memory access error in cudf::shift for sliced strings ([#13894](https://github.com/NVIDIA/cudf/pull/13894)) [@davidwendt](https://github.com/davidwendt) +- Raise error when trying to construct a `DataFrame` with mixed types ([#13889](https://github.com/NVIDIA/cudf/pull/13889)) [@galipremsagar](https://github.com/galipremsagar) +- Return `nan` when one variable to be correlated has zero variance in JIT GroupBy Apply ([#13884](https://github.com/NVIDIA/cudf/pull/13884)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Correctly detect the BOM mark in `read_csv` with compressed input ([#13881](https://github.com/NVIDIA/cudf/pull/13881)) [@vuule](https://github.com/vuule) +- Check for the presence of all values in `MultiIndex.isin` ([#13879](https://github.com/NVIDIA/cudf/pull/13879)) [@galipremsagar](https://github.com/galipremsagar) +- Fix nvtext::generate_character_ngrams performance regression for longer strings ([#13874](https://github.com/NVIDIA/cudf/pull/13874)) [@davidwendt](https://github.com/davidwendt) +- Fix return type of `MultiIndex.levels` ([#13870](https://github.com/NVIDIA/cudf/pull/13870)) [@galipremsagar](https://github.com/galipremsagar) +- Fix List's missing children metadata in JSON writer ([#13869](https://github.com/NVIDIA/cudf/pull/13869)) [@karthikeyann](https://github.com/karthikeyann) +- Disable construction of Index when `freq` is set in pandas-compatibility mode ([#13857](https://github.com/NVIDIA/cudf/pull/13857)) [@galipremsagar](https://github.com/galipremsagar) +- Fix an issue with fetching `NA` from a `TimedeltaColumn` ([#13853](https://github.com/NVIDIA/cudf/pull/13853)) [@galipremsagar](https://github.com/galipremsagar) +- Simplify implementation of interval_range() and fix behaviour for floating `freq` ([#13844](https://github.com/NVIDIA/cudf/pull/13844)) [@shwina](https://github.com/shwina) +- Fix binary operations between `Series` and `Index` ([#13842](https://github.com/NVIDIA/cudf/pull/13842)) [@galipremsagar](https://github.com/galipremsagar) +- Update make_lists_column_from_scalar to use make_offsets_child_column utility ([#13841](https://github.com/NVIDIA/cudf/pull/13841)) [@davidwendt](https://github.com/davidwendt) +- Fix read out of bounds in string concatenate ([#13838](https://github.com/NVIDIA/cudf/pull/13838)) [@pentschev](https://github.com/pentschev) +- Raise error for more cases when `timezone-aware` data is passed to `as_column` ([#13835](https://github.com/NVIDIA/cudf/pull/13835)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `any`, `all` reduction behavior for `axis=None` and warn for other reductions ([#13831](https://github.com/NVIDIA/cudf/pull/13831)) [@galipremsagar](https://github.com/galipremsagar) +- Raise error when trying to construct time-zone aware timestamps ([#13830](https://github.com/NVIDIA/cudf/pull/13830)) [@galipremsagar](https://github.com/galipremsagar) +- Fix cuFile I/O factories ([#13829](https://github.com/NVIDIA/cudf/pull/13829)) [@vuule](https://github.com/vuule) +- DataFrame with namedtuples uses ._field as column names ([#13824](https://github.com/NVIDIA/cudf/pull/13824)) [@mroeschke](https://github.com/mroeschke) +- Branch 23.10 merge 23.08 ([#13822](https://github.com/NVIDIA/cudf/pull/13822)) [@vyasr](https://github.com/vyasr) +- Return a Series from JIT GroupBy apply, rather than a DataFrame ([#13820](https://github.com/NVIDIA/cudf/pull/13820)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- No need to dlsym EnsureS3Finalized we can call it directly ([#13819](https://github.com/NVIDIA/cudf/pull/13819)) [@robertmaynard](https://github.com/robertmaynard) +- Raise error when mixed types are being constructed ([#13816](https://github.com/NVIDIA/cudf/pull/13816)) [@galipremsagar](https://github.com/galipremsagar) +- Fix unbounded sequence issue in `DataFrame` constructor ([#13811](https://github.com/NVIDIA/cudf/pull/13811)) [@galipremsagar](https://github.com/galipremsagar) +- Fix Byte-Pair-Encoding usage of cuco static-map for storing merge-pairs ([#13807](https://github.com/NVIDIA/cudf/pull/13807)) [@davidwendt](https://github.com/davidwendt) +- Fix for Parquet writer when requested pages per row is smaller than fragment size ([#13806](https://github.com/NVIDIA/cudf/pull/13806)) [@etseidl](https://github.com/etseidl) +- Remove hangs from trying to construct un-bounded sequences ([#13799](https://github.com/NVIDIA/cudf/pull/13799)) [@galipremsagar](https://github.com/galipremsagar) +- Bug/update libcudf to handle arrow12 changes ([#13794](https://github.com/NVIDIA/cudf/pull/13794)) [@robertmaynard](https://github.com/robertmaynard) +- Update get_arrow to arrows 12 CMake target name of arrow::xsimd ([#13790](https://github.com/NVIDIA/cudf/pull/13790)) [@robertmaynard](https://github.com/robertmaynard) +- Raise error when trying to join `datetime` and `timedelta` types with other types ([#13786](https://github.com/NVIDIA/cudf/pull/13786)) [@galipremsagar](https://github.com/galipremsagar) +- Fix negative unary operation for boolean type ([#13780](https://github.com/NVIDIA/cudf/pull/13780)) [@galipremsagar](https://github.com/galipremsagar) +- Fix contains(`in`) method for `Series` ([#13779](https://github.com/NVIDIA/cudf/pull/13779)) [@galipremsagar](https://github.com/galipremsagar) +- Fix binary operation column ordering and missing column issues ([#13778](https://github.com/NVIDIA/cudf/pull/13778)) [@galipremsagar](https://github.com/galipremsagar) +- Cast only time of day to nanos to avoid an overflow in Parquet INT96 write ([#13776](https://github.com/NVIDIA/cudf/pull/13776)) [@gerashegalov](https://github.com/gerashegalov) +- Preserve names of column object in various APIs ([#13772](https://github.com/NVIDIA/cudf/pull/13772)) [@galipremsagar](https://github.com/galipremsagar) +- Raise error on constructing an array from mixed type inputs ([#13768](https://github.com/NVIDIA/cudf/pull/13768)) [@galipremsagar](https://github.com/galipremsagar) +- Fix construction of DataFrames from dict when columns are provided ([#13766](https://github.com/NVIDIA/cudf/pull/13766)) [@wence-](https://github.com/wence-) +- Provide our own Cython declaration for make_unique ([#13746](https://github.com/NVIDIA/cudf/pull/13746)) [@wence-](https://github.com/wence-) ## 📖 Documentation -- Fix typo in docstring: metadata. ([#14025](https://github.com/rapidsai/cudf/pull/14025)) [@bdice](https://github.com/bdice) -- Fix typo in parquet/page_decode.cuh ([#13849](https://github.com/rapidsai/cudf/pull/13849)) [@XinyuZeng](https://github.com/XinyuZeng) -- Simplify Python doc configuration ([#13826](https://github.com/rapidsai/cudf/pull/13826)) [@vyasr](https://github.com/vyasr) -- Update documentation to reflect recent changes in JSON reader and writer ([#13791](https://github.com/rapidsai/cudf/pull/13791)) [@vuule](https://github.com/vuule) -- Fix all warnings in Python docs ([#13789](https://github.com/rapidsai/cudf/pull/13789)) [@vyasr](https://github.com/vyasr) +- Fix typo in docstring: metadata. ([#14025](https://github.com/NVIDIA/cudf/pull/14025)) [@bdice](https://github.com/bdice) +- Fix typo in parquet/page_decode.cuh ([#13849](https://github.com/NVIDIA/cudf/pull/13849)) [@XinyuZeng](https://github.com/XinyuZeng) +- Simplify Python doc configuration ([#13826](https://github.com/NVIDIA/cudf/pull/13826)) [@vyasr](https://github.com/vyasr) +- Update documentation to reflect recent changes in JSON reader and writer ([#13791](https://github.com/NVIDIA/cudf/pull/13791)) [@vuule](https://github.com/vuule) +- Fix all warnings in Python docs ([#13789](https://github.com/NVIDIA/cudf/pull/13789)) [@vyasr](https://github.com/vyasr) ## 🚀 New Features -- [Java] Add JNI bindings for `integers_to_hex` ([#14205](https://github.com/rapidsai/cudf/pull/14205)) [@razajafri](https://github.com/razajafri) -- Propagate errors from Parquet reader kernels back to host ([#14167](https://github.com/rapidsai/cudf/pull/14167)) [@vuule](https://github.com/vuule) -- JNI for `HISTOGRAM` and `MERGE_HISTOGRAM` aggregations ([#14154](https://github.com/rapidsai/cudf/pull/14154)) [@ttnghia](https://github.com/ttnghia) -- Expose streams in all public sorting APIs ([#14146](https://github.com/rapidsai/cudf/pull/14146)) [@vyasr](https://github.com/vyasr) -- Enable direct ingestion and production of Arrow scalars ([#14121](https://github.com/rapidsai/cudf/pull/14121)) [@vyasr](https://github.com/vyasr) -- Implement `GroupBy.value_counts` to match pandas API ([#14114](https://github.com/rapidsai/cudf/pull/14114)) [@stmio](https://github.com/stmio) -- Refactor parquet thrift reader ([#14097](https://github.com/rapidsai/cudf/pull/14097)) [@etseidl](https://github.com/etseidl) -- Refactor `hash_reduce_by_row` ([#14095](https://github.com/rapidsai/cudf/pull/14095)) [@ttnghia](https://github.com/ttnghia) -- Support negative preceding/following for ROW window functions ([#14093](https://github.com/rapidsai/cudf/pull/14093)) [@mythrocks](https://github.com/mythrocks) -- Support for progressive parquet chunked reading. ([#14079](https://github.com/rapidsai/cudf/pull/14079)) [@nvdbaranec](https://github.com/nvdbaranec) -- Implement `HISTOGRAM` and `MERGE_HISTOGRAM` aggregations ([#14045](https://github.com/rapidsai/cudf/pull/14045)) [@ttnghia](https://github.com/ttnghia) -- Expose streams in public search APIs ([#14034](https://github.com/rapidsai/cudf/pull/14034)) [@vyasr](https://github.com/vyasr) -- Expose streams in public replace APIs ([#14010](https://github.com/rapidsai/cudf/pull/14010)) [@vyasr](https://github.com/vyasr) -- Add stream parameter to public cudf::strings::split APIs ([#13997](https://github.com/rapidsai/cudf/pull/13997)) [@davidwendt](https://github.com/davidwendt) -- Expose streams in public filling APIs ([#13990](https://github.com/rapidsai/cudf/pull/13990)) [@vyasr](https://github.com/vyasr) -- Expose streams in public concatenate APIs ([#13987](https://github.com/rapidsai/cudf/pull/13987)) [@vyasr](https://github.com/vyasr) -- Use HostMemoryAllocator in jni::allocate_host_buffer ([#13975](https://github.com/rapidsai/cudf/pull/13975)) [@gerashegalov](https://github.com/gerashegalov) -- Enable fractional null probability for hashing benchmark ([#13967](https://github.com/rapidsai/cudf/pull/13967)) [@Blonck](https://github.com/Blonck) -- Switch pylibcudf-enabled types to use enum class in Cython ([#13931](https://github.com/rapidsai/cudf/pull/13931)) [@vyasr](https://github.com/vyasr) -- Add nvtext::tokenize_with_vocabulary API ([#13930](https://github.com/rapidsai/cudf/pull/13930)) [@davidwendt](https://github.com/davidwendt) -- Rewrite `DataFrame.stack` to support multi level column names ([#13927](https://github.com/rapidsai/cudf/pull/13927)) [@isVoid](https://github.com/isVoid) -- Add HostMemoryAllocator interface ([#13924](https://github.com/rapidsai/cudf/pull/13924)) [@gerashegalov](https://github.com/gerashegalov) -- Global stream pool ([#13922](https://github.com/rapidsai/cudf/pull/13922)) [@etseidl](https://github.com/etseidl) -- Create table_input_metadata from a table_metadata ([#13920](https://github.com/rapidsai/cudf/pull/13920)) [@etseidl](https://github.com/etseidl) -- Translate column size overflow exception to JNI ([#13911](https://github.com/rapidsai/cudf/pull/13911)) [@mythrocks](https://github.com/mythrocks) -- Enable RLE boolean encoding for v2 Parquet files ([#13886](https://github.com/rapidsai/cudf/pull/13886)) [@etseidl](https://github.com/etseidl) -- Exclude some tests from running with the compute sanitizer ([#13872](https://github.com/rapidsai/cudf/pull/13872)) [@firestarman](https://github.com/firestarman) -- Expand statistics support in ORC writer ([#13848](https://github.com/rapidsai/cudf/pull/13848)) [@vuule](https://github.com/vuule) -- Register the memory mapped buffer in `datasource` to improve H2D throughput ([#13814](https://github.com/rapidsai/cudf/pull/13814)) [@vuule](https://github.com/vuule) -- Add cudf::strings::find function with target per row ([#13808](https://github.com/rapidsai/cudf/pull/13808)) [@davidwendt](https://github.com/davidwendt) -- Add minhash support for MurmurHash3_x64_128 ([#13796](https://github.com/rapidsai/cudf/pull/13796)) [@davidwendt](https://github.com/davidwendt) -- Remove unnecessary pointer copying in JIT GroupBy Apply ([#13792](https://github.com/rapidsai/cudf/pull/13792)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add 'poll' function to custreamz kafka consumer ([#13782](https://github.com/rapidsai/cudf/pull/13782)) [@jdye64](https://github.com/jdye64) -- Support `corr` in `GroupBy.apply` through the jit engine ([#13767](https://github.com/rapidsai/cudf/pull/13767)) [@shwina](https://github.com/shwina) -- Optionally write version 2 page headers in Parquet writer ([#13751](https://github.com/rapidsai/cudf/pull/13751)) [@etseidl](https://github.com/etseidl) -- Support more numeric types in `Groupby.apply` with `engine='jit'` ([#13729](https://github.com/rapidsai/cudf/pull/13729)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- [FEA] Add DELTA_BINARY_PACKED decoding support to Parquet reader ([#13637](https://github.com/rapidsai/cudf/pull/13637)) [@etseidl](https://github.com/etseidl) -- Read FIXED_LEN_BYTE_ARRAY as binary in parquet reader ([#13437](https://github.com/rapidsai/cudf/pull/13437)) [@PointKernel](https://github.com/PointKernel) +- [Java] Add JNI bindings for `integers_to_hex` ([#14205](https://github.com/NVIDIA/cudf/pull/14205)) [@razajafri](https://github.com/razajafri) +- Propagate errors from Parquet reader kernels back to host ([#14167](https://github.com/NVIDIA/cudf/pull/14167)) [@vuule](https://github.com/vuule) +- JNI for `HISTOGRAM` and `MERGE_HISTOGRAM` aggregations ([#14154](https://github.com/NVIDIA/cudf/pull/14154)) [@ttnghia](https://github.com/ttnghia) +- Expose streams in all public sorting APIs ([#14146](https://github.com/NVIDIA/cudf/pull/14146)) [@vyasr](https://github.com/vyasr) +- Enable direct ingestion and production of Arrow scalars ([#14121](https://github.com/NVIDIA/cudf/pull/14121)) [@vyasr](https://github.com/vyasr) +- Implement `GroupBy.value_counts` to match pandas API ([#14114](https://github.com/NVIDIA/cudf/pull/14114)) [@stmio](https://github.com/stmio) +- Refactor parquet thrift reader ([#14097](https://github.com/NVIDIA/cudf/pull/14097)) [@etseidl](https://github.com/etseidl) +- Refactor `hash_reduce_by_row` ([#14095](https://github.com/NVIDIA/cudf/pull/14095)) [@ttnghia](https://github.com/ttnghia) +- Support negative preceding/following for ROW window functions ([#14093](https://github.com/NVIDIA/cudf/pull/14093)) [@mythrocks](https://github.com/mythrocks) +- Support for progressive parquet chunked reading. ([#14079](https://github.com/NVIDIA/cudf/pull/14079)) [@nvdbaranec](https://github.com/nvdbaranec) +- Implement `HISTOGRAM` and `MERGE_HISTOGRAM` aggregations ([#14045](https://github.com/NVIDIA/cudf/pull/14045)) [@ttnghia](https://github.com/ttnghia) +- Expose streams in public search APIs ([#14034](https://github.com/NVIDIA/cudf/pull/14034)) [@vyasr](https://github.com/vyasr) +- Expose streams in public replace APIs ([#14010](https://github.com/NVIDIA/cudf/pull/14010)) [@vyasr](https://github.com/vyasr) +- Add stream parameter to public cudf::strings::split APIs ([#13997](https://github.com/NVIDIA/cudf/pull/13997)) [@davidwendt](https://github.com/davidwendt) +- Expose streams in public filling APIs ([#13990](https://github.com/NVIDIA/cudf/pull/13990)) [@vyasr](https://github.com/vyasr) +- Expose streams in public concatenate APIs ([#13987](https://github.com/NVIDIA/cudf/pull/13987)) [@vyasr](https://github.com/vyasr) +- Use HostMemoryAllocator in jni::allocate_host_buffer ([#13975](https://github.com/NVIDIA/cudf/pull/13975)) [@gerashegalov](https://github.com/gerashegalov) +- Enable fractional null probability for hashing benchmark ([#13967](https://github.com/NVIDIA/cudf/pull/13967)) [@Blonck](https://github.com/Blonck) +- Switch pylibcudf-enabled types to use enum class in Cython ([#13931](https://github.com/NVIDIA/cudf/pull/13931)) [@vyasr](https://github.com/vyasr) +- Add nvtext::tokenize_with_vocabulary API ([#13930](https://github.com/NVIDIA/cudf/pull/13930)) [@davidwendt](https://github.com/davidwendt) +- Rewrite `DataFrame.stack` to support multi level column names ([#13927](https://github.com/NVIDIA/cudf/pull/13927)) [@isVoid](https://github.com/isVoid) +- Add HostMemoryAllocator interface ([#13924](https://github.com/NVIDIA/cudf/pull/13924)) [@gerashegalov](https://github.com/gerashegalov) +- Global stream pool ([#13922](https://github.com/NVIDIA/cudf/pull/13922)) [@etseidl](https://github.com/etseidl) +- Create table_input_metadata from a table_metadata ([#13920](https://github.com/NVIDIA/cudf/pull/13920)) [@etseidl](https://github.com/etseidl) +- Translate column size overflow exception to JNI ([#13911](https://github.com/NVIDIA/cudf/pull/13911)) [@mythrocks](https://github.com/mythrocks) +- Enable RLE boolean encoding for v2 Parquet files ([#13886](https://github.com/NVIDIA/cudf/pull/13886)) [@etseidl](https://github.com/etseidl) +- Exclude some tests from running with the compute sanitizer ([#13872](https://github.com/NVIDIA/cudf/pull/13872)) [@firestarman](https://github.com/firestarman) +- Expand statistics support in ORC writer ([#13848](https://github.com/NVIDIA/cudf/pull/13848)) [@vuule](https://github.com/vuule) +- Register the memory mapped buffer in `datasource` to improve H2D throughput ([#13814](https://github.com/NVIDIA/cudf/pull/13814)) [@vuule](https://github.com/vuule) +- Add cudf::strings::find function with target per row ([#13808](https://github.com/NVIDIA/cudf/pull/13808)) [@davidwendt](https://github.com/davidwendt) +- Add minhash support for MurmurHash3_x64_128 ([#13796](https://github.com/NVIDIA/cudf/pull/13796)) [@davidwendt](https://github.com/davidwendt) +- Remove unnecessary pointer copying in JIT GroupBy Apply ([#13792](https://github.com/NVIDIA/cudf/pull/13792)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add 'poll' function to custreamz kafka consumer ([#13782](https://github.com/NVIDIA/cudf/pull/13782)) [@jdye64](https://github.com/jdye64) +- Support `corr` in `GroupBy.apply` through the jit engine ([#13767](https://github.com/NVIDIA/cudf/pull/13767)) [@shwina](https://github.com/shwina) +- Optionally write version 2 page headers in Parquet writer ([#13751](https://github.com/NVIDIA/cudf/pull/13751)) [@etseidl](https://github.com/etseidl) +- Support more numeric types in `Groupby.apply` with `engine='jit'` ([#13729](https://github.com/NVIDIA/cudf/pull/13729)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- [FEA] Add DELTA_BINARY_PACKED decoding support to Parquet reader ([#13637](https://github.com/NVIDIA/cudf/pull/13637)) [@etseidl](https://github.com/etseidl) +- Read FIXED_LEN_BYTE_ARRAY as binary in parquet reader ([#13437](https://github.com/NVIDIA/cudf/pull/13437)) [@PointKernel](https://github.com/PointKernel) ## 🛠️ Improvements -- Pin `dask` and `distributed` for `23.10` release ([#14225](https://github.com/rapidsai/cudf/pull/14225)) [@galipremsagar](https://github.com/galipremsagar) -- update rmm tag path ([#14195](https://github.com/rapidsai/cudf/pull/14195)) [@AyodeAwe](https://github.com/AyodeAwe) -- Disable `Recently Updated` Check ([#14193](https://github.com/rapidsai/cudf/pull/14193)) [@ajschmidt8](https://github.com/ajschmidt8) -- Move cpp/src/hash/hash_allocator.cuh to include/cudf/hashing/detail ([#14163](https://github.com/rapidsai/cudf/pull/14163)) [@davidwendt](https://github.com/davidwendt) -- Add Parquet reader benchmarks for row selection ([#14147](https://github.com/rapidsai/cudf/pull/14147)) [@vuule](https://github.com/vuule) -- Update image names ([#14145](https://github.com/rapidsai/cudf/pull/14145)) [@AyodeAwe](https://github.com/AyodeAwe) -- Support callables in DataFrame.assign ([#14142](https://github.com/rapidsai/cudf/pull/14142)) [@wence-](https://github.com/wence-) -- Reduce memory usage of as_categorical_column ([#14138](https://github.com/rapidsai/cudf/pull/14138)) [@wence-](https://github.com/wence-) -- Replace Python scalar conversions with libcudf ([#14124](https://github.com/rapidsai/cudf/pull/14124)) [@vyasr](https://github.com/vyasr) -- Update to clang 16.0.6. ([#14120](https://github.com/rapidsai/cudf/pull/14120)) [@bdice](https://github.com/bdice) -- Fix type of empty `Index` and raise warning in `Series` constructor ([#14116](https://github.com/rapidsai/cudf/pull/14116)) [@galipremsagar](https://github.com/galipremsagar) -- Add stream parameter to external dict APIs ([#14115](https://github.com/rapidsai/cudf/pull/14115)) [@SurajAralihalli](https://github.com/SurajAralihalli) -- Add fallback matrix for nvcomp. ([#14082](https://github.com/rapidsai/cudf/pull/14082)) [@bdice](https://github.com/bdice) -- [Java] Add recoverWithNull to JSONOptions and pass to Table.readJSON ([#14078](https://github.com/rapidsai/cudf/pull/14078)) [@andygrove](https://github.com/andygrove) -- Remove header tests ([#14072](https://github.com/rapidsai/cudf/pull/14072)) [@ajschmidt8](https://github.com/ajschmidt8) -- Refactor `contains_table` with cuco::static_set ([#14064](https://github.com/rapidsai/cudf/pull/14064)) [@PointKernel](https://github.com/PointKernel) -- Remove debug print in a Parquet test ([#14063](https://github.com/rapidsai/cudf/pull/14063)) [@vuule](https://github.com/vuule) -- Expose stream parameter in public nvtext ngram APIs ([#14061](https://github.com/rapidsai/cudf/pull/14061)) [@davidwendt](https://github.com/davidwendt) -- Expose stream parameter in public strings find APIs ([#14060](https://github.com/rapidsai/cudf/pull/14060)) [@davidwendt](https://github.com/davidwendt) -- Update doxygen to 1.9.1 ([#14059](https://github.com/rapidsai/cudf/pull/14059)) [@vyasr](https://github.com/vyasr) -- Remove the mr from the base fixture ([#14057](https://github.com/rapidsai/cudf/pull/14057)) [@vyasr](https://github.com/vyasr) -- Expose streams in public strings case APIs ([#14056](https://github.com/rapidsai/cudf/pull/14056)) [@davidwendt](https://github.com/davidwendt) -- Refactor libcudf indexalator to typed normalator ([#14043](https://github.com/rapidsai/cudf/pull/14043)) [@davidwendt](https://github.com/davidwendt) -- Use cudf::make_empty_column instead of column_view constructor ([#14030](https://github.com/rapidsai/cudf/pull/14030)) [@davidwendt](https://github.com/davidwendt) -- Remove quadratic runtime due to accessing Frame._dtypes in loop ([#14028](https://github.com/rapidsai/cudf/pull/14028)) [@wence-](https://github.com/wence-) -- Explicitly depend on zlib in conda recipes ([#14018](https://github.com/rapidsai/cudf/pull/14018)) [@wence-](https://github.com/wence-) -- Use grid_stride for stride computations. ([#13996](https://github.com/rapidsai/cudf/pull/13996)) [@bdice](https://github.com/bdice) -- Fix an issue where casting null-array to `object` dtype will result in a failure ([#13994](https://github.com/rapidsai/cudf/pull/13994)) [@galipremsagar](https://github.com/galipremsagar) -- Add tab as literal to cudf::test::to_string output ([#13993](https://github.com/rapidsai/cudf/pull/13993)) [@davidwendt](https://github.com/davidwendt) -- Enable `codes` dtype parity in pandas-compatibility mode for `factorize` API ([#13982](https://github.com/rapidsai/cudf/pull/13982)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `CategoricalIndex` ordering in `Groupby.agg` when pandas-compatibility mode is enabled ([#13978](https://github.com/rapidsai/cudf/pull/13978)) [@galipremsagar](https://github.com/galipremsagar) -- Produce a fatal error if cudf is unable to find pyarrow include directory ([#13976](https://github.com/rapidsai/cudf/pull/13976)) [@cwharris](https://github.com/cwharris) -- Use `thread_index_type` in `partitioning.cu` ([#13973](https://github.com/rapidsai/cudf/pull/13973)) [@divyegala](https://github.com/divyegala) -- Use `cudf::thread_index_type` in `merge.cu` ([#13972](https://github.com/rapidsai/cudf/pull/13972)) [@divyegala](https://github.com/divyegala) -- Use `copy-pr-bot` ([#13970](https://github.com/rapidsai/cudf/pull/13970)) [@ajschmidt8](https://github.com/ajschmidt8) -- Use cudf::thread_index_type in strings custom kernels ([#13968](https://github.com/rapidsai/cudf/pull/13968)) [@davidwendt](https://github.com/davidwendt) -- Add `bytes_per_second` to hash_partition benchmark ([#13965](https://github.com/rapidsai/cudf/pull/13965)) [@Blonck](https://github.com/Blonck) -- Added pinned pool reservation API for java ([#13964](https://github.com/rapidsai/cudf/pull/13964)) [@revans2](https://github.com/revans2) -- Simplify wheel build scripts and allow alphas of RAPIDS dependencies ([#13963](https://github.com/rapidsai/cudf/pull/13963)) [@vyasr](https://github.com/vyasr) -- Add `bytes_per_second` to copy_if_else benchmark ([#13960](https://github.com/rapidsai/cudf/pull/13960)) [@Blonck](https://github.com/Blonck) -- Add pandas compatible output to `Series.unique` ([#13959](https://github.com/rapidsai/cudf/pull/13959)) [@galipremsagar](https://github.com/galipremsagar) -- Add `bytes_per_second` to compiled binaryop benchmark ([#13938](https://github.com/rapidsai/cudf/pull/13938)) [@Blonck](https://github.com/Blonck) -- Unpin `dask` and `distributed` for `23.10` development ([#13935](https://github.com/rapidsai/cudf/pull/13935)) [@galipremsagar](https://github.com/galipremsagar) -- Make HostColumnVector.getRefCount public ([#13934](https://github.com/rapidsai/cudf/pull/13934)) [@abellina](https://github.com/abellina) -- Use cuco::static_set in JSON tree algorithm ([#13928](https://github.com/rapidsai/cudf/pull/13928)) [@karthikeyann](https://github.com/karthikeyann) -- Add java API to get size of host memory needed to copy column view ([#13919](https://github.com/rapidsai/cudf/pull/13919)) [@revans2](https://github.com/revans2) -- Use cudf::size_type instead of int32 where appropriate in nvtext functions ([#13915](https://github.com/rapidsai/cudf/pull/13915)) [@davidwendt](https://github.com/davidwendt) -- Enable hugepage for arrow host allocations ([#13914](https://github.com/rapidsai/cudf/pull/13914)) [@madsbk](https://github.com/madsbk) -- Improve performance of nvtext::edit_distance ([#13912](https://github.com/rapidsai/cudf/pull/13912)) [@davidwendt](https://github.com/davidwendt) -- Ensure cudf internals use pylibcudf in pure Python mode ([#13909](https://github.com/rapidsai/cudf/pull/13909)) [@vyasr](https://github.com/vyasr) -- Use `empty()` instead of `size()` where possible ([#13908](https://github.com/rapidsai/cudf/pull/13908)) [@vuule](https://github.com/vuule) -- [JNI] Adds HostColumnVector.EventHandler for spillability checks ([#13898](https://github.com/rapidsai/cudf/pull/13898)) [@abellina](https://github.com/abellina) -- Return `Timestamp` & `Timedelta` for fetching scalars in `DatetimeIndex` & `TimedeltaIndex` ([#13896](https://github.com/rapidsai/cudf/pull/13896)) [@galipremsagar](https://github.com/galipremsagar) -- Allow explicit `shuffle="p2p"` within dask-cudf API ([#13893](https://github.com/rapidsai/cudf/pull/13893)) [@rjzamora](https://github.com/rjzamora) -- Disable creation of `DatetimeIndex` when `freq` is passed to `cudf.date_range` ([#13890](https://github.com/rapidsai/cudf/pull/13890)) [@galipremsagar](https://github.com/galipremsagar) -- Bring parity with pandas for `datetime` & `timedelta` comparison operations ([#13877](https://github.com/rapidsai/cudf/pull/13877)) [@galipremsagar](https://github.com/galipremsagar) -- Change `NA` to `NaT` for `datetime` and `timedelta` types ([#13868](https://github.com/rapidsai/cudf/pull/13868)) [@galipremsagar](https://github.com/galipremsagar) -- Raise error when `astype(object)` is called in pandas compatibility mode ([#13862](https://github.com/rapidsai/cudf/pull/13862)) [@galipremsagar](https://github.com/galipremsagar) -- Fixes a performance regression in FST ([#13850](https://github.com/rapidsai/cudf/pull/13850)) [@elstehle](https://github.com/elstehle) -- Set native handles to null on close in Java wrapper classes ([#13818](https://github.com/rapidsai/cudf/pull/13818)) [@jlowe](https://github.com/jlowe) -- Avoid use of CUDF_EXPECTS in libcudf unit tests outside of helper functions with return values ([#13812](https://github.com/rapidsai/cudf/pull/13812)) [@vuule](https://github.com/vuule) -- Update `lists::contains` to experimental row comparator ([#13810](https://github.com/rapidsai/cudf/pull/13810)) [@divyegala](https://github.com/divyegala) -- Reduce `lists::contains` dispatches for scalars ([#13805](https://github.com/rapidsai/cudf/pull/13805)) [@divyegala](https://github.com/divyegala) -- Long string optimization for string column parsing in JSON reader ([#13803](https://github.com/rapidsai/cudf/pull/13803)) [@karthikeyann](https://github.com/karthikeyann) -- Raise NotImplementedError for pd.SparseDtype ([#13798](https://github.com/rapidsai/cudf/pull/13798)) [@mroeschke](https://github.com/mroeschke) -- Remove the libcudf cudf::offset_type type ([#13788](https://github.com/rapidsai/cudf/pull/13788)) [@davidwendt](https://github.com/davidwendt) -- Move Spark-indpendent Table debug to cudf Java ([#13783](https://github.com/rapidsai/cudf/pull/13783)) [@gerashegalov](https://github.com/gerashegalov) -- Update to Cython 3.0.0 ([#13777](https://github.com/rapidsai/cudf/pull/13777)) [@vyasr](https://github.com/vyasr) -- Refactor Parquet reader handling of V2 page header info ([#13775](https://github.com/rapidsai/cudf/pull/13775)) [@etseidl](https://github.com/etseidl) -- Branch 23.10 merge 23.08 ([#13773](https://github.com/rapidsai/cudf/pull/13773)) [@vyasr](https://github.com/vyasr) -- Restructure JSON code to correctly reflect legacy/experimental status ([#13757](https://github.com/rapidsai/cudf/pull/13757)) [@vuule](https://github.com/vuule) -- Branch 23.10 merge 23.08 ([#13753](https://github.com/rapidsai/cudf/pull/13753)) [@vyasr](https://github.com/vyasr) -- Enforce deprecations in `23.10` ([#13732](https://github.com/rapidsai/cudf/pull/13732)) [@galipremsagar](https://github.com/galipremsagar) -- Upgrade to arrow 12 ([#13728](https://github.com/rapidsai/cudf/pull/13728)) [@galipremsagar](https://github.com/galipremsagar) -- Refactors JSON reader's pushdown automaton ([#13716](https://github.com/rapidsai/cudf/pull/13716)) [@elstehle](https://github.com/elstehle) -- Remove Arrow dependency from the `datasource.hpp` public header ([#13698](https://github.com/rapidsai/cudf/pull/13698)) [@vuule](https://github.com/vuule) +- Pin `dask` and `distributed` for `23.10` release ([#14225](https://github.com/NVIDIA/cudf/pull/14225)) [@galipremsagar](https://github.com/galipremsagar) +- update rmm tag path ([#14195](https://github.com/NVIDIA/cudf/pull/14195)) [@AyodeAwe](https://github.com/AyodeAwe) +- Disable `Recently Updated` Check ([#14193](https://github.com/NVIDIA/cudf/pull/14193)) [@ajschmidt8](https://github.com/ajschmidt8) +- Move cpp/src/hash/hash_allocator.cuh to include/cudf/hashing/detail ([#14163](https://github.com/NVIDIA/cudf/pull/14163)) [@davidwendt](https://github.com/davidwendt) +- Add Parquet reader benchmarks for row selection ([#14147](https://github.com/NVIDIA/cudf/pull/14147)) [@vuule](https://github.com/vuule) +- Update image names ([#14145](https://github.com/NVIDIA/cudf/pull/14145)) [@AyodeAwe](https://github.com/AyodeAwe) +- Support callables in DataFrame.assign ([#14142](https://github.com/NVIDIA/cudf/pull/14142)) [@wence-](https://github.com/wence-) +- Reduce memory usage of as_categorical_column ([#14138](https://github.com/NVIDIA/cudf/pull/14138)) [@wence-](https://github.com/wence-) +- Replace Python scalar conversions with libcudf ([#14124](https://github.com/NVIDIA/cudf/pull/14124)) [@vyasr](https://github.com/vyasr) +- Update to clang 16.0.6. ([#14120](https://github.com/NVIDIA/cudf/pull/14120)) [@bdice](https://github.com/bdice) +- Fix type of empty `Index` and raise warning in `Series` constructor ([#14116](https://github.com/NVIDIA/cudf/pull/14116)) [@galipremsagar](https://github.com/galipremsagar) +- Add stream parameter to external dict APIs ([#14115](https://github.com/NVIDIA/cudf/pull/14115)) [@SurajAralihalli](https://github.com/SurajAralihalli) +- Add fallback matrix for nvcomp. ([#14082](https://github.com/NVIDIA/cudf/pull/14082)) [@bdice](https://github.com/bdice) +- [Java] Add recoverWithNull to JSONOptions and pass to Table.readJSON ([#14078](https://github.com/NVIDIA/cudf/pull/14078)) [@andygrove](https://github.com/andygrove) +- Remove header tests ([#14072](https://github.com/NVIDIA/cudf/pull/14072)) [@ajschmidt8](https://github.com/ajschmidt8) +- Refactor `contains_table` with cuco::static_set ([#14064](https://github.com/NVIDIA/cudf/pull/14064)) [@PointKernel](https://github.com/PointKernel) +- Remove debug print in a Parquet test ([#14063](https://github.com/NVIDIA/cudf/pull/14063)) [@vuule](https://github.com/vuule) +- Expose stream parameter in public nvtext ngram APIs ([#14061](https://github.com/NVIDIA/cudf/pull/14061)) [@davidwendt](https://github.com/davidwendt) +- Expose stream parameter in public strings find APIs ([#14060](https://github.com/NVIDIA/cudf/pull/14060)) [@davidwendt](https://github.com/davidwendt) +- Update doxygen to 1.9.1 ([#14059](https://github.com/NVIDIA/cudf/pull/14059)) [@vyasr](https://github.com/vyasr) +- Remove the mr from the base fixture ([#14057](https://github.com/NVIDIA/cudf/pull/14057)) [@vyasr](https://github.com/vyasr) +- Expose streams in public strings case APIs ([#14056](https://github.com/NVIDIA/cudf/pull/14056)) [@davidwendt](https://github.com/davidwendt) +- Refactor libcudf indexalator to typed normalator ([#14043](https://github.com/NVIDIA/cudf/pull/14043)) [@davidwendt](https://github.com/davidwendt) +- Use cudf::make_empty_column instead of column_view constructor ([#14030](https://github.com/NVIDIA/cudf/pull/14030)) [@davidwendt](https://github.com/davidwendt) +- Remove quadratic runtime due to accessing Frame._dtypes in loop ([#14028](https://github.com/NVIDIA/cudf/pull/14028)) [@wence-](https://github.com/wence-) +- Explicitly depend on zlib in conda recipes ([#14018](https://github.com/NVIDIA/cudf/pull/14018)) [@wence-](https://github.com/wence-) +- Use grid_stride for stride computations. ([#13996](https://github.com/NVIDIA/cudf/pull/13996)) [@bdice](https://github.com/bdice) +- Fix an issue where casting null-array to `object` dtype will result in a failure ([#13994](https://github.com/NVIDIA/cudf/pull/13994)) [@galipremsagar](https://github.com/galipremsagar) +- Add tab as literal to cudf::test::to_string output ([#13993](https://github.com/NVIDIA/cudf/pull/13993)) [@davidwendt](https://github.com/davidwendt) +- Enable `codes` dtype parity in pandas-compatibility mode for `factorize` API ([#13982](https://github.com/NVIDIA/cudf/pull/13982)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `CategoricalIndex` ordering in `Groupby.agg` when pandas-compatibility mode is enabled ([#13978](https://github.com/NVIDIA/cudf/pull/13978)) [@galipremsagar](https://github.com/galipremsagar) +- Produce a fatal error if cudf is unable to find pyarrow include directory ([#13976](https://github.com/NVIDIA/cudf/pull/13976)) [@cwharris](https://github.com/cwharris) +- Use `thread_index_type` in `partitioning.cu` ([#13973](https://github.com/NVIDIA/cudf/pull/13973)) [@divyegala](https://github.com/divyegala) +- Use `cudf::thread_index_type` in `merge.cu` ([#13972](https://github.com/NVIDIA/cudf/pull/13972)) [@divyegala](https://github.com/divyegala) +- Use `copy-pr-bot` ([#13970](https://github.com/NVIDIA/cudf/pull/13970)) [@ajschmidt8](https://github.com/ajschmidt8) +- Use cudf::thread_index_type in strings custom kernels ([#13968](https://github.com/NVIDIA/cudf/pull/13968)) [@davidwendt](https://github.com/davidwendt) +- Add `bytes_per_second` to hash_partition benchmark ([#13965](https://github.com/NVIDIA/cudf/pull/13965)) [@Blonck](https://github.com/Blonck) +- Added pinned pool reservation API for java ([#13964](https://github.com/NVIDIA/cudf/pull/13964)) [@revans2](https://github.com/revans2) +- Simplify wheel build scripts and allow alphas of RAPIDS dependencies ([#13963](https://github.com/NVIDIA/cudf/pull/13963)) [@vyasr](https://github.com/vyasr) +- Add `bytes_per_second` to copy_if_else benchmark ([#13960](https://github.com/NVIDIA/cudf/pull/13960)) [@Blonck](https://github.com/Blonck) +- Add pandas compatible output to `Series.unique` ([#13959](https://github.com/NVIDIA/cudf/pull/13959)) [@galipremsagar](https://github.com/galipremsagar) +- Add `bytes_per_second` to compiled binaryop benchmark ([#13938](https://github.com/NVIDIA/cudf/pull/13938)) [@Blonck](https://github.com/Blonck) +- Unpin `dask` and `distributed` for `23.10` development ([#13935](https://github.com/NVIDIA/cudf/pull/13935)) [@galipremsagar](https://github.com/galipremsagar) +- Make HostColumnVector.getRefCount public ([#13934](https://github.com/NVIDIA/cudf/pull/13934)) [@abellina](https://github.com/abellina) +- Use cuco::static_set in JSON tree algorithm ([#13928](https://github.com/NVIDIA/cudf/pull/13928)) [@karthikeyann](https://github.com/karthikeyann) +- Add java API to get size of host memory needed to copy column view ([#13919](https://github.com/NVIDIA/cudf/pull/13919)) [@revans2](https://github.com/revans2) +- Use cudf::size_type instead of int32 where appropriate in nvtext functions ([#13915](https://github.com/NVIDIA/cudf/pull/13915)) [@davidwendt](https://github.com/davidwendt) +- Enable hugepage for arrow host allocations ([#13914](https://github.com/NVIDIA/cudf/pull/13914)) [@madsbk](https://github.com/madsbk) +- Improve performance of nvtext::edit_distance ([#13912](https://github.com/NVIDIA/cudf/pull/13912)) [@davidwendt](https://github.com/davidwendt) +- Ensure cudf internals use pylibcudf in pure Python mode ([#13909](https://github.com/NVIDIA/cudf/pull/13909)) [@vyasr](https://github.com/vyasr) +- Use `empty()` instead of `size()` where possible ([#13908](https://github.com/NVIDIA/cudf/pull/13908)) [@vuule](https://github.com/vuule) +- [JNI] Adds HostColumnVector.EventHandler for spillability checks ([#13898](https://github.com/NVIDIA/cudf/pull/13898)) [@abellina](https://github.com/abellina) +- Return `Timestamp` & `Timedelta` for fetching scalars in `DatetimeIndex` & `TimedeltaIndex` ([#13896](https://github.com/NVIDIA/cudf/pull/13896)) [@galipremsagar](https://github.com/galipremsagar) +- Allow explicit `shuffle="p2p"` within dask-cudf API ([#13893](https://github.com/NVIDIA/cudf/pull/13893)) [@rjzamora](https://github.com/rjzamora) +- Disable creation of `DatetimeIndex` when `freq` is passed to `cudf.date_range` ([#13890](https://github.com/NVIDIA/cudf/pull/13890)) [@galipremsagar](https://github.com/galipremsagar) +- Bring parity with pandas for `datetime` & `timedelta` comparison operations ([#13877](https://github.com/NVIDIA/cudf/pull/13877)) [@galipremsagar](https://github.com/galipremsagar) +- Change `NA` to `NaT` for `datetime` and `timedelta` types ([#13868](https://github.com/NVIDIA/cudf/pull/13868)) [@galipremsagar](https://github.com/galipremsagar) +- Raise error when `astype(object)` is called in pandas compatibility mode ([#13862](https://github.com/NVIDIA/cudf/pull/13862)) [@galipremsagar](https://github.com/galipremsagar) +- Fixes a performance regression in FST ([#13850](https://github.com/NVIDIA/cudf/pull/13850)) [@elstehle](https://github.com/elstehle) +- Set native handles to null on close in Java wrapper classes ([#13818](https://github.com/NVIDIA/cudf/pull/13818)) [@jlowe](https://github.com/jlowe) +- Avoid use of CUDF_EXPECTS in libcudf unit tests outside of helper functions with return values ([#13812](https://github.com/NVIDIA/cudf/pull/13812)) [@vuule](https://github.com/vuule) +- Update `lists::contains` to experimental row comparator ([#13810](https://github.com/NVIDIA/cudf/pull/13810)) [@divyegala](https://github.com/divyegala) +- Reduce `lists::contains` dispatches for scalars ([#13805](https://github.com/NVIDIA/cudf/pull/13805)) [@divyegala](https://github.com/divyegala) +- Long string optimization for string column parsing in JSON reader ([#13803](https://github.com/NVIDIA/cudf/pull/13803)) [@karthikeyann](https://github.com/karthikeyann) +- Raise NotImplementedError for pd.SparseDtype ([#13798](https://github.com/NVIDIA/cudf/pull/13798)) [@mroeschke](https://github.com/mroeschke) +- Remove the libcudf cudf::offset_type type ([#13788](https://github.com/NVIDIA/cudf/pull/13788)) [@davidwendt](https://github.com/davidwendt) +- Move Spark-indpendent Table debug to cudf Java ([#13783](https://github.com/NVIDIA/cudf/pull/13783)) [@gerashegalov](https://github.com/gerashegalov) +- Update to Cython 3.0.0 ([#13777](https://github.com/NVIDIA/cudf/pull/13777)) [@vyasr](https://github.com/vyasr) +- Refactor Parquet reader handling of V2 page header info ([#13775](https://github.com/NVIDIA/cudf/pull/13775)) [@etseidl](https://github.com/etseidl) +- Branch 23.10 merge 23.08 ([#13773](https://github.com/NVIDIA/cudf/pull/13773)) [@vyasr](https://github.com/vyasr) +- Restructure JSON code to correctly reflect legacy/experimental status ([#13757](https://github.com/NVIDIA/cudf/pull/13757)) [@vuule](https://github.com/vuule) +- Branch 23.10 merge 23.08 ([#13753](https://github.com/NVIDIA/cudf/pull/13753)) [@vyasr](https://github.com/vyasr) +- Enforce deprecations in `23.10` ([#13732](https://github.com/NVIDIA/cudf/pull/13732)) [@galipremsagar](https://github.com/galipremsagar) +- Upgrade to arrow 12 ([#13728](https://github.com/NVIDIA/cudf/pull/13728)) [@galipremsagar](https://github.com/galipremsagar) +- Refactors JSON reader's pushdown automaton ([#13716](https://github.com/NVIDIA/cudf/pull/13716)) [@elstehle](https://github.com/elstehle) +- Remove Arrow dependency from the `datasource.hpp` public header ([#13698](https://github.com/NVIDIA/cudf/pull/13698)) [@vuule](https://github.com/vuule) # cuDF 23.08.00 (9 Aug 2023) ## 🚨 Breaking Changes -- Enforce deprecations and add clarifications around existing deprecations ([#13710](https://github.com/rapidsai/cudf/pull/13710)) [@galipremsagar](https://github.com/galipremsagar) -- Separate MurmurHash32 from hash_functions.cuh ([#13681](https://github.com/rapidsai/cudf/pull/13681)) [@davidwendt](https://github.com/davidwendt) -- Avoid storing metadata in pointers in ORC and Parquet writers ([#13648](https://github.com/rapidsai/cudf/pull/13648)) [@vuule](https://github.com/vuule) -- Expose streams in all public copying APIs ([#13629](https://github.com/rapidsai/cudf/pull/13629)) [@vyasr](https://github.com/vyasr) -- Remove deprecated cudf::strings::slice_strings (by delimiter) functions ([#13628](https://github.com/rapidsai/cudf/pull/13628)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated cudf.set_allocator. ([#13591](https://github.com/rapidsai/cudf/pull/13591)) [@bdice](https://github.com/bdice) -- Change build.sh to use pip install instead of setup.py ([#13507](https://github.com/rapidsai/cudf/pull/13507)) [@vyasr](https://github.com/vyasr) -- Remove unused max_rows_tensor parameter from subword tokenizer ([#13463](https://github.com/rapidsai/cudf/pull/13463)) [@davidwendt](https://github.com/davidwendt) -- Fix decimal scale reductions in `_get_decimal_type` ([#13224](https://github.com/rapidsai/cudf/pull/13224)) [@charlesbluca](https://github.com/charlesbluca) +- Enforce deprecations and add clarifications around existing deprecations ([#13710](https://github.com/NVIDIA/cudf/pull/13710)) [@galipremsagar](https://github.com/galipremsagar) +- Separate MurmurHash32 from hash_functions.cuh ([#13681](https://github.com/NVIDIA/cudf/pull/13681)) [@davidwendt](https://github.com/davidwendt) +- Avoid storing metadata in pointers in ORC and Parquet writers ([#13648](https://github.com/NVIDIA/cudf/pull/13648)) [@vuule](https://github.com/vuule) +- Expose streams in all public copying APIs ([#13629](https://github.com/NVIDIA/cudf/pull/13629)) [@vyasr](https://github.com/vyasr) +- Remove deprecated cudf::strings::slice_strings (by delimiter) functions ([#13628](https://github.com/NVIDIA/cudf/pull/13628)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated cudf.set_allocator. ([#13591](https://github.com/NVIDIA/cudf/pull/13591)) [@bdice](https://github.com/bdice) +- Change build.sh to use pip install instead of setup.py ([#13507](https://github.com/NVIDIA/cudf/pull/13507)) [@vyasr](https://github.com/vyasr) +- Remove unused max_rows_tensor parameter from subword tokenizer ([#13463](https://github.com/NVIDIA/cudf/pull/13463)) [@davidwendt](https://github.com/davidwendt) +- Fix decimal scale reductions in `_get_decimal_type` ([#13224](https://github.com/NVIDIA/cudf/pull/13224)) [@charlesbluca](https://github.com/charlesbluca) ## 🐛 Bug Fixes -- Add CUDA version to cudf_kafka and libcudf-example build strings. ([#13769](https://github.com/rapidsai/cudf/pull/13769)) [@bdice](https://github.com/bdice) -- Fix typo in wheels-test.yaml. ([#13763](https://github.com/rapidsai/cudf/pull/13763)) [@bdice](https://github.com/bdice) -- Don't test strings shorter than the requested ngram size ([#13758](https://github.com/rapidsai/cudf/pull/13758)) [@vyasr](https://github.com/vyasr) -- Add CUDA version to custreamz build string. ([#13754](https://github.com/rapidsai/cudf/pull/13754)) [@bdice](https://github.com/bdice) -- Fix writing of ORC files with empty child string columns ([#13745](https://github.com/rapidsai/cudf/pull/13745)) [@vuule](https://github.com/vuule) -- Remove the erroneous "empty level" short-circuit from ORC reader ([#13722](https://github.com/rapidsai/cudf/pull/13722)) [@vuule](https://github.com/vuule) -- Fix character counting when writing sliced tables into ORC ([#13721](https://github.com/rapidsai/cudf/pull/13721)) [@vuule](https://github.com/vuule) -- Parquet uses row group row count if missing from header ([#13712](https://github.com/rapidsai/cudf/pull/13712)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix reading of RLE encoded boolean data from parquet files with V2 page headers ([#13707](https://github.com/rapidsai/cudf/pull/13707)) [@etseidl](https://github.com/etseidl) -- Fix a corner case of list lexicographic comparator ([#13701](https://github.com/rapidsai/cudf/pull/13701)) [@ttnghia](https://github.com/ttnghia) -- Fix combined filtering and column projection in `dask_cudf.read_parquet` ([#13697](https://github.com/rapidsai/cudf/pull/13697)) [@rjzamora](https://github.com/rjzamora) -- Revert fetch-rapids changes ([#13696](https://github.com/rapidsai/cudf/pull/13696)) [@vyasr](https://github.com/vyasr) -- Data generator - include offsets in the size estimate of list elments ([#13688](https://github.com/rapidsai/cudf/pull/13688)) [@vuule](https://github.com/vuule) -- Add `cuda-nvcc-impl` to `cudf` for `numba` CUDA 12 ([#13673](https://github.com/rapidsai/cudf/pull/13673)) [@jakirkham](https://github.com/jakirkham) -- Fix combined filtering and column projection in `read_parquet` ([#13666](https://github.com/rapidsai/cudf/pull/13666)) [@rjzamora](https://github.com/rjzamora) -- Use `thrust::identity` as hash functions for byte pair encoding ([#13665](https://github.com/rapidsai/cudf/pull/13665)) [@PointKernel](https://github.com/PointKernel) -- Fix loc-getitem ordering when index contains duplicate labels ([#13659](https://github.com/rapidsai/cudf/pull/13659)) [@wence-](https://github.com/wence-) -- [REVIEW] Introduce parity with pandas for `MultiIndex.loc` ordering & fix a bug in `Groupby` with `as_index` ([#13657](https://github.com/rapidsai/cudf/pull/13657)) [@galipremsagar](https://github.com/galipremsagar) -- Fix memcheck error found in nvtext tokenize functions ([#13649](https://github.com/rapidsai/cudf/pull/13649)) [@davidwendt](https://github.com/davidwendt) -- Fix `has_nonempty_nulls` ignoring column offset ([#13647](https://github.com/rapidsai/cudf/pull/13647)) [@ttnghia](https://github.com/ttnghia) -- [Java] Avoid double-free corruption in case of an Exception while creating a ColumnView ([#13645](https://github.com/rapidsai/cudf/pull/13645)) [@razajafri](https://github.com/razajafri) -- Fix memcheck error in ORC reader call to cudf::io::copy_uncompressed_kernel ([#13643](https://github.com/rapidsai/cudf/pull/13643)) [@davidwendt](https://github.com/davidwendt) -- Fix CUDA 12 conda environment to remove cubinlinker and ptxcompiler. ([#13636](https://github.com/rapidsai/cudf/pull/13636)) [@bdice](https://github.com/bdice) -- Fix inf/NaN comparisons for FLOAT orderby in window functions ([#13635](https://github.com/rapidsai/cudf/pull/13635)) [@mythrocks](https://github.com/mythrocks) -- Refactor `Index` search to simplify code and increase correctness ([#13625](https://github.com/rapidsai/cudf/pull/13625)) [@wence-](https://github.com/wence-) -- Fix compile warning for unused variable in split_re.cu ([#13621](https://github.com/rapidsai/cudf/pull/13621)) [@davidwendt](https://github.com/davidwendt) -- Fix tz_localize for dask_cudf Series ([#13610](https://github.com/rapidsai/cudf/pull/13610)) [@shwina](https://github.com/shwina) -- Fix issue with no decompressed data in ORC reader ([#13609](https://github.com/rapidsai/cudf/pull/13609)) [@vuule](https://github.com/vuule) -- Fix floating point window range extents. ([#13606](https://github.com/rapidsai/cudf/pull/13606)) [@mythrocks](https://github.com/mythrocks) -- Fix `localize(None)` for timezone-naive columns ([#13603](https://github.com/rapidsai/cudf/pull/13603)) [@shwina](https://github.com/shwina) -- Fixed a memory leak caused by Exception thrown while constructing a ColumnView ([#13597](https://github.com/rapidsai/cudf/pull/13597)) [@razajafri](https://github.com/razajafri) -- Handle nullptr return value from bitmask_or in distinct_count ([#13590](https://github.com/rapidsai/cudf/pull/13590)) [@wence-](https://github.com/wence-) -- Bring parity with pandas in Index.join ([#13589](https://github.com/rapidsai/cudf/pull/13589)) [@galipremsagar](https://github.com/galipremsagar) -- Fix cudf.melt when there are more than 255 columns ([#13588](https://github.com/rapidsai/cudf/pull/13588)) [@hcho3](https://github.com/hcho3) -- Fix memory issues in cuIO due to removal of memory padding ([#13586](https://github.com/rapidsai/cudf/pull/13586)) [@ttnghia](https://github.com/ttnghia) -- Fix Parquet multi-file reading ([#13584](https://github.com/rapidsai/cudf/pull/13584)) [@etseidl](https://github.com/etseidl) -- Fix memcheck error found in LISTS_TEST ([#13579](https://github.com/rapidsai/cudf/pull/13579)) [@davidwendt](https://github.com/davidwendt) -- Fix memcheck error found in STRINGS_TEST ([#13578](https://github.com/rapidsai/cudf/pull/13578)) [@davidwendt](https://github.com/davidwendt) -- Fix memcheck error found in INTEROP_TEST ([#13577](https://github.com/rapidsai/cudf/pull/13577)) [@davidwendt](https://github.com/davidwendt) -- Fix memcheck errors found in REDUCTION_TEST ([#13574](https://github.com/rapidsai/cudf/pull/13574)) [@davidwendt](https://github.com/davidwendt) -- Preemptive fix for hive-partitioning change in dask ([#13564](https://github.com/rapidsai/cudf/pull/13564)) [@rjzamora](https://github.com/rjzamora) -- Fix an issue with `dask_cudf.read_csv` when lines are needed to be skipped ([#13555](https://github.com/rapidsai/cudf/pull/13555)) [@galipremsagar](https://github.com/galipremsagar) -- Fix out-of-bounds memory write in cudf::dictionary::detail::concatenate ([#13554](https://github.com/rapidsai/cudf/pull/13554)) [@davidwendt](https://github.com/davidwendt) -- Fix the null mask size in json reader ([#13537](https://github.com/rapidsai/cudf/pull/13537)) [@karthikeyann](https://github.com/karthikeyann) -- Fix cudf::strings::strip for all-empty input column ([#13533](https://github.com/rapidsai/cudf/pull/13533)) [@davidwendt](https://github.com/davidwendt) -- Make sure to build without isolation or installing dependencies ([#13524](https://github.com/rapidsai/cudf/pull/13524)) [@vyasr](https://github.com/vyasr) -- Remove preload lib from CMake for now ([#13519](https://github.com/rapidsai/cudf/pull/13519)) [@vyasr](https://github.com/vyasr) -- Fix missing separator after null values in JSON writer ([#13503](https://github.com/rapidsai/cudf/pull/13503)) [@karthikeyann](https://github.com/karthikeyann) -- Ensure `single_lane_block_sum_reduce` is safe to call in a loop ([#13488](https://github.com/rapidsai/cudf/pull/13488)) [@wence-](https://github.com/wence-) -- Update all versions in pyproject.toml files. ([#13486](https://github.com/rapidsai/cudf/pull/13486)) [@bdice](https://github.com/bdice) -- Remove applying nvbench that doesn't exist in 23.08 ([#13484](https://github.com/rapidsai/cudf/pull/13484)) [@robertmaynard](https://github.com/robertmaynard) -- Fix chunked Parquet reader benchmark ([#13482](https://github.com/rapidsai/cudf/pull/13482)) [@vuule](https://github.com/vuule) -- Update JNI JSON reader column compatability for Spark ([#13477](https://github.com/rapidsai/cudf/pull/13477)) [@revans2](https://github.com/revans2) -- Fix unsanitized output of scan with strings ([#13455](https://github.com/rapidsai/cudf/pull/13455)) [@davidwendt](https://github.com/davidwendt) -- Reject functions without bytecode from `_can_be_jitted` in GroupBy Apply ([#13429](https://github.com/rapidsai/cudf/pull/13429)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix decimal scale reductions in `_get_decimal_type` ([#13224](https://github.com/rapidsai/cudf/pull/13224)) [@charlesbluca](https://github.com/charlesbluca) +- Add CUDA version to cudf_kafka and libcudf-example build strings. ([#13769](https://github.com/NVIDIA/cudf/pull/13769)) [@bdice](https://github.com/bdice) +- Fix typo in wheels-test.yaml. ([#13763](https://github.com/NVIDIA/cudf/pull/13763)) [@bdice](https://github.com/bdice) +- Don't test strings shorter than the requested ngram size ([#13758](https://github.com/NVIDIA/cudf/pull/13758)) [@vyasr](https://github.com/vyasr) +- Add CUDA version to custreamz build string. ([#13754](https://github.com/NVIDIA/cudf/pull/13754)) [@bdice](https://github.com/bdice) +- Fix writing of ORC files with empty child string columns ([#13745](https://github.com/NVIDIA/cudf/pull/13745)) [@vuule](https://github.com/vuule) +- Remove the erroneous "empty level" short-circuit from ORC reader ([#13722](https://github.com/NVIDIA/cudf/pull/13722)) [@vuule](https://github.com/vuule) +- Fix character counting when writing sliced tables into ORC ([#13721](https://github.com/NVIDIA/cudf/pull/13721)) [@vuule](https://github.com/vuule) +- Parquet uses row group row count if missing from header ([#13712](https://github.com/NVIDIA/cudf/pull/13712)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix reading of RLE encoded boolean data from parquet files with V2 page headers ([#13707](https://github.com/NVIDIA/cudf/pull/13707)) [@etseidl](https://github.com/etseidl) +- Fix a corner case of list lexicographic comparator ([#13701](https://github.com/NVIDIA/cudf/pull/13701)) [@ttnghia](https://github.com/ttnghia) +- Fix combined filtering and column projection in `dask_cudf.read_parquet` ([#13697](https://github.com/NVIDIA/cudf/pull/13697)) [@rjzamora](https://github.com/rjzamora) +- Revert fetch-rapids changes ([#13696](https://github.com/NVIDIA/cudf/pull/13696)) [@vyasr](https://github.com/vyasr) +- Data generator - include offsets in the size estimate of list elments ([#13688](https://github.com/NVIDIA/cudf/pull/13688)) [@vuule](https://github.com/vuule) +- Add `cuda-nvcc-impl` to `cudf` for `numba` CUDA 12 ([#13673](https://github.com/NVIDIA/cudf/pull/13673)) [@jakirkham](https://github.com/jakirkham) +- Fix combined filtering and column projection in `read_parquet` ([#13666](https://github.com/NVIDIA/cudf/pull/13666)) [@rjzamora](https://github.com/rjzamora) +- Use `thrust::identity` as hash functions for byte pair encoding ([#13665](https://github.com/NVIDIA/cudf/pull/13665)) [@PointKernel](https://github.com/PointKernel) +- Fix loc-getitem ordering when index contains duplicate labels ([#13659](https://github.com/NVIDIA/cudf/pull/13659)) [@wence-](https://github.com/wence-) +- [REVIEW] Introduce parity with pandas for `MultiIndex.loc` ordering & fix a bug in `Groupby` with `as_index` ([#13657](https://github.com/NVIDIA/cudf/pull/13657)) [@galipremsagar](https://github.com/galipremsagar) +- Fix memcheck error found in nvtext tokenize functions ([#13649](https://github.com/NVIDIA/cudf/pull/13649)) [@davidwendt](https://github.com/davidwendt) +- Fix `has_nonempty_nulls` ignoring column offset ([#13647](https://github.com/NVIDIA/cudf/pull/13647)) [@ttnghia](https://github.com/ttnghia) +- [Java] Avoid double-free corruption in case of an Exception while creating a ColumnView ([#13645](https://github.com/NVIDIA/cudf/pull/13645)) [@razajafri](https://github.com/razajafri) +- Fix memcheck error in ORC reader call to cudf::io::copy_uncompressed_kernel ([#13643](https://github.com/NVIDIA/cudf/pull/13643)) [@davidwendt](https://github.com/davidwendt) +- Fix CUDA 12 conda environment to remove cubinlinker and ptxcompiler. ([#13636](https://github.com/NVIDIA/cudf/pull/13636)) [@bdice](https://github.com/bdice) +- Fix inf/NaN comparisons for FLOAT orderby in window functions ([#13635](https://github.com/NVIDIA/cudf/pull/13635)) [@mythrocks](https://github.com/mythrocks) +- Refactor `Index` search to simplify code and increase correctness ([#13625](https://github.com/NVIDIA/cudf/pull/13625)) [@wence-](https://github.com/wence-) +- Fix compile warning for unused variable in split_re.cu ([#13621](https://github.com/NVIDIA/cudf/pull/13621)) [@davidwendt](https://github.com/davidwendt) +- Fix tz_localize for dask_cudf Series ([#13610](https://github.com/NVIDIA/cudf/pull/13610)) [@shwina](https://github.com/shwina) +- Fix issue with no decompressed data in ORC reader ([#13609](https://github.com/NVIDIA/cudf/pull/13609)) [@vuule](https://github.com/vuule) +- Fix floating point window range extents. ([#13606](https://github.com/NVIDIA/cudf/pull/13606)) [@mythrocks](https://github.com/mythrocks) +- Fix `localize(None)` for timezone-naive columns ([#13603](https://github.com/NVIDIA/cudf/pull/13603)) [@shwina](https://github.com/shwina) +- Fixed a memory leak caused by Exception thrown while constructing a ColumnView ([#13597](https://github.com/NVIDIA/cudf/pull/13597)) [@razajafri](https://github.com/razajafri) +- Handle nullptr return value from bitmask_or in distinct_count ([#13590](https://github.com/NVIDIA/cudf/pull/13590)) [@wence-](https://github.com/wence-) +- Bring parity with pandas in Index.join ([#13589](https://github.com/NVIDIA/cudf/pull/13589)) [@galipremsagar](https://github.com/galipremsagar) +- Fix cudf.melt when there are more than 255 columns ([#13588](https://github.com/NVIDIA/cudf/pull/13588)) [@hcho3](https://github.com/hcho3) +- Fix memory issues in cuIO due to removal of memory padding ([#13586](https://github.com/NVIDIA/cudf/pull/13586)) [@ttnghia](https://github.com/ttnghia) +- Fix Parquet multi-file reading ([#13584](https://github.com/NVIDIA/cudf/pull/13584)) [@etseidl](https://github.com/etseidl) +- Fix memcheck error found in LISTS_TEST ([#13579](https://github.com/NVIDIA/cudf/pull/13579)) [@davidwendt](https://github.com/davidwendt) +- Fix memcheck error found in STRINGS_TEST ([#13578](https://github.com/NVIDIA/cudf/pull/13578)) [@davidwendt](https://github.com/davidwendt) +- Fix memcheck error found in INTEROP_TEST ([#13577](https://github.com/NVIDIA/cudf/pull/13577)) [@davidwendt](https://github.com/davidwendt) +- Fix memcheck errors found in REDUCTION_TEST ([#13574](https://github.com/NVIDIA/cudf/pull/13574)) [@davidwendt](https://github.com/davidwendt) +- Preemptive fix for hive-partitioning change in dask ([#13564](https://github.com/NVIDIA/cudf/pull/13564)) [@rjzamora](https://github.com/rjzamora) +- Fix an issue with `dask_cudf.read_csv` when lines are needed to be skipped ([#13555](https://github.com/NVIDIA/cudf/pull/13555)) [@galipremsagar](https://github.com/galipremsagar) +- Fix out-of-bounds memory write in cudf::dictionary::detail::concatenate ([#13554](https://github.com/NVIDIA/cudf/pull/13554)) [@davidwendt](https://github.com/davidwendt) +- Fix the null mask size in json reader ([#13537](https://github.com/NVIDIA/cudf/pull/13537)) [@karthikeyann](https://github.com/karthikeyann) +- Fix cudf::strings::strip for all-empty input column ([#13533](https://github.com/NVIDIA/cudf/pull/13533)) [@davidwendt](https://github.com/davidwendt) +- Make sure to build without isolation or installing dependencies ([#13524](https://github.com/NVIDIA/cudf/pull/13524)) [@vyasr](https://github.com/vyasr) +- Remove preload lib from CMake for now ([#13519](https://github.com/NVIDIA/cudf/pull/13519)) [@vyasr](https://github.com/vyasr) +- Fix missing separator after null values in JSON writer ([#13503](https://github.com/NVIDIA/cudf/pull/13503)) [@karthikeyann](https://github.com/karthikeyann) +- Ensure `single_lane_block_sum_reduce` is safe to call in a loop ([#13488](https://github.com/NVIDIA/cudf/pull/13488)) [@wence-](https://github.com/wence-) +- Update all versions in pyproject.toml files. ([#13486](https://github.com/NVIDIA/cudf/pull/13486)) [@bdice](https://github.com/bdice) +- Remove applying nvbench that doesn't exist in 23.08 ([#13484](https://github.com/NVIDIA/cudf/pull/13484)) [@robertmaynard](https://github.com/robertmaynard) +- Fix chunked Parquet reader benchmark ([#13482](https://github.com/NVIDIA/cudf/pull/13482)) [@vuule](https://github.com/vuule) +- Update JNI JSON reader column compatability for Spark ([#13477](https://github.com/NVIDIA/cudf/pull/13477)) [@revans2](https://github.com/revans2) +- Fix unsanitized output of scan with strings ([#13455](https://github.com/NVIDIA/cudf/pull/13455)) [@davidwendt](https://github.com/davidwendt) +- Reject functions without bytecode from `_can_be_jitted` in GroupBy Apply ([#13429](https://github.com/NVIDIA/cudf/pull/13429)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix decimal scale reductions in `_get_decimal_type` ([#13224](https://github.com/NVIDIA/cudf/pull/13224)) [@charlesbluca](https://github.com/charlesbluca) ## 📖 Documentation -- Fix doxygen groups for io data sources and sinks ([#13718](https://github.com/rapidsai/cudf/pull/13718)) [@davidwendt](https://github.com/davidwendt) -- Add pandas compatibility note to DataFrame.query docstring ([#13693](https://github.com/rapidsai/cudf/pull/13693)) [@beckernick](https://github.com/beckernick) -- Add pylibcudf to developer guide ([#13639](https://github.com/rapidsai/cudf/pull/13639)) [@vyasr](https://github.com/vyasr) -- Fix repeated words in doxygen text ([#13598](https://github.com/rapidsai/cudf/pull/13598)) [@karthikeyann](https://github.com/karthikeyann) -- Update docs for top-level API. ([#13592](https://github.com/rapidsai/cudf/pull/13592)) [@bdice](https://github.com/bdice) -- Fix the the doxygen text for cudf::concatenate and other places ([#13561](https://github.com/rapidsai/cudf/pull/13561)) [@davidwendt](https://github.com/davidwendt) -- Document stream validation approach used in testing ([#13556](https://github.com/rapidsai/cudf/pull/13556)) [@vyasr](https://github.com/vyasr) -- Cleanup doc repetitions in libcudf ([#13470](https://github.com/rapidsai/cudf/pull/13470)) [@karthikeyann](https://github.com/karthikeyann) +- Fix doxygen groups for io data sources and sinks ([#13718](https://github.com/NVIDIA/cudf/pull/13718)) [@davidwendt](https://github.com/davidwendt) +- Add pandas compatibility note to DataFrame.query docstring ([#13693](https://github.com/NVIDIA/cudf/pull/13693)) [@beckernick](https://github.com/beckernick) +- Add pylibcudf to developer guide ([#13639](https://github.com/NVIDIA/cudf/pull/13639)) [@vyasr](https://github.com/vyasr) +- Fix repeated words in doxygen text ([#13598](https://github.com/NVIDIA/cudf/pull/13598)) [@karthikeyann](https://github.com/karthikeyann) +- Update docs for top-level API. ([#13592](https://github.com/NVIDIA/cudf/pull/13592)) [@bdice](https://github.com/bdice) +- Fix the the doxygen text for cudf::concatenate and other places ([#13561](https://github.com/NVIDIA/cudf/pull/13561)) [@davidwendt](https://github.com/davidwendt) +- Document stream validation approach used in testing ([#13556](https://github.com/NVIDIA/cudf/pull/13556)) [@vyasr](https://github.com/vyasr) +- Cleanup doc repetitions in libcudf ([#13470](https://github.com/NVIDIA/cudf/pull/13470)) [@karthikeyann](https://github.com/karthikeyann) ## 🚀 New Features -- Support `min` and `max` aggregations for list type in groupby and reduction ([#13676](https://github.com/rapidsai/cudf/pull/13676)) [@ttnghia](https://github.com/ttnghia) -- Add nvtext::jaccard_index API for strings columns ([#13669](https://github.com/rapidsai/cudf/pull/13669)) [@davidwendt](https://github.com/davidwendt) -- Add read_parquet_metadata libcudf API ([#13663](https://github.com/rapidsai/cudf/pull/13663)) [@karthikeyann](https://github.com/karthikeyann) -- Expose streams in all public copying APIs ([#13629](https://github.com/rapidsai/cudf/pull/13629)) [@vyasr](https://github.com/vyasr) -- Add XXHash_64 hash function to cudf ([#13612](https://github.com/rapidsai/cudf/pull/13612)) [@davidwendt](https://github.com/davidwendt) -- Java support: Floating point order-by columns for RANGE window functions ([#13595](https://github.com/rapidsai/cudf/pull/13595)) [@mythrocks](https://github.com/mythrocks) -- Use `cuco::static_map` to build string dictionaries in ORC writer ([#13580](https://github.com/rapidsai/cudf/pull/13580)) [@vuule](https://github.com/vuule) -- Add pylibcudf subpackage with gather implementation ([#13562](https://github.com/rapidsai/cudf/pull/13562)) [@vyasr](https://github.com/vyasr) -- Add JNI for `lists::concatenate_list_elements` ([#13547](https://github.com/rapidsai/cudf/pull/13547)) [@ttnghia](https://github.com/ttnghia) -- Enable nested types for `lists::concatenate_list_elements` ([#13545](https://github.com/rapidsai/cudf/pull/13545)) [@ttnghia](https://github.com/ttnghia) -- Add unicode encoding for string columns in JSON writer ([#13539](https://github.com/rapidsai/cudf/pull/13539)) [@karthikeyann](https://github.com/karthikeyann) -- Remove numba kernels from `find_index_of_val` ([#13517](https://github.com/rapidsai/cudf/pull/13517)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Floating point order-by columns for RANGE window functions ([#13512](https://github.com/rapidsai/cudf/pull/13512)) [@mythrocks](https://github.com/mythrocks) -- Parse column chunk metadata statistics in parquet reader ([#13472](https://github.com/rapidsai/cudf/pull/13472)) [@karthikeyann](https://github.com/karthikeyann) -- Add `abs` function to apply ([#13408](https://github.com/rapidsai/cudf/pull/13408)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- [FEA] AST filtering in parquet reader ([#13348](https://github.com/rapidsai/cudf/pull/13348)) [@karthikeyann](https://github.com/karthikeyann) -- [FEA] Adds option to recover from invalid JSON lines in JSON tokenizer ([#13344](https://github.com/rapidsai/cudf/pull/13344)) [@elstehle](https://github.com/elstehle) -- Ensure cccl packages don't clash with upstream version ([#13235](https://github.com/rapidsai/cudf/pull/13235)) [@robertmaynard](https://github.com/robertmaynard) -- Update `struct_minmax_util` to experimental row comparator ([#13069](https://github.com/rapidsai/cudf/pull/13069)) [@divyegala](https://github.com/divyegala) -- Add stream parameter to hashing APIs ([#12090](https://github.com/rapidsai/cudf/pull/12090)) [@vyasr](https://github.com/vyasr) +- Support `min` and `max` aggregations for list type in groupby and reduction ([#13676](https://github.com/NVIDIA/cudf/pull/13676)) [@ttnghia](https://github.com/ttnghia) +- Add nvtext::jaccard_index API for strings columns ([#13669](https://github.com/NVIDIA/cudf/pull/13669)) [@davidwendt](https://github.com/davidwendt) +- Add read_parquet_metadata libcudf API ([#13663](https://github.com/NVIDIA/cudf/pull/13663)) [@karthikeyann](https://github.com/karthikeyann) +- Expose streams in all public copying APIs ([#13629](https://github.com/NVIDIA/cudf/pull/13629)) [@vyasr](https://github.com/vyasr) +- Add XXHash_64 hash function to cudf ([#13612](https://github.com/NVIDIA/cudf/pull/13612)) [@davidwendt](https://github.com/davidwendt) +- Java support: Floating point order-by columns for RANGE window functions ([#13595](https://github.com/NVIDIA/cudf/pull/13595)) [@mythrocks](https://github.com/mythrocks) +- Use `cuco::static_map` to build string dictionaries in ORC writer ([#13580](https://github.com/NVIDIA/cudf/pull/13580)) [@vuule](https://github.com/vuule) +- Add pylibcudf subpackage with gather implementation ([#13562](https://github.com/NVIDIA/cudf/pull/13562)) [@vyasr](https://github.com/vyasr) +- Add JNI for `lists::concatenate_list_elements` ([#13547](https://github.com/NVIDIA/cudf/pull/13547)) [@ttnghia](https://github.com/ttnghia) +- Enable nested types for `lists::concatenate_list_elements` ([#13545](https://github.com/NVIDIA/cudf/pull/13545)) [@ttnghia](https://github.com/ttnghia) +- Add unicode encoding for string columns in JSON writer ([#13539](https://github.com/NVIDIA/cudf/pull/13539)) [@karthikeyann](https://github.com/karthikeyann) +- Remove numba kernels from `find_index_of_val` ([#13517](https://github.com/NVIDIA/cudf/pull/13517)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Floating point order-by columns for RANGE window functions ([#13512](https://github.com/NVIDIA/cudf/pull/13512)) [@mythrocks](https://github.com/mythrocks) +- Parse column chunk metadata statistics in parquet reader ([#13472](https://github.com/NVIDIA/cudf/pull/13472)) [@karthikeyann](https://github.com/karthikeyann) +- Add `abs` function to apply ([#13408](https://github.com/NVIDIA/cudf/pull/13408)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- [FEA] AST filtering in parquet reader ([#13348](https://github.com/NVIDIA/cudf/pull/13348)) [@karthikeyann](https://github.com/karthikeyann) +- [FEA] Adds option to recover from invalid JSON lines in JSON tokenizer ([#13344](https://github.com/NVIDIA/cudf/pull/13344)) [@elstehle](https://github.com/elstehle) +- Ensure cccl packages don't clash with upstream version ([#13235](https://github.com/NVIDIA/cudf/pull/13235)) [@robertmaynard](https://github.com/robertmaynard) +- Update `struct_minmax_util` to experimental row comparator ([#13069](https://github.com/NVIDIA/cudf/pull/13069)) [@divyegala](https://github.com/divyegala) +- Add stream parameter to hashing APIs ([#12090](https://github.com/NVIDIA/cudf/pull/12090)) [@vyasr](https://github.com/vyasr) ## 🛠️ Improvements -- Pin `dask` and `distributed` for `23.08` release ([#13802](https://github.com/rapidsai/cudf/pull/13802)) [@galipremsagar](https://github.com/galipremsagar) -- Relax protobuf pinnings. ([#13770](https://github.com/rapidsai/cudf/pull/13770)) [@bdice](https://github.com/bdice) -- Switch fully unbounded window functions to use aggregations ([#13727](https://github.com/rapidsai/cudf/pull/13727)) [@mythrocks](https://github.com/mythrocks) -- Switch to new wheel building pipeline ([#13723](https://github.com/rapidsai/cudf/pull/13723)) [@vyasr](https://github.com/vyasr) -- Revert CUDA 12.0 CI workflows to branch-23.08. ([#13719](https://github.com/rapidsai/cudf/pull/13719)) [@bdice](https://github.com/bdice) -- Adding identify minimum version requirement ([#13713](https://github.com/rapidsai/cudf/pull/13713)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Enforce deprecations and add clarifications around existing deprecations ([#13710](https://github.com/rapidsai/cudf/pull/13710)) [@galipremsagar](https://github.com/galipremsagar) -- Optimize ORC reader performance for list data ([#13708](https://github.com/rapidsai/cudf/pull/13708)) [@vyasr](https://github.com/vyasr) -- fix limit overflow message in a docstring ([#13703](https://github.com/rapidsai/cudf/pull/13703)) [@ahmet-uyar](https://github.com/ahmet-uyar) -- Alleviates JSON parser's need for multi-file sources to end with a newline ([#13702](https://github.com/rapidsai/cudf/pull/13702)) [@elstehle](https://github.com/elstehle) -- Update cython-lint and replace flake8 with ruff ([#13699](https://github.com/rapidsai/cudf/pull/13699)) [@vyasr](https://github.com/vyasr) -- Add `__dask_tokenize__` definitions to cudf classes ([#13695](https://github.com/rapidsai/cudf/pull/13695)) [@rjzamora](https://github.com/rjzamora) -- Convert libcudf hashing benchmarks to nvbench ([#13694](https://github.com/rapidsai/cudf/pull/13694)) [@davidwendt](https://github.com/davidwendt) -- Separate MurmurHash32 from hash_functions.cuh ([#13681](https://github.com/rapidsai/cudf/pull/13681)) [@davidwendt](https://github.com/davidwendt) -- Improve performance of cudf::strings::split on whitespace ([#13680](https://github.com/rapidsai/cudf/pull/13680)) [@davidwendt](https://github.com/davidwendt) -- Allow ORC and Parquet writers to write nullable columns without nulls as non-nullable ([#13675](https://github.com/rapidsai/cudf/pull/13675)) [@vuule](https://github.com/vuule) -- Raise a NotImplementedError in to_datetime when utc is passed ([#13670](https://github.com/rapidsai/cudf/pull/13670)) [@shwina](https://github.com/shwina) -- Add rmm_mode parameter to nvbench base fixture ([#13668](https://github.com/rapidsai/cudf/pull/13668)) [@davidwendt](https://github.com/davidwendt) -- Fix multiindex loc ordering in pandas-compat mode ([#13660](https://github.com/rapidsai/cudf/pull/13660)) [@wence-](https://github.com/wence-) -- Add nvtext hash_character_ngrams function ([#13654](https://github.com/rapidsai/cudf/pull/13654)) [@davidwendt](https://github.com/davidwendt) -- Avoid storing metadata in pointers in ORC and Parquet writers ([#13648](https://github.com/rapidsai/cudf/pull/13648)) [@vuule](https://github.com/vuule) -- Acquire spill lock in to/from_arrow ([#13646](https://github.com/rapidsai/cudf/pull/13646)) [@shwina](https://github.com/shwina) -- Expose stable versions of libcudf sort routines ([#13634](https://github.com/rapidsai/cudf/pull/13634)) [@wence-](https://github.com/wence-) -- Separate out hash_test.cpp source for each hash API ([#13633](https://github.com/rapidsai/cudf/pull/13633)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated cudf::strings::slice_strings (by delimiter) functions ([#13628](https://github.com/rapidsai/cudf/pull/13628)) [@davidwendt](https://github.com/davidwendt) -- Create separate libcudf hash APIs for each supported hash function ([#13626](https://github.com/rapidsai/cudf/pull/13626)) [@davidwendt](https://github.com/davidwendt) -- Add convert_dtypes API ([#13623](https://github.com/rapidsai/cudf/pull/13623)) [@shwina](https://github.com/shwina) -- Clean up cupy in dependencies.yaml. ([#13617](https://github.com/rapidsai/cudf/pull/13617)) [@bdice](https://github.com/bdice) -- Use cuda-version to constrain cudatoolkit. ([#13615](https://github.com/rapidsai/cudf/pull/13615)) [@bdice](https://github.com/bdice) -- Add murmurhash3_x64_128 function to libcudf ([#13604](https://github.com/rapidsai/cudf/pull/13604)) [@davidwendt](https://github.com/davidwendt) -- Performance improvement for cudf::strings::like ([#13594](https://github.com/rapidsai/cudf/pull/13594)) [@davidwendt](https://github.com/davidwendt) -- Remove deprecated cudf.set_allocator. ([#13591](https://github.com/rapidsai/cudf/pull/13591)) [@bdice](https://github.com/bdice) -- Clean up cudf device atomic with `cuda::atomic_ref` ([#13583](https://github.com/rapidsai/cudf/pull/13583)) [@PointKernel](https://github.com/PointKernel) -- Add java bindings for distinct count ([#13573](https://github.com/rapidsai/cudf/pull/13573)) [@revans2](https://github.com/revans2) -- Use nvcomp conda package. ([#13566](https://github.com/rapidsai/cudf/pull/13566)) [@bdice](https://github.com/bdice) -- Add exception to string_scalar if input string exceeds size_type ([#13560](https://github.com/rapidsai/cudf/pull/13560)) [@davidwendt](https://github.com/davidwendt) -- Add dispatch for `cudf.Dataframe` to/from `pyarrow.Table` conversion ([#13558](https://github.com/rapidsai/cudf/pull/13558)) [@rjzamora](https://github.com/rjzamora) -- Get rid of `cuco::pair_type` aliases ([#13553](https://github.com/rapidsai/cudf/pull/13553)) [@PointKernel](https://github.com/PointKernel) -- Introduce parity with pandas when `sort=False` in `Groupby` ([#13551](https://github.com/rapidsai/cudf/pull/13551)) [@galipremsagar](https://github.com/galipremsagar) -- Update CMake in docker to 3.26.4 ([#13550](https://github.com/rapidsai/cudf/pull/13550)) [@NvTimLiu](https://github.com/NvTimLiu) -- Clarify source of error message in stream testing. ([#13541](https://github.com/rapidsai/cudf/pull/13541)) [@bdice](https://github.com/bdice) -- Deprecate `strings_to_categorical` in `cudf.read_parquet` ([#13540](https://github.com/rapidsai/cudf/pull/13540)) [@galipremsagar](https://github.com/galipremsagar) -- Update to CMake 3.26.4 ([#13538](https://github.com/rapidsai/cudf/pull/13538)) [@vyasr](https://github.com/vyasr) -- s3 folder naming fix ([#13536](https://github.com/rapidsai/cudf/pull/13536)) [@AyodeAwe](https://github.com/AyodeAwe) -- Implement iloc-getitem using parse-don't-validate approach ([#13534](https://github.com/rapidsai/cudf/pull/13534)) [@wence-](https://github.com/wence-) -- Make synchronization explicit in the names of `hostdevice_*` copying APIs ([#13530](https://github.com/rapidsai/cudf/pull/13530)) [@ttnghia](https://github.com/ttnghia) -- Add benchmark (Google Benchmark) dependency to conda packages. ([#13528](https://github.com/rapidsai/cudf/pull/13528)) [@bdice](https://github.com/bdice) -- Add libcufile to dependencies.yaml. ([#13523](https://github.com/rapidsai/cudf/pull/13523)) [@bdice](https://github.com/bdice) -- Fix some memoization logic in groupby/sort/sort_helper.cu ([#13521](https://github.com/rapidsai/cudf/pull/13521)) [@davidwendt](https://github.com/davidwendt) -- Use sizes_to_offsets_iterator in cudf::gather for strings ([#13520](https://github.com/rapidsai/cudf/pull/13520)) [@davidwendt](https://github.com/davidwendt) -- use rapids-upload-docs script ([#13518](https://github.com/rapidsai/cudf/pull/13518)) [@AyodeAwe](https://github.com/AyodeAwe) -- Support UTF-8 BOM in CSV reader ([#13516](https://github.com/rapidsai/cudf/pull/13516)) [@davidwendt](https://github.com/davidwendt) -- Move stream-related test configuration to CMake ([#13513](https://github.com/rapidsai/cudf/pull/13513)) [@vyasr](https://github.com/vyasr) -- Implement `cudf.option_context` ([#13511](https://github.com/rapidsai/cudf/pull/13511)) [@galipremsagar](https://github.com/galipremsagar) -- Unpin `dask` and `distributed` for development ([#13508](https://github.com/rapidsai/cudf/pull/13508)) [@galipremsagar](https://github.com/galipremsagar) -- Change build.sh to use pip install instead of setup.py ([#13507](https://github.com/rapidsai/cudf/pull/13507)) [@vyasr](https://github.com/vyasr) -- Use test default stream ([#13506](https://github.com/rapidsai/cudf/pull/13506)) [@vyasr](https://github.com/vyasr) -- Remove documentation build scripts for Jenkins ([#13495](https://github.com/rapidsai/cudf/pull/13495)) [@ajschmidt8](https://github.com/ajschmidt8) -- Use east const in include files ([#13494](https://github.com/rapidsai/cudf/pull/13494)) [@karthikeyann](https://github.com/karthikeyann) -- Use east const in src files ([#13493](https://github.com/rapidsai/cudf/pull/13493)) [@karthikeyann](https://github.com/karthikeyann) -- Use east const in tests files ([#13492](https://github.com/rapidsai/cudf/pull/13492)) [@karthikeyann](https://github.com/karthikeyann) -- Use east const in benchmarks files ([#13491](https://github.com/rapidsai/cudf/pull/13491)) [@karthikeyann](https://github.com/karthikeyann) -- Performance improvement for nvtext tokenize/token functions ([#13480](https://github.com/rapidsai/cudf/pull/13480)) [@davidwendt](https://github.com/davidwendt) -- Add pd.Float*Dtype to Avro and ORC mappings ([#13475](https://github.com/rapidsai/cudf/pull/13475)) [@mroeschke](https://github.com/mroeschke) -- Use pandas public APIs where available ([#13467](https://github.com/rapidsai/cudf/pull/13467)) [@mroeschke](https://github.com/mroeschke) -- Allow pd.ArrowDtype in cudf.from_pandas ([#13465](https://github.com/rapidsai/cudf/pull/13465)) [@mroeschke](https://github.com/mroeschke) -- Rework libcudf regex benchmarks with nvbench ([#13464](https://github.com/rapidsai/cudf/pull/13464)) [@davidwendt](https://github.com/davidwendt) -- Remove unused max_rows_tensor parameter from subword tokenizer ([#13463](https://github.com/rapidsai/cudf/pull/13463)) [@davidwendt](https://github.com/davidwendt) -- Separate io-text and nvtext pytests into different files ([#13435](https://github.com/rapidsai/cudf/pull/13435)) [@davidwendt](https://github.com/davidwendt) -- Add a move_to function to cudf::string_view::const_iterator ([#13428](https://github.com/rapidsai/cudf/pull/13428)) [@davidwendt](https://github.com/davidwendt) -- Allow newer scikit-build ([#13424](https://github.com/rapidsai/cudf/pull/13424)) [@vyasr](https://github.com/vyasr) -- Refactor sort_by_values to sort_values, drop indices from return values. ([#13419](https://github.com/rapidsai/cudf/pull/13419)) [@bdice](https://github.com/bdice) -- Inline Cython exception handler ([#13411](https://github.com/rapidsai/cudf/pull/13411)) [@vyasr](https://github.com/vyasr) -- Init JNI version 23.08.0-SNAPSHOT ([#13401](https://github.com/rapidsai/cudf/pull/13401)) [@pxLi](https://github.com/pxLi) -- Refactor ORC reader ([#13396](https://github.com/rapidsai/cudf/pull/13396)) [@ttnghia](https://github.com/ttnghia) -- JNI: Remove cleaned objects in memory cleaner ([#13378](https://github.com/rapidsai/cudf/pull/13378)) [@res-life](https://github.com/res-life) -- Add tests of currently unsupported indexing ([#13338](https://github.com/rapidsai/cudf/pull/13338)) [@wence-](https://github.com/wence-) -- Performance improvement for some libcudf regex functions for long strings ([#13322](https://github.com/rapidsai/cudf/pull/13322)) [@davidwendt](https://github.com/davidwendt) -- Exposure Tracked Buffer (first step towards unifying copy-on-write and spilling) ([#13307](https://github.com/rapidsai/cudf/pull/13307)) [@madsbk](https://github.com/madsbk) -- Write string data directly to column_buffer in Parquet reader ([#13302](https://github.com/rapidsai/cudf/pull/13302)) [@etseidl](https://github.com/etseidl) -- Add stacktrace into cudf exception types ([#13298](https://github.com/rapidsai/cudf/pull/13298)) [@ttnghia](https://github.com/ttnghia) -- cuDF: Build CUDA 12 packages ([#12922](https://github.com/rapidsai/cudf/pull/12922)) [@bdice](https://github.com/bdice) +- Pin `dask` and `distributed` for `23.08` release ([#13802](https://github.com/NVIDIA/cudf/pull/13802)) [@galipremsagar](https://github.com/galipremsagar) +- Relax protobuf pinnings. ([#13770](https://github.com/NVIDIA/cudf/pull/13770)) [@bdice](https://github.com/bdice) +- Switch fully unbounded window functions to use aggregations ([#13727](https://github.com/NVIDIA/cudf/pull/13727)) [@mythrocks](https://github.com/mythrocks) +- Switch to new wheel building pipeline ([#13723](https://github.com/NVIDIA/cudf/pull/13723)) [@vyasr](https://github.com/vyasr) +- Revert CUDA 12.0 CI workflows to branch-23.08. ([#13719](https://github.com/NVIDIA/cudf/pull/13719)) [@bdice](https://github.com/bdice) +- Adding identify minimum version requirement ([#13713](https://github.com/NVIDIA/cudf/pull/13713)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Enforce deprecations and add clarifications around existing deprecations ([#13710](https://github.com/NVIDIA/cudf/pull/13710)) [@galipremsagar](https://github.com/galipremsagar) +- Optimize ORC reader performance for list data ([#13708](https://github.com/NVIDIA/cudf/pull/13708)) [@vyasr](https://github.com/vyasr) +- fix limit overflow message in a docstring ([#13703](https://github.com/NVIDIA/cudf/pull/13703)) [@ahmet-uyar](https://github.com/ahmet-uyar) +- Alleviates JSON parser's need for multi-file sources to end with a newline ([#13702](https://github.com/NVIDIA/cudf/pull/13702)) [@elstehle](https://github.com/elstehle) +- Update cython-lint and replace flake8 with ruff ([#13699](https://github.com/NVIDIA/cudf/pull/13699)) [@vyasr](https://github.com/vyasr) +- Add `__dask_tokenize__` definitions to cudf classes ([#13695](https://github.com/NVIDIA/cudf/pull/13695)) [@rjzamora](https://github.com/rjzamora) +- Convert libcudf hashing benchmarks to nvbench ([#13694](https://github.com/NVIDIA/cudf/pull/13694)) [@davidwendt](https://github.com/davidwendt) +- Separate MurmurHash32 from hash_functions.cuh ([#13681](https://github.com/NVIDIA/cudf/pull/13681)) [@davidwendt](https://github.com/davidwendt) +- Improve performance of cudf::strings::split on whitespace ([#13680](https://github.com/NVIDIA/cudf/pull/13680)) [@davidwendt](https://github.com/davidwendt) +- Allow ORC and Parquet writers to write nullable columns without nulls as non-nullable ([#13675](https://github.com/NVIDIA/cudf/pull/13675)) [@vuule](https://github.com/vuule) +- Raise a NotImplementedError in to_datetime when utc is passed ([#13670](https://github.com/NVIDIA/cudf/pull/13670)) [@shwina](https://github.com/shwina) +- Add rmm_mode parameter to nvbench base fixture ([#13668](https://github.com/NVIDIA/cudf/pull/13668)) [@davidwendt](https://github.com/davidwendt) +- Fix multiindex loc ordering in pandas-compat mode ([#13660](https://github.com/NVIDIA/cudf/pull/13660)) [@wence-](https://github.com/wence-) +- Add nvtext hash_character_ngrams function ([#13654](https://github.com/NVIDIA/cudf/pull/13654)) [@davidwendt](https://github.com/davidwendt) +- Avoid storing metadata in pointers in ORC and Parquet writers ([#13648](https://github.com/NVIDIA/cudf/pull/13648)) [@vuule](https://github.com/vuule) +- Acquire spill lock in to/from_arrow ([#13646](https://github.com/NVIDIA/cudf/pull/13646)) [@shwina](https://github.com/shwina) +- Expose stable versions of libcudf sort routines ([#13634](https://github.com/NVIDIA/cudf/pull/13634)) [@wence-](https://github.com/wence-) +- Separate out hash_test.cpp source for each hash API ([#13633](https://github.com/NVIDIA/cudf/pull/13633)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated cudf::strings::slice_strings (by delimiter) functions ([#13628](https://github.com/NVIDIA/cudf/pull/13628)) [@davidwendt](https://github.com/davidwendt) +- Create separate libcudf hash APIs for each supported hash function ([#13626](https://github.com/NVIDIA/cudf/pull/13626)) [@davidwendt](https://github.com/davidwendt) +- Add convert_dtypes API ([#13623](https://github.com/NVIDIA/cudf/pull/13623)) [@shwina](https://github.com/shwina) +- Clean up cupy in dependencies.yaml. ([#13617](https://github.com/NVIDIA/cudf/pull/13617)) [@bdice](https://github.com/bdice) +- Use cuda-version to constrain cudatoolkit. ([#13615](https://github.com/NVIDIA/cudf/pull/13615)) [@bdice](https://github.com/bdice) +- Add murmurhash3_x64_128 function to libcudf ([#13604](https://github.com/NVIDIA/cudf/pull/13604)) [@davidwendt](https://github.com/davidwendt) +- Performance improvement for cudf::strings::like ([#13594](https://github.com/NVIDIA/cudf/pull/13594)) [@davidwendt](https://github.com/davidwendt) +- Remove deprecated cudf.set_allocator. ([#13591](https://github.com/NVIDIA/cudf/pull/13591)) [@bdice](https://github.com/bdice) +- Clean up cudf device atomic with `cuda::atomic_ref` ([#13583](https://github.com/NVIDIA/cudf/pull/13583)) [@PointKernel](https://github.com/PointKernel) +- Add java bindings for distinct count ([#13573](https://github.com/NVIDIA/cudf/pull/13573)) [@revans2](https://github.com/revans2) +- Use nvcomp conda package. ([#13566](https://github.com/NVIDIA/cudf/pull/13566)) [@bdice](https://github.com/bdice) +- Add exception to string_scalar if input string exceeds size_type ([#13560](https://github.com/NVIDIA/cudf/pull/13560)) [@davidwendt](https://github.com/davidwendt) +- Add dispatch for `cudf.Dataframe` to/from `pyarrow.Table` conversion ([#13558](https://github.com/NVIDIA/cudf/pull/13558)) [@rjzamora](https://github.com/rjzamora) +- Get rid of `cuco::pair_type` aliases ([#13553](https://github.com/NVIDIA/cudf/pull/13553)) [@PointKernel](https://github.com/PointKernel) +- Introduce parity with pandas when `sort=False` in `Groupby` ([#13551](https://github.com/NVIDIA/cudf/pull/13551)) [@galipremsagar](https://github.com/galipremsagar) +- Update CMake in docker to 3.26.4 ([#13550](https://github.com/NVIDIA/cudf/pull/13550)) [@NvTimLiu](https://github.com/NvTimLiu) +- Clarify source of error message in stream testing. ([#13541](https://github.com/NVIDIA/cudf/pull/13541)) [@bdice](https://github.com/bdice) +- Deprecate `strings_to_categorical` in `cudf.read_parquet` ([#13540](https://github.com/NVIDIA/cudf/pull/13540)) [@galipremsagar](https://github.com/galipremsagar) +- Update to CMake 3.26.4 ([#13538](https://github.com/NVIDIA/cudf/pull/13538)) [@vyasr](https://github.com/vyasr) +- s3 folder naming fix ([#13536](https://github.com/NVIDIA/cudf/pull/13536)) [@AyodeAwe](https://github.com/AyodeAwe) +- Implement iloc-getitem using parse-don't-validate approach ([#13534](https://github.com/NVIDIA/cudf/pull/13534)) [@wence-](https://github.com/wence-) +- Make synchronization explicit in the names of `hostdevice_*` copying APIs ([#13530](https://github.com/NVIDIA/cudf/pull/13530)) [@ttnghia](https://github.com/ttnghia) +- Add benchmark (Google Benchmark) dependency to conda packages. ([#13528](https://github.com/NVIDIA/cudf/pull/13528)) [@bdice](https://github.com/bdice) +- Add libcufile to dependencies.yaml. ([#13523](https://github.com/NVIDIA/cudf/pull/13523)) [@bdice](https://github.com/bdice) +- Fix some memoization logic in groupby/sort/sort_helper.cu ([#13521](https://github.com/NVIDIA/cudf/pull/13521)) [@davidwendt](https://github.com/davidwendt) +- Use sizes_to_offsets_iterator in cudf::gather for strings ([#13520](https://github.com/NVIDIA/cudf/pull/13520)) [@davidwendt](https://github.com/davidwendt) +- use rapids-upload-docs script ([#13518](https://github.com/NVIDIA/cudf/pull/13518)) [@AyodeAwe](https://github.com/AyodeAwe) +- Support UTF-8 BOM in CSV reader ([#13516](https://github.com/NVIDIA/cudf/pull/13516)) [@davidwendt](https://github.com/davidwendt) +- Move stream-related test configuration to CMake ([#13513](https://github.com/NVIDIA/cudf/pull/13513)) [@vyasr](https://github.com/vyasr) +- Implement `cudf.option_context` ([#13511](https://github.com/NVIDIA/cudf/pull/13511)) [@galipremsagar](https://github.com/galipremsagar) +- Unpin `dask` and `distributed` for development ([#13508](https://github.com/NVIDIA/cudf/pull/13508)) [@galipremsagar](https://github.com/galipremsagar) +- Change build.sh to use pip install instead of setup.py ([#13507](https://github.com/NVIDIA/cudf/pull/13507)) [@vyasr](https://github.com/vyasr) +- Use test default stream ([#13506](https://github.com/NVIDIA/cudf/pull/13506)) [@vyasr](https://github.com/vyasr) +- Remove documentation build scripts for Jenkins ([#13495](https://github.com/NVIDIA/cudf/pull/13495)) [@ajschmidt8](https://github.com/ajschmidt8) +- Use east const in include files ([#13494](https://github.com/NVIDIA/cudf/pull/13494)) [@karthikeyann](https://github.com/karthikeyann) +- Use east const in src files ([#13493](https://github.com/NVIDIA/cudf/pull/13493)) [@karthikeyann](https://github.com/karthikeyann) +- Use east const in tests files ([#13492](https://github.com/NVIDIA/cudf/pull/13492)) [@karthikeyann](https://github.com/karthikeyann) +- Use east const in benchmarks files ([#13491](https://github.com/NVIDIA/cudf/pull/13491)) [@karthikeyann](https://github.com/karthikeyann) +- Performance improvement for nvtext tokenize/token functions ([#13480](https://github.com/NVIDIA/cudf/pull/13480)) [@davidwendt](https://github.com/davidwendt) +- Add pd.Float*Dtype to Avro and ORC mappings ([#13475](https://github.com/NVIDIA/cudf/pull/13475)) [@mroeschke](https://github.com/mroeschke) +- Use pandas public APIs where available ([#13467](https://github.com/NVIDIA/cudf/pull/13467)) [@mroeschke](https://github.com/mroeschke) +- Allow pd.ArrowDtype in cudf.from_pandas ([#13465](https://github.com/NVIDIA/cudf/pull/13465)) [@mroeschke](https://github.com/mroeschke) +- Rework libcudf regex benchmarks with nvbench ([#13464](https://github.com/NVIDIA/cudf/pull/13464)) [@davidwendt](https://github.com/davidwendt) +- Remove unused max_rows_tensor parameter from subword tokenizer ([#13463](https://github.com/NVIDIA/cudf/pull/13463)) [@davidwendt](https://github.com/davidwendt) +- Separate io-text and nvtext pytests into different files ([#13435](https://github.com/NVIDIA/cudf/pull/13435)) [@davidwendt](https://github.com/davidwendt) +- Add a move_to function to cudf::string_view::const_iterator ([#13428](https://github.com/NVIDIA/cudf/pull/13428)) [@davidwendt](https://github.com/davidwendt) +- Allow newer scikit-build ([#13424](https://github.com/NVIDIA/cudf/pull/13424)) [@vyasr](https://github.com/vyasr) +- Refactor sort_by_values to sort_values, drop indices from return values. ([#13419](https://github.com/NVIDIA/cudf/pull/13419)) [@bdice](https://github.com/bdice) +- Inline Cython exception handler ([#13411](https://github.com/NVIDIA/cudf/pull/13411)) [@vyasr](https://github.com/vyasr) +- Init JNI version 23.08.0-SNAPSHOT ([#13401](https://github.com/NVIDIA/cudf/pull/13401)) [@pxLi](https://github.com/pxLi) +- Refactor ORC reader ([#13396](https://github.com/NVIDIA/cudf/pull/13396)) [@ttnghia](https://github.com/ttnghia) +- JNI: Remove cleaned objects in memory cleaner ([#13378](https://github.com/NVIDIA/cudf/pull/13378)) [@res-life](https://github.com/res-life) +- Add tests of currently unsupported indexing ([#13338](https://github.com/NVIDIA/cudf/pull/13338)) [@wence-](https://github.com/wence-) +- Performance improvement for some libcudf regex functions for long strings ([#13322](https://github.com/NVIDIA/cudf/pull/13322)) [@davidwendt](https://github.com/davidwendt) +- Exposure Tracked Buffer (first step towards unifying copy-on-write and spilling) ([#13307](https://github.com/NVIDIA/cudf/pull/13307)) [@madsbk](https://github.com/madsbk) +- Write string data directly to column_buffer in Parquet reader ([#13302](https://github.com/NVIDIA/cudf/pull/13302)) [@etseidl](https://github.com/etseidl) +- Add stacktrace into cudf exception types ([#13298](https://github.com/NVIDIA/cudf/pull/13298)) [@ttnghia](https://github.com/ttnghia) +- cuDF: Build CUDA 12 packages ([#12922](https://github.com/NVIDIA/cudf/pull/12922)) [@bdice](https://github.com/bdice) # cuDF 23.06.00 (7 Jun 2023) ## 🚨 Breaking Changes -- Fix batch processing for parquet writer ([#13438](https://github.com/rapidsai/cudf/pull/13438)) [@ttnghia](https://github.com/ttnghia) -- Use <NA> instead of null to match pandas. ([#13415](https://github.com/rapidsai/cudf/pull/13415)) [@bdice](https://github.com/bdice) -- Remove UNKNOWN_NULL_COUNT ([#13372](https://github.com/rapidsai/cudf/pull/13372)) [@vyasr](https://github.com/vyasr) -- Remove default UNKNOWN_NULL_COUNT from cudf::column member functions ([#13341](https://github.com/rapidsai/cudf/pull/13341)) [@davidwendt](https://github.com/davidwendt) -- Use std::overflow_error when output would exceed column size limit ([#13323](https://github.com/rapidsai/cudf/pull/13323)) [@davidwendt](https://github.com/davidwendt) -- Remove null mask and null count from column_view constructors ([#13311](https://github.com/rapidsai/cudf/pull/13311)) [@vyasr](https://github.com/vyasr) -- Change default value of the `observed=` argument in groupby to `True` to reflect the actual behaviour ([#13296](https://github.com/rapidsai/cudf/pull/13296)) [@shwina](https://github.com/shwina) -- Throw error if UNINITIALIZED is passed to cudf::state_null_count ([#13292](https://github.com/rapidsai/cudf/pull/13292)) [@davidwendt](https://github.com/davidwendt) -- Remove default null-count parameter from cudf::make_strings_column factory ([#13227](https://github.com/rapidsai/cudf/pull/13227)) [@davidwendt](https://github.com/davidwendt) -- Remove UNKNOWN_NULL_COUNT where it can be easily computed ([#13205](https://github.com/rapidsai/cudf/pull/13205)) [@vyasr](https://github.com/vyasr) -- Update minimum Python version to Python 3.9 ([#13196](https://github.com/rapidsai/cudf/pull/13196)) [@shwina](https://github.com/shwina) -- Refactor contiguous_split API into contiguous_split.hpp ([#13186](https://github.com/rapidsai/cudf/pull/13186)) [@abellina](https://github.com/abellina) -- Cleanup Parquet chunked writer ([#13094](https://github.com/rapidsai/cudf/pull/13094)) [@ttnghia](https://github.com/ttnghia) -- Cleanup ORC chunked writer ([#13091](https://github.com/rapidsai/cudf/pull/13091)) [@ttnghia](https://github.com/ttnghia) -- Raise `NotImplementedError` when attempting to construct cuDF objects from timezone-aware datetimes ([#13086](https://github.com/rapidsai/cudf/pull/13086)) [@shwina](https://github.com/shwina) -- Remove deprecated regex functions from libcudf ([#13067](https://github.com/rapidsai/cudf/pull/13067)) [@davidwendt](https://github.com/davidwendt) -- [REVIEW] Upgrade to `arrow-11` ([#12757](https://github.com/rapidsai/cudf/pull/12757)) [@galipremsagar](https://github.com/galipremsagar) -- Implement Python drop_duplicates with cudf::stable_distinct. ([#11656](https://github.com/rapidsai/cudf/pull/11656)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix batch processing for parquet writer ([#13438](https://github.com/NVIDIA/cudf/pull/13438)) [@ttnghia](https://github.com/ttnghia) +- Use <NA> instead of null to match pandas. ([#13415](https://github.com/NVIDIA/cudf/pull/13415)) [@bdice](https://github.com/bdice) +- Remove UNKNOWN_NULL_COUNT ([#13372](https://github.com/NVIDIA/cudf/pull/13372)) [@vyasr](https://github.com/vyasr) +- Remove default UNKNOWN_NULL_COUNT from cudf::column member functions ([#13341](https://github.com/NVIDIA/cudf/pull/13341)) [@davidwendt](https://github.com/davidwendt) +- Use std::overflow_error when output would exceed column size limit ([#13323](https://github.com/NVIDIA/cudf/pull/13323)) [@davidwendt](https://github.com/davidwendt) +- Remove null mask and null count from column_view constructors ([#13311](https://github.com/NVIDIA/cudf/pull/13311)) [@vyasr](https://github.com/vyasr) +- Change default value of the `observed=` argument in groupby to `True` to reflect the actual behaviour ([#13296](https://github.com/NVIDIA/cudf/pull/13296)) [@shwina](https://github.com/shwina) +- Throw error if UNINITIALIZED is passed to cudf::state_null_count ([#13292](https://github.com/NVIDIA/cudf/pull/13292)) [@davidwendt](https://github.com/davidwendt) +- Remove default null-count parameter from cudf::make_strings_column factory ([#13227](https://github.com/NVIDIA/cudf/pull/13227)) [@davidwendt](https://github.com/davidwendt) +- Remove UNKNOWN_NULL_COUNT where it can be easily computed ([#13205](https://github.com/NVIDIA/cudf/pull/13205)) [@vyasr](https://github.com/vyasr) +- Update minimum Python version to Python 3.9 ([#13196](https://github.com/NVIDIA/cudf/pull/13196)) [@shwina](https://github.com/shwina) +- Refactor contiguous_split API into contiguous_split.hpp ([#13186](https://github.com/NVIDIA/cudf/pull/13186)) [@abellina](https://github.com/abellina) +- Cleanup Parquet chunked writer ([#13094](https://github.com/NVIDIA/cudf/pull/13094)) [@ttnghia](https://github.com/ttnghia) +- Cleanup ORC chunked writer ([#13091](https://github.com/NVIDIA/cudf/pull/13091)) [@ttnghia](https://github.com/ttnghia) +- Raise `NotImplementedError` when attempting to construct cuDF objects from timezone-aware datetimes ([#13086](https://github.com/NVIDIA/cudf/pull/13086)) [@shwina](https://github.com/shwina) +- Remove deprecated regex functions from libcudf ([#13067](https://github.com/NVIDIA/cudf/pull/13067)) [@davidwendt](https://github.com/davidwendt) +- [REVIEW] Upgrade to `arrow-11` ([#12757](https://github.com/NVIDIA/cudf/pull/12757)) [@galipremsagar](https://github.com/galipremsagar) +- Implement Python drop_duplicates with cudf::stable_distinct. ([#11656](https://github.com/NVIDIA/cudf/pull/11656)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🐛 Bug Fixes -- Fix valid count computation in offset_bitmask_binop kernel ([#13489](https://github.com/rapidsai/cudf/pull/13489)) [@davidwendt](https://github.com/davidwendt) -- Fix writing of ORC files with empty rowgroups ([#13466](https://github.com/rapidsai/cudf/pull/13466)) [@vuule](https://github.com/vuule) -- Fix cudf::repeat logic when count is zero ([#13459](https://github.com/rapidsai/cudf/pull/13459)) [@davidwendt](https://github.com/davidwendt) -- Fix batch processing for parquet writer ([#13438](https://github.com/rapidsai/cudf/pull/13438)) [@ttnghia](https://github.com/ttnghia) -- Fix invalid use of std::exclusive_scan in Parquet writer ([#13434](https://github.com/rapidsai/cudf/pull/13434)) [@etseidl](https://github.com/etseidl) -- Patch numba if it is imported first to ensure minor version compatibility works. ([#13433](https://github.com/rapidsai/cudf/pull/13433)) [@bdice](https://github.com/bdice) -- Fix cudf::strings::replace_with_backrefs hang on empty match result ([#13418](https://github.com/rapidsai/cudf/pull/13418)) [@davidwendt](https://github.com/davidwendt) -- Use <NA> instead of null to match pandas. ([#13415](https://github.com/rapidsai/cudf/pull/13415)) [@bdice](https://github.com/bdice) -- Fix tokenize with non-space delimiter ([#13403](https://github.com/rapidsai/cudf/pull/13403)) [@shwina](https://github.com/shwina) -- Fix groupby head/tail for empty dataframe ([#13398](https://github.com/rapidsai/cudf/pull/13398)) [@shwina](https://github.com/shwina) -- Default to closed="right" in `IntervalIndex` constructor ([#13394](https://github.com/rapidsai/cudf/pull/13394)) [@shwina](https://github.com/shwina) -- Correctly reorder and reindex scan groupbys with null keys ([#13389](https://github.com/rapidsai/cudf/pull/13389)) [@wence-](https://github.com/wence-) -- Fix unused argument errors in nvcc 11.5 ([#13387](https://github.com/rapidsai/cudf/pull/13387)) [@abellina](https://github.com/abellina) -- Updates needed to work with jitify that leverages libcudacxx ([#13383](https://github.com/rapidsai/cudf/pull/13383)) [@robertmaynard](https://github.com/robertmaynard) -- Fix unused parameter warning/error in parquet/page_data.cu ([#13367](https://github.com/rapidsai/cudf/pull/13367)) [@davidwendt](https://github.com/davidwendt) -- Fix page size estimation in Parquet writer ([#13364](https://github.com/rapidsai/cudf/pull/13364)) [@etseidl](https://github.com/etseidl) -- Fix subword_tokenize error when input contains no tokens ([#13320](https://github.com/rapidsai/cudf/pull/13320)) [@davidwendt](https://github.com/davidwendt) -- Support gcc 12 as the C++ compiler ([#13316](https://github.com/rapidsai/cudf/pull/13316)) [@robertmaynard](https://github.com/robertmaynard) -- Correctly set bitmask size in `from_column_view` ([#13315](https://github.com/rapidsai/cudf/pull/13315)) [@wence-](https://github.com/wence-) -- Fix approach to detecting assignment for gte/lte operators ([#13285](https://github.com/rapidsai/cudf/pull/13285)) [@vyasr](https://github.com/vyasr) -- Fix parquet schema interpretation issue ([#13277](https://github.com/rapidsai/cudf/pull/13277)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix 64bit shift bug in avro reader ([#13276](https://github.com/rapidsai/cudf/pull/13276)) [@karthikeyann](https://github.com/karthikeyann) -- Fix unused variables/parameters in parquet/writer_impl.cu ([#13263](https://github.com/rapidsai/cudf/pull/13263)) [@davidwendt](https://github.com/davidwendt) -- Clean up buffers in case AssertionError ([#13262](https://github.com/rapidsai/cudf/pull/13262)) [@razajafri](https://github.com/razajafri) -- Allow empty input table in ast `compute_column` ([#13245](https://github.com/rapidsai/cudf/pull/13245)) [@wence-](https://github.com/wence-) -- Fix structs_column_wrapper constructors to copy input column wrappers ([#13243](https://github.com/rapidsai/cudf/pull/13243)) [@davidwendt](https://github.com/davidwendt) -- Fix the row index stream order in ORC reader ([#13242](https://github.com/rapidsai/cudf/pull/13242)) [@vuule](https://github.com/vuule) -- Make `is_decompression_disabled` and `is_compression_disabled` thread-safe ([#13240](https://github.com/rapidsai/cudf/pull/13240)) [@vuule](https://github.com/vuule) -- Add [[maybe_unused]] to nvbench environment. ([#13219](https://github.com/rapidsai/cudf/pull/13219)) [@bdice](https://github.com/bdice) -- Fix race in ORC string dictionary creation ([#13214](https://github.com/rapidsai/cudf/pull/13214)) [@revans2](https://github.com/revans2) -- Add scalar argtypes to udf cache keys ([#13194](https://github.com/rapidsai/cudf/pull/13194)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix unused parameter warning/error in grouped_rolling.cu ([#13192](https://github.com/rapidsai/cudf/pull/13192)) [@davidwendt](https://github.com/davidwendt) -- Avoid skbuild 0.17.2 which affected the cmake -DPython_LIBRARY string ([#13188](https://github.com/rapidsai/cudf/pull/13188)) [@sevagh](https://github.com/sevagh) -- Fix `hostdevice_vector::subspan` ([#13187](https://github.com/rapidsai/cudf/pull/13187)) [@ttnghia](https://github.com/ttnghia) -- Use custom nvbench entry point to ensure `cudf::nvbench_base_fixture` usage ([#13183](https://github.com/rapidsai/cudf/pull/13183)) [@robertmaynard](https://github.com/robertmaynard) -- Fix slice_strings to return empty strings for stop < start indices ([#13178](https://github.com/rapidsai/cudf/pull/13178)) [@davidwendt](https://github.com/davidwendt) -- Allow compilation with any GTest version 1.11+ ([#13153](https://github.com/rapidsai/cudf/pull/13153)) [@robertmaynard](https://github.com/robertmaynard) -- Fix a few clang-format style check errors ([#13146](https://github.com/rapidsai/cudf/pull/13146)) [@davidwendt](https://github.com/davidwendt) -- [REVIEW] Fix `Series` and `DataFrame` constructors to validate index lengths ([#13122](https://github.com/rapidsai/cudf/pull/13122)) [@galipremsagar](https://github.com/galipremsagar) -- Fix hash join when the input tables have nulls on only one side ([#13120](https://github.com/rapidsai/cudf/pull/13120)) [@ttnghia](https://github.com/ttnghia) -- Fix GPU_ARCHS setting in Java CMake build and CMAKE_CUDA_ARCHITECTURES in Python package build. ([#13117](https://github.com/rapidsai/cudf/pull/13117)) [@davidwendt](https://github.com/davidwendt) -- Adds checks to make sure json reader won't overflow ([#13115](https://github.com/rapidsai/cudf/pull/13115)) [@elstehle](https://github.com/elstehle) -- Fix `null_count` of columns returned by `chunked_parquet_reader` ([#13111](https://github.com/rapidsai/cudf/pull/13111)) [@vuule](https://github.com/vuule) -- Fixes sliced list and struct column bug in JSON chunked writer ([#13108](https://github.com/rapidsai/cudf/pull/13108)) [@karthikeyann](https://github.com/karthikeyann) -- [REVIEW] Fix missing confluent kafka version ([#13101](https://github.com/rapidsai/cudf/pull/13101)) [@galipremsagar](https://github.com/galipremsagar) -- Use make_empty_lists_column instead of make_empty_column(type_id::LIST) ([#13099](https://github.com/rapidsai/cudf/pull/13099)) [@davidwendt](https://github.com/davidwendt) -- Raise `NotImplementedError` when attempting to construct cuDF objects from timezone-aware datetimes ([#13086](https://github.com/rapidsai/cudf/pull/13086)) [@shwina](https://github.com/shwina) -- Fix column selection `read_parquet` benchmarks ([#13082](https://github.com/rapidsai/cudf/pull/13082)) [@vuule](https://github.com/vuule) -- Fix bugs in iterative groupby apply algorithm ([#13078](https://github.com/rapidsai/cudf/pull/13078)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add algorithm include in data_sink.hpp ([#13068](https://github.com/rapidsai/cudf/pull/13068)) [@ahendriksen](https://github.com/ahendriksen) -- Fix tests/identify_stream_usage.cpp ([#13066](https://github.com/rapidsai/cudf/pull/13066)) [@ahendriksen](https://github.com/ahendriksen) -- Prevent overflow with `skip_rows` in ORC and Parquet readers ([#13063](https://github.com/rapidsai/cudf/pull/13063)) [@vuule](https://github.com/vuule) -- Add except declaration in Cython interface for regex_program::create ([#13054](https://github.com/rapidsai/cudf/pull/13054)) [@davidwendt](https://github.com/davidwendt) -- [REVIEW] Fix branch version in CI scripts ([#13029](https://github.com/rapidsai/cudf/pull/13029)) [@galipremsagar](https://github.com/galipremsagar) -- Fix OOB memory access in CSV reader when reading without NA values ([#13011](https://github.com/rapidsai/cudf/pull/13011)) [@vuule](https://github.com/vuule) -- Fix read_avro() skip_rows and num_rows. ([#12912](https://github.com/rapidsai/cudf/pull/12912)) [@tpn](https://github.com/tpn) -- Purge nonempty nulls from byte_cast list outputs. ([#11971](https://github.com/rapidsai/cudf/pull/11971)) [@bdice](https://github.com/bdice) -- Fix consumption of CPU-backed interchange protocol dataframes ([#11392](https://github.com/rapidsai/cudf/pull/11392)) [@shwina](https://github.com/shwina) +- Fix valid count computation in offset_bitmask_binop kernel ([#13489](https://github.com/NVIDIA/cudf/pull/13489)) [@davidwendt](https://github.com/davidwendt) +- Fix writing of ORC files with empty rowgroups ([#13466](https://github.com/NVIDIA/cudf/pull/13466)) [@vuule](https://github.com/vuule) +- Fix cudf::repeat logic when count is zero ([#13459](https://github.com/NVIDIA/cudf/pull/13459)) [@davidwendt](https://github.com/davidwendt) +- Fix batch processing for parquet writer ([#13438](https://github.com/NVIDIA/cudf/pull/13438)) [@ttnghia](https://github.com/ttnghia) +- Fix invalid use of std::exclusive_scan in Parquet writer ([#13434](https://github.com/NVIDIA/cudf/pull/13434)) [@etseidl](https://github.com/etseidl) +- Patch numba if it is imported first to ensure minor version compatibility works. ([#13433](https://github.com/NVIDIA/cudf/pull/13433)) [@bdice](https://github.com/bdice) +- Fix cudf::strings::replace_with_backrefs hang on empty match result ([#13418](https://github.com/NVIDIA/cudf/pull/13418)) [@davidwendt](https://github.com/davidwendt) +- Use <NA> instead of null to match pandas. ([#13415](https://github.com/NVIDIA/cudf/pull/13415)) [@bdice](https://github.com/bdice) +- Fix tokenize with non-space delimiter ([#13403](https://github.com/NVIDIA/cudf/pull/13403)) [@shwina](https://github.com/shwina) +- Fix groupby head/tail for empty dataframe ([#13398](https://github.com/NVIDIA/cudf/pull/13398)) [@shwina](https://github.com/shwina) +- Default to closed="right" in `IntervalIndex` constructor ([#13394](https://github.com/NVIDIA/cudf/pull/13394)) [@shwina](https://github.com/shwina) +- Correctly reorder and reindex scan groupbys with null keys ([#13389](https://github.com/NVIDIA/cudf/pull/13389)) [@wence-](https://github.com/wence-) +- Fix unused argument errors in nvcc 11.5 ([#13387](https://github.com/NVIDIA/cudf/pull/13387)) [@abellina](https://github.com/abellina) +- Updates needed to work with jitify that leverages libcudacxx ([#13383](https://github.com/NVIDIA/cudf/pull/13383)) [@robertmaynard](https://github.com/robertmaynard) +- Fix unused parameter warning/error in parquet/page_data.cu ([#13367](https://github.com/NVIDIA/cudf/pull/13367)) [@davidwendt](https://github.com/davidwendt) +- Fix page size estimation in Parquet writer ([#13364](https://github.com/NVIDIA/cudf/pull/13364)) [@etseidl](https://github.com/etseidl) +- Fix subword_tokenize error when input contains no tokens ([#13320](https://github.com/NVIDIA/cudf/pull/13320)) [@davidwendt](https://github.com/davidwendt) +- Support gcc 12 as the C++ compiler ([#13316](https://github.com/NVIDIA/cudf/pull/13316)) [@robertmaynard](https://github.com/robertmaynard) +- Correctly set bitmask size in `from_column_view` ([#13315](https://github.com/NVIDIA/cudf/pull/13315)) [@wence-](https://github.com/wence-) +- Fix approach to detecting assignment for gte/lte operators ([#13285](https://github.com/NVIDIA/cudf/pull/13285)) [@vyasr](https://github.com/vyasr) +- Fix parquet schema interpretation issue ([#13277](https://github.com/NVIDIA/cudf/pull/13277)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix 64bit shift bug in avro reader ([#13276](https://github.com/NVIDIA/cudf/pull/13276)) [@karthikeyann](https://github.com/karthikeyann) +- Fix unused variables/parameters in parquet/writer_impl.cu ([#13263](https://github.com/NVIDIA/cudf/pull/13263)) [@davidwendt](https://github.com/davidwendt) +- Clean up buffers in case AssertionError ([#13262](https://github.com/NVIDIA/cudf/pull/13262)) [@razajafri](https://github.com/razajafri) +- Allow empty input table in ast `compute_column` ([#13245](https://github.com/NVIDIA/cudf/pull/13245)) [@wence-](https://github.com/wence-) +- Fix structs_column_wrapper constructors to copy input column wrappers ([#13243](https://github.com/NVIDIA/cudf/pull/13243)) [@davidwendt](https://github.com/davidwendt) +- Fix the row index stream order in ORC reader ([#13242](https://github.com/NVIDIA/cudf/pull/13242)) [@vuule](https://github.com/vuule) +- Make `is_decompression_disabled` and `is_compression_disabled` thread-safe ([#13240](https://github.com/NVIDIA/cudf/pull/13240)) [@vuule](https://github.com/vuule) +- Add [[maybe_unused]] to nvbench environment. ([#13219](https://github.com/NVIDIA/cudf/pull/13219)) [@bdice](https://github.com/bdice) +- Fix race in ORC string dictionary creation ([#13214](https://github.com/NVIDIA/cudf/pull/13214)) [@revans2](https://github.com/revans2) +- Add scalar argtypes to udf cache keys ([#13194](https://github.com/NVIDIA/cudf/pull/13194)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix unused parameter warning/error in grouped_rolling.cu ([#13192](https://github.com/NVIDIA/cudf/pull/13192)) [@davidwendt](https://github.com/davidwendt) +- Avoid skbuild 0.17.2 which affected the cmake -DPython_LIBRARY string ([#13188](https://github.com/NVIDIA/cudf/pull/13188)) [@sevagh](https://github.com/sevagh) +- Fix `hostdevice_vector::subspan` ([#13187](https://github.com/NVIDIA/cudf/pull/13187)) [@ttnghia](https://github.com/ttnghia) +- Use custom nvbench entry point to ensure `cudf::nvbench_base_fixture` usage ([#13183](https://github.com/NVIDIA/cudf/pull/13183)) [@robertmaynard](https://github.com/robertmaynard) +- Fix slice_strings to return empty strings for stop < start indices ([#13178](https://github.com/NVIDIA/cudf/pull/13178)) [@davidwendt](https://github.com/davidwendt) +- Allow compilation with any GTest version 1.11+ ([#13153](https://github.com/NVIDIA/cudf/pull/13153)) [@robertmaynard](https://github.com/robertmaynard) +- Fix a few clang-format style check errors ([#13146](https://github.com/NVIDIA/cudf/pull/13146)) [@davidwendt](https://github.com/davidwendt) +- [REVIEW] Fix `Series` and `DataFrame` constructors to validate index lengths ([#13122](https://github.com/NVIDIA/cudf/pull/13122)) [@galipremsagar](https://github.com/galipremsagar) +- Fix hash join when the input tables have nulls on only one side ([#13120](https://github.com/NVIDIA/cudf/pull/13120)) [@ttnghia](https://github.com/ttnghia) +- Fix GPU_ARCHS setting in Java CMake build and CMAKE_CUDA_ARCHITECTURES in Python package build. ([#13117](https://github.com/NVIDIA/cudf/pull/13117)) [@davidwendt](https://github.com/davidwendt) +- Adds checks to make sure json reader won't overflow ([#13115](https://github.com/NVIDIA/cudf/pull/13115)) [@elstehle](https://github.com/elstehle) +- Fix `null_count` of columns returned by `chunked_parquet_reader` ([#13111](https://github.com/NVIDIA/cudf/pull/13111)) [@vuule](https://github.com/vuule) +- Fixes sliced list and struct column bug in JSON chunked writer ([#13108](https://github.com/NVIDIA/cudf/pull/13108)) [@karthikeyann](https://github.com/karthikeyann) +- [REVIEW] Fix missing confluent kafka version ([#13101](https://github.com/NVIDIA/cudf/pull/13101)) [@galipremsagar](https://github.com/galipremsagar) +- Use make_empty_lists_column instead of make_empty_column(type_id::LIST) ([#13099](https://github.com/NVIDIA/cudf/pull/13099)) [@davidwendt](https://github.com/davidwendt) +- Raise `NotImplementedError` when attempting to construct cuDF objects from timezone-aware datetimes ([#13086](https://github.com/NVIDIA/cudf/pull/13086)) [@shwina](https://github.com/shwina) +- Fix column selection `read_parquet` benchmarks ([#13082](https://github.com/NVIDIA/cudf/pull/13082)) [@vuule](https://github.com/vuule) +- Fix bugs in iterative groupby apply algorithm ([#13078](https://github.com/NVIDIA/cudf/pull/13078)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add algorithm include in data_sink.hpp ([#13068](https://github.com/NVIDIA/cudf/pull/13068)) [@ahendriksen](https://github.com/ahendriksen) +- Fix tests/identify_stream_usage.cpp ([#13066](https://github.com/NVIDIA/cudf/pull/13066)) [@ahendriksen](https://github.com/ahendriksen) +- Prevent overflow with `skip_rows` in ORC and Parquet readers ([#13063](https://github.com/NVIDIA/cudf/pull/13063)) [@vuule](https://github.com/vuule) +- Add except declaration in Cython interface for regex_program::create ([#13054](https://github.com/NVIDIA/cudf/pull/13054)) [@davidwendt](https://github.com/davidwendt) +- [REVIEW] Fix branch version in CI scripts ([#13029](https://github.com/NVIDIA/cudf/pull/13029)) [@galipremsagar](https://github.com/galipremsagar) +- Fix OOB memory access in CSV reader when reading without NA values ([#13011](https://github.com/NVIDIA/cudf/pull/13011)) [@vuule](https://github.com/vuule) +- Fix read_avro() skip_rows and num_rows. ([#12912](https://github.com/NVIDIA/cudf/pull/12912)) [@tpn](https://github.com/tpn) +- Purge nonempty nulls from byte_cast list outputs. ([#11971](https://github.com/NVIDIA/cudf/pull/11971)) [@bdice](https://github.com/bdice) +- Fix consumption of CPU-backed interchange protocol dataframes ([#11392](https://github.com/NVIDIA/cudf/pull/11392)) [@shwina](https://github.com/shwina) ## 🚀 New Features -- Remove numba JIT kernel usage from dataframe copy tests ([#13385](https://github.com/rapidsai/cudf/pull/13385)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add JNI for ORC/Parquet writer compression statistics ([#13376](https://github.com/rapidsai/cudf/pull/13376)) [@ttnghia](https://github.com/ttnghia) -- Use _compile_or_get in JIT groupby apply ([#13350](https://github.com/rapidsai/cudf/pull/13350)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- cuDF numba cuda 12 updates ([#13337](https://github.com/rapidsai/cudf/pull/13337)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add tz_convert method to convert between timestamps ([#13328](https://github.com/rapidsai/cudf/pull/13328)) [@shwina](https://github.com/shwina) -- Optionally return compression statistics from ORC and Parquet writers ([#13294](https://github.com/rapidsai/cudf/pull/13294)) [@vuule](https://github.com/vuule) -- Support the case=False argument to str.contains ([#13290](https://github.com/rapidsai/cudf/pull/13290)) [@shwina](https://github.com/shwina) -- Add an event handler for ColumnVector.close ([#13279](https://github.com/rapidsai/cudf/pull/13279)) [@abellina](https://github.com/abellina) -- JNI api for cudf::chunked_pack ([#13278](https://github.com/rapidsai/cudf/pull/13278)) [@abellina](https://github.com/abellina) -- Implement a chunked_pack API ([#13260](https://github.com/rapidsai/cudf/pull/13260)) [@abellina](https://github.com/abellina) -- Update cudf recipes to use GTest version to >=1.13 ([#13207](https://github.com/rapidsai/cudf/pull/13207)) [@robertmaynard](https://github.com/robertmaynard) -- JNI changes for range-extents in window functions. ([#13199](https://github.com/rapidsai/cudf/pull/13199)) [@mythrocks](https://github.com/mythrocks) -- Add support for DatetimeTZDtype and tz_localize ([#13163](https://github.com/rapidsai/cudf/pull/13163)) [@shwina](https://github.com/shwina) -- Add IS_NULL operator to AST ([#13145](https://github.com/rapidsai/cudf/pull/13145)) [@karthikeyann](https://github.com/karthikeyann) -- STRING order-by column for RANGE window functions ([#13143](https://github.com/rapidsai/cudf/pull/13143)) [@mythrocks](https://github.com/mythrocks) -- Update `contains_table` to experimental row hasher and equality comparator ([#13119](https://github.com/rapidsai/cudf/pull/13119)) [@divyegala](https://github.com/divyegala) -- Automatically select `GroupBy.apply` algorithm based on if the UDF is jittable ([#13113](https://github.com/rapidsai/cudf/pull/13113)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Refactor Parquet chunked writer ([#13076](https://github.com/rapidsai/cudf/pull/13076)) [@ttnghia](https://github.com/ttnghia) -- Add Python bindings for string literal support in AST ([#13073](https://github.com/rapidsai/cudf/pull/13073)) [@karthikeyann](https://github.com/karthikeyann) -- Add Java bindings for string literal support in AST ([#13072](https://github.com/rapidsai/cudf/pull/13072)) [@karthikeyann](https://github.com/karthikeyann) -- Add string scalar support in AST ([#13061](https://github.com/rapidsai/cudf/pull/13061)) [@karthikeyann](https://github.com/karthikeyann) -- Log cuIO warnings using the libcudf logger ([#13043](https://github.com/rapidsai/cudf/pull/13043)) [@vuule](https://github.com/vuule) -- Update `mixed_join` to use experimental row hasher and comparator ([#13028](https://github.com/rapidsai/cudf/pull/13028)) [@divyegala](https://github.com/divyegala) -- Support structs of lists in row lexicographic comparator ([#13005](https://github.com/rapidsai/cudf/pull/13005)) [@ttnghia](https://github.com/ttnghia) -- Adding `hostdevice_span` that is a span createable from `hostdevice_vector` ([#12981](https://github.com/rapidsai/cudf/pull/12981)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add nvtext::minhash function ([#12961](https://github.com/rapidsai/cudf/pull/12961)) [@davidwendt](https://github.com/davidwendt) -- Support lists of structs in row lexicographic comparator ([#12953](https://github.com/rapidsai/cudf/pull/12953)) [@ttnghia](https://github.com/ttnghia) -- Update `join` to use experimental row hasher and comparator ([#12787](https://github.com/rapidsai/cudf/pull/12787)) [@divyegala](https://github.com/divyegala) -- Implement Python drop_duplicates with cudf::stable_distinct. ([#11656](https://github.com/rapidsai/cudf/pull/11656)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Remove numba JIT kernel usage from dataframe copy tests ([#13385](https://github.com/NVIDIA/cudf/pull/13385)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add JNI for ORC/Parquet writer compression statistics ([#13376](https://github.com/NVIDIA/cudf/pull/13376)) [@ttnghia](https://github.com/ttnghia) +- Use _compile_or_get in JIT groupby apply ([#13350](https://github.com/NVIDIA/cudf/pull/13350)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- cuDF numba cuda 12 updates ([#13337](https://github.com/NVIDIA/cudf/pull/13337)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add tz_convert method to convert between timestamps ([#13328](https://github.com/NVIDIA/cudf/pull/13328)) [@shwina](https://github.com/shwina) +- Optionally return compression statistics from ORC and Parquet writers ([#13294](https://github.com/NVIDIA/cudf/pull/13294)) [@vuule](https://github.com/vuule) +- Support the case=False argument to str.contains ([#13290](https://github.com/NVIDIA/cudf/pull/13290)) [@shwina](https://github.com/shwina) +- Add an event handler for ColumnVector.close ([#13279](https://github.com/NVIDIA/cudf/pull/13279)) [@abellina](https://github.com/abellina) +- JNI api for cudf::chunked_pack ([#13278](https://github.com/NVIDIA/cudf/pull/13278)) [@abellina](https://github.com/abellina) +- Implement a chunked_pack API ([#13260](https://github.com/NVIDIA/cudf/pull/13260)) [@abellina](https://github.com/abellina) +- Update cudf recipes to use GTest version to >=1.13 ([#13207](https://github.com/NVIDIA/cudf/pull/13207)) [@robertmaynard](https://github.com/robertmaynard) +- JNI changes for range-extents in window functions. ([#13199](https://github.com/NVIDIA/cudf/pull/13199)) [@mythrocks](https://github.com/mythrocks) +- Add support for DatetimeTZDtype and tz_localize ([#13163](https://github.com/NVIDIA/cudf/pull/13163)) [@shwina](https://github.com/shwina) +- Add IS_NULL operator to AST ([#13145](https://github.com/NVIDIA/cudf/pull/13145)) [@karthikeyann](https://github.com/karthikeyann) +- STRING order-by column for RANGE window functions ([#13143](https://github.com/NVIDIA/cudf/pull/13143)) [@mythrocks](https://github.com/mythrocks) +- Update `contains_table` to experimental row hasher and equality comparator ([#13119](https://github.com/NVIDIA/cudf/pull/13119)) [@divyegala](https://github.com/divyegala) +- Automatically select `GroupBy.apply` algorithm based on if the UDF is jittable ([#13113](https://github.com/NVIDIA/cudf/pull/13113)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Refactor Parquet chunked writer ([#13076](https://github.com/NVIDIA/cudf/pull/13076)) [@ttnghia](https://github.com/ttnghia) +- Add Python bindings for string literal support in AST ([#13073](https://github.com/NVIDIA/cudf/pull/13073)) [@karthikeyann](https://github.com/karthikeyann) +- Add Java bindings for string literal support in AST ([#13072](https://github.com/NVIDIA/cudf/pull/13072)) [@karthikeyann](https://github.com/karthikeyann) +- Add string scalar support in AST ([#13061](https://github.com/NVIDIA/cudf/pull/13061)) [@karthikeyann](https://github.com/karthikeyann) +- Log cuIO warnings using the libcudf logger ([#13043](https://github.com/NVIDIA/cudf/pull/13043)) [@vuule](https://github.com/vuule) +- Update `mixed_join` to use experimental row hasher and comparator ([#13028](https://github.com/NVIDIA/cudf/pull/13028)) [@divyegala](https://github.com/divyegala) +- Support structs of lists in row lexicographic comparator ([#13005](https://github.com/NVIDIA/cudf/pull/13005)) [@ttnghia](https://github.com/ttnghia) +- Adding `hostdevice_span` that is a span createable from `hostdevice_vector` ([#12981](https://github.com/NVIDIA/cudf/pull/12981)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add nvtext::minhash function ([#12961](https://github.com/NVIDIA/cudf/pull/12961)) [@davidwendt](https://github.com/davidwendt) +- Support lists of structs in row lexicographic comparator ([#12953](https://github.com/NVIDIA/cudf/pull/12953)) [@ttnghia](https://github.com/ttnghia) +- Update `join` to use experimental row hasher and comparator ([#12787](https://github.com/NVIDIA/cudf/pull/12787)) [@divyegala](https://github.com/divyegala) +- Implement Python drop_duplicates with cudf::stable_distinct. ([#11656](https://github.com/NVIDIA/cudf/pull/11656)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 🛠️ Improvements -- Drop extraneous dependencies from cudf conda recipe. ([#13406](https://github.com/rapidsai/cudf/pull/13406)) [@bdice](https://github.com/bdice) -- Handle some corner-cases in indexing with boolean masks ([#13402](https://github.com/rapidsai/cudf/pull/13402)) [@wence-](https://github.com/wence-) -- Add cudf::stable_distinct public API, tests, and benchmarks. ([#13392](https://github.com/rapidsai/cudf/pull/13392)) [@bdice](https://github.com/bdice) -- [JNI] Pass this ColumnVector to the onClosed event handler ([#13386](https://github.com/rapidsai/cudf/pull/13386)) [@abellina](https://github.com/abellina) -- Fix JNI method with mismatched parameter list ([#13384](https://github.com/rapidsai/cudf/pull/13384)) [@ttnghia](https://github.com/ttnghia) -- Split up experimental_row_operator_tests.cu to improve its compile time ([#13382](https://github.com/rapidsai/cudf/pull/13382)) [@davidwendt](https://github.com/davidwendt) -- Deprecate cudf::strings::slice_strings APIs that accept delimiters ([#13373](https://github.com/rapidsai/cudf/pull/13373)) [@davidwendt](https://github.com/davidwendt) -- Remove UNKNOWN_NULL_COUNT ([#13372](https://github.com/rapidsai/cudf/pull/13372)) [@vyasr](https://github.com/vyasr) -- Move some nvtext benchmarks to nvbench ([#13368](https://github.com/rapidsai/cudf/pull/13368)) [@davidwendt](https://github.com/davidwendt) -- run docs nightly too ([#13366](https://github.com/rapidsai/cudf/pull/13366)) [@AyodeAwe](https://github.com/AyodeAwe) -- Add warning for default `dtype` parameter in `get_dummies` ([#13365](https://github.com/rapidsai/cudf/pull/13365)) [@galipremsagar](https://github.com/galipremsagar) -- Add log messages about kvikIO compatibility mode ([#13363](https://github.com/rapidsai/cudf/pull/13363)) [@vuule](https://github.com/vuule) -- Switch back to using primary shared-action-workflows branch ([#13362](https://github.com/rapidsai/cudf/pull/13362)) [@vyasr](https://github.com/vyasr) -- Deprecate `StringIndex` and use `Index` instead ([#13361](https://github.com/rapidsai/cudf/pull/13361)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure columns have valid null counts in CUDF JNI. ([#13355](https://github.com/rapidsai/cudf/pull/13355)) [@mythrocks](https://github.com/mythrocks) -- Expunge most uses of `TypeVar(bound="Foo")` ([#13346](https://github.com/rapidsai/cudf/pull/13346)) [@wence-](https://github.com/wence-) -- Remove all references to UNKNOWN_NULL_COUNT in Python ([#13345](https://github.com/rapidsai/cudf/pull/13345)) [@vyasr](https://github.com/vyasr) -- Improve `distinct_count` with `cuco::static_set` ([#13343](https://github.com/rapidsai/cudf/pull/13343)) [@PointKernel](https://github.com/PointKernel) -- Fix `contiguous_split` performance ([#13342](https://github.com/rapidsai/cudf/pull/13342)) [@ttnghia](https://github.com/ttnghia) -- Remove default UNKNOWN_NULL_COUNT from cudf::column member functions ([#13341](https://github.com/rapidsai/cudf/pull/13341)) [@davidwendt](https://github.com/davidwendt) -- Update mypy to 1.3 ([#13340](https://github.com/rapidsai/cudf/pull/13340)) [@wence-](https://github.com/wence-) -- [Java] Purge non-empty nulls when setting validity ([#13335](https://github.com/rapidsai/cudf/pull/13335)) [@razajafri](https://github.com/razajafri) -- Add row-wise filtering step to `read_parquet` ([#13334](https://github.com/rapidsai/cudf/pull/13334)) [@rjzamora](https://github.com/rjzamora) -- Performance improvement for nvtext::minhash ([#13333](https://github.com/rapidsai/cudf/pull/13333)) [@davidwendt](https://github.com/davidwendt) -- Fix some libcudf functions to set the null count on returning columns ([#13331](https://github.com/rapidsai/cudf/pull/13331)) [@davidwendt](https://github.com/davidwendt) -- Change cudf::detail::concatenate_masks to return null-count ([#13330](https://github.com/rapidsai/cudf/pull/13330)) [@davidwendt](https://github.com/davidwendt) -- Move `meta` calculation in `dask_cudf.read_parquet` ([#13327](https://github.com/rapidsai/cudf/pull/13327)) [@rjzamora](https://github.com/rjzamora) -- Changes to support Numpy >= 1.24 ([#13325](https://github.com/rapidsai/cudf/pull/13325)) [@shwina](https://github.com/shwina) -- Use std::overflow_error when output would exceed column size limit ([#13323](https://github.com/rapidsai/cudf/pull/13323)) [@davidwendt](https://github.com/davidwendt) -- Clean up `distinct_count` benchmark ([#13321](https://github.com/rapidsai/cudf/pull/13321)) [@PointKernel](https://github.com/PointKernel) -- Fix gtest pinning to 1.13.0. ([#13319](https://github.com/rapidsai/cudf/pull/13319)) [@bdice](https://github.com/bdice) -- Remove null mask and null count from column_view constructors ([#13311](https://github.com/rapidsai/cudf/pull/13311)) [@vyasr](https://github.com/vyasr) -- Address feedback from 13289 ([#13306](https://github.com/rapidsai/cudf/pull/13306)) [@vyasr](https://github.com/vyasr) -- Change default value of the `observed=` argument in groupby to `True` to reflect the actual behaviour ([#13296](https://github.com/rapidsai/cudf/pull/13296)) [@shwina](https://github.com/shwina) -- First check for `BaseDtype` when infering the data type of an arbitrary object ([#13295](https://github.com/rapidsai/cudf/pull/13295)) [@shwina](https://github.com/shwina) -- Throw error if UNINITIALIZED is passed to cudf::state_null_count ([#13292](https://github.com/rapidsai/cudf/pull/13292)) [@davidwendt](https://github.com/davidwendt) -- Support CUDA 12.0 for pip wheels ([#13289](https://github.com/rapidsai/cudf/pull/13289)) [@divyegala](https://github.com/divyegala) -- Refactor `transform_lists_of_structs` in `row_operators.cu` ([#13288](https://github.com/rapidsai/cudf/pull/13288)) [@ttnghia](https://github.com/ttnghia) -- Branch 23.06 merge 23.04 ([#13286](https://github.com/rapidsai/cudf/pull/13286)) [@vyasr](https://github.com/vyasr) -- Update cupy dependency ([#13284](https://github.com/rapidsai/cudf/pull/13284)) [@vyasr](https://github.com/vyasr) -- Performance improvement in cudf::strings::join_strings for long strings ([#13283](https://github.com/rapidsai/cudf/pull/13283)) [@davidwendt](https://github.com/davidwendt) -- Fix unused variables and functions ([#13275](https://github.com/rapidsai/cudf/pull/13275)) [@karthikeyann](https://github.com/karthikeyann) -- Fix integer overflow in `partition` `scatter_map` construction ([#13272](https://github.com/rapidsai/cudf/pull/13272)) [@wence-](https://github.com/wence-) -- Numba 0.57 compatibility fixes ([#13271](https://github.com/rapidsai/cudf/pull/13271)) [@gmarkall](https://github.com/gmarkall) -- Performance improvement in cudf::strings::all_characters_of_type ([#13259](https://github.com/rapidsai/cudf/pull/13259)) [@davidwendt](https://github.com/davidwendt) -- Remove default null-count parameter from some libcudf factory functions ([#13258](https://github.com/rapidsai/cudf/pull/13258)) [@davidwendt](https://github.com/davidwendt) -- Roll our own generate_string() because mimesis' has gone away ([#13257](https://github.com/rapidsai/cudf/pull/13257)) [@shwina](https://github.com/shwina) -- Build wheels using new single image workflow ([#13249](https://github.com/rapidsai/cudf/pull/13249)) [@vyasr](https://github.com/vyasr) -- Enable sccache hits from local builds ([#13248](https://github.com/rapidsai/cudf/pull/13248)) [@AyodeAwe](https://github.com/AyodeAwe) -- Revert to branch-23.06 for shared-action-workflows ([#13247](https://github.com/rapidsai/cudf/pull/13247)) [@shwina](https://github.com/shwina) -- Introduce `pandas_compatible` option in `cudf` ([#13241](https://github.com/rapidsai/cudf/pull/13241)) [@galipremsagar](https://github.com/galipremsagar) -- Add metadata_builder helper class ([#13232](https://github.com/rapidsai/cudf/pull/13232)) [@abellina](https://github.com/abellina) -- Use libkvikio conda packages in libcudf, add explicit libcufile dependency. ([#13231](https://github.com/rapidsai/cudf/pull/13231)) [@bdice](https://github.com/bdice) -- Remove default null-count parameter from cudf::make_strings_column factory ([#13227](https://github.com/rapidsai/cudf/pull/13227)) [@davidwendt](https://github.com/davidwendt) -- Performance improvement in cudf::strings::find/rfind for long strings ([#13226](https://github.com/rapidsai/cudf/pull/13226)) [@davidwendt](https://github.com/davidwendt) -- Add chunked reader benchmark ([#13223](https://github.com/rapidsai/cudf/pull/13223)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Set the null count in output columns in the CSV reader ([#13221](https://github.com/rapidsai/cudf/pull/13221)) [@vuule](https://github.com/vuule) -- Skip Non-Empty nulls tests for the nightly build just like we skip CuFileTest and CudaFatalTest ([#13213](https://github.com/rapidsai/cudf/pull/13213)) [@razajafri](https://github.com/razajafri) -- Fix string_scalar stream usage in write_json.cu ([#13212](https://github.com/rapidsai/cudf/pull/13212)) [@davidwendt](https://github.com/davidwendt) -- Use canonicalized name for dlopen'd libraries (libcufile) ([#13210](https://github.com/rapidsai/cudf/pull/13210)) [@shwina](https://github.com/shwina) -- Refactor pinned memory vector and ORC+Parquet writers ([#13206](https://github.com/rapidsai/cudf/pull/13206)) [@ttnghia](https://github.com/ttnghia) -- Remove UNKNOWN_NULL_COUNT where it can be easily computed ([#13205](https://github.com/rapidsai/cudf/pull/13205)) [@vyasr](https://github.com/vyasr) -- Optimization to decoding of parquet level streams ([#13203](https://github.com/rapidsai/cudf/pull/13203)) [@nvdbaranec](https://github.com/nvdbaranec) -- Clean up and simplify `gpuDecideCompression` ([#13202](https://github.com/rapidsai/cudf/pull/13202)) [@vuule](https://github.com/vuule) -- Use std::array for a statically sized vector in `create_serialized_trie` ([#13201](https://github.com/rapidsai/cudf/pull/13201)) [@vuule](https://github.com/vuule) -- Update minimum Python version to Python 3.9 ([#13196](https://github.com/rapidsai/cudf/pull/13196)) [@shwina](https://github.com/shwina) -- Refactor contiguous_split API into contiguous_split.hpp ([#13186](https://github.com/rapidsai/cudf/pull/13186)) [@abellina](https://github.com/abellina) -- Remove usage of rapids-get-rapids-version-from-git ([#13184](https://github.com/rapidsai/cudf/pull/13184)) [@jjacobelli](https://github.com/jjacobelli) -- Enable mixed-dtype decimal/scalar binary operations ([#13171](https://github.com/rapidsai/cudf/pull/13171)) [@shwina](https://github.com/shwina) -- Split up unique_count.cu to improve build time ([#13169](https://github.com/rapidsai/cudf/pull/13169)) [@davidwendt](https://github.com/davidwendt) -- Use nvtx3 includes in string examples. ([#13165](https://github.com/rapidsai/cudf/pull/13165)) [@bdice](https://github.com/bdice) -- Change some .cu gtest files to .cpp ([#13155](https://github.com/rapidsai/cudf/pull/13155)) [@davidwendt](https://github.com/davidwendt) -- Remove wheel pytest verbosity ([#13151](https://github.com/rapidsai/cudf/pull/13151)) [@sevagh](https://github.com/sevagh) -- Fix libcudf to always pass null-count to set_null_mask ([#13149](https://github.com/rapidsai/cudf/pull/13149)) [@davidwendt](https://github.com/davidwendt) -- Fix gtests to always pass null-count to set_null_mask calls ([#13148](https://github.com/rapidsai/cudf/pull/13148)) [@davidwendt](https://github.com/davidwendt) -- Optimize JSON writer ([#13144](https://github.com/rapidsai/cudf/pull/13144)) [@karthikeyann](https://github.com/karthikeyann) -- Performance improvement for libcudf upper/lower conversion for long strings ([#13142](https://github.com/rapidsai/cudf/pull/13142)) [@davidwendt](https://github.com/davidwendt) -- [REVIEW] Deprecate `pad` and `backfill` methods ([#13140](https://github.com/rapidsai/cudf/pull/13140)) [@galipremsagar](https://github.com/galipremsagar) -- Use CTAD instead of functions in ProtobufReader ([#13135](https://github.com/rapidsai/cudf/pull/13135)) [@vuule](https://github.com/vuule) -- Remove more instances of `UNKNOWN_NULL_COUNT` ([#13134](https://github.com/rapidsai/cudf/pull/13134)) [@vyasr](https://github.com/vyasr) -- Update clang-format to 16.0.1. ([#13133](https://github.com/rapidsai/cudf/pull/13133)) [@bdice](https://github.com/bdice) -- Add log messages about cuIO's nvCOMP and cuFile use ([#13132](https://github.com/rapidsai/cudf/pull/13132)) [@vuule](https://github.com/vuule) -- Branch 23.06 merge 23.04 ([#13131](https://github.com/rapidsai/cudf/pull/13131)) [@vyasr](https://github.com/vyasr) -- Compute null-count in cudf::detail::slice ([#13124](https://github.com/rapidsai/cudf/pull/13124)) [@davidwendt](https://github.com/davidwendt) -- Use ARC V2 self-hosted runners for GPU jobs ([#13123](https://github.com/rapidsai/cudf/pull/13123)) [@jjacobelli](https://github.com/jjacobelli) -- Set null-count in linked_column_view conversion operator ([#13121](https://github.com/rapidsai/cudf/pull/13121)) [@davidwendt](https://github.com/davidwendt) -- Adding ifdefs around nvcc-specific pragmas ([#13110](https://github.com/rapidsai/cudf/pull/13110)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add null-count parameter to json experimental parse_data utility ([#13107](https://github.com/rapidsai/cudf/pull/13107)) [@davidwendt](https://github.com/davidwendt) -- Remove uses-setup-env-vars ([#13105](https://github.com/rapidsai/cudf/pull/13105)) [@vyasr](https://github.com/vyasr) -- Explicitly compute null count in concatenate APIs ([#13104](https://github.com/rapidsai/cudf/pull/13104)) [@vyasr](https://github.com/vyasr) -- Replace unnecessary uses of `UNKNOWN_NULL_COUNT` ([#13102](https://github.com/rapidsai/cudf/pull/13102)) [@vyasr](https://github.com/vyasr) -- Performance improvement for cudf::string_view::find functions ([#13100](https://github.com/rapidsai/cudf/pull/13100)) [@davidwendt](https://github.com/davidwendt) -- Use `.element()` instead of `.data()` for window range calculations ([#13095](https://github.com/rapidsai/cudf/pull/13095)) [@mythrocks](https://github.com/mythrocks) -- Cleanup Parquet chunked writer ([#13094](https://github.com/rapidsai/cudf/pull/13094)) [@ttnghia](https://github.com/ttnghia) -- Fix unused variable error/warning in page_data.cu ([#13093](https://github.com/rapidsai/cudf/pull/13093)) [@davidwendt](https://github.com/davidwendt) -- Cleanup ORC chunked writer ([#13091](https://github.com/rapidsai/cudf/pull/13091)) [@ttnghia](https://github.com/ttnghia) -- Remove using namespace cudf; from libcudf gtests source ([#13089](https://github.com/rapidsai/cudf/pull/13089)) [@davidwendt](https://github.com/davidwendt) -- Change cudf::test::make_null_mask to also return null-count ([#13081](https://github.com/rapidsai/cudf/pull/13081)) [@davidwendt](https://github.com/davidwendt) -- Resolved automerger from `branch-23.04` to `branch-23.06` ([#13080](https://github.com/rapidsai/cudf/pull/13080)) [@galipremsagar](https://github.com/galipremsagar) -- Assert for non-empty nulls ([#13071](https://github.com/rapidsai/cudf/pull/13071)) [@razajafri](https://github.com/razajafri) -- Remove deprecated regex functions from libcudf ([#13067](https://github.com/rapidsai/cudf/pull/13067)) [@davidwendt](https://github.com/davidwendt) -- Refactor `cudf::detail::sorted_order` ([#13062](https://github.com/rapidsai/cudf/pull/13062)) [@ttnghia](https://github.com/ttnghia) -- Improve performance of slice_strings for long strings ([#13057](https://github.com/rapidsai/cudf/pull/13057)) [@davidwendt](https://github.com/davidwendt) -- Reduce shared memory usage in gpuComputePageSizes by 50% ([#13047](https://github.com/rapidsai/cudf/pull/13047)) [@nvdbaranec](https://github.com/nvdbaranec) -- [REVIEW] Add notes to performance comparisons notebook ([#13044](https://github.com/rapidsai/cudf/pull/13044)) [@galipremsagar](https://github.com/galipremsagar) -- Enable binary operations between scalars and columns of differing decimal types ([#13034](https://github.com/rapidsai/cudf/pull/13034)) [@shwina](https://github.com/shwina) -- Remove console output from some libcudf gtests ([#13027](https://github.com/rapidsai/cudf/pull/13027)) [@davidwendt](https://github.com/davidwendt) -- Remove underscore in build string. ([#13025](https://github.com/rapidsai/cudf/pull/13025)) [@bdice](https://github.com/bdice) -- Bump up JNI version 23.06.0-SNAPSHOT ([#13021](https://github.com/rapidsai/cudf/pull/13021)) [@pxLi](https://github.com/pxLi) -- Fix auto merger from `branch-23.04` to `branch-23.06` ([#13009](https://github.com/rapidsai/cudf/pull/13009)) [@galipremsagar](https://github.com/galipremsagar) -- Reduce peak memory use when writing compressed ORC files. ([#12963](https://github.com/rapidsai/cudf/pull/12963)) [@vuule](https://github.com/vuule) -- Add nvtx annotatations to groupby methods ([#12941](https://github.com/rapidsai/cudf/pull/12941)) [@wence-](https://github.com/wence-) -- Compute column sizes in Parquet preprocess with single kernel ([#12931](https://github.com/rapidsai/cudf/pull/12931)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Add Python bindings for time zone data (TZiF) reader ([#12826](https://github.com/rapidsai/cudf/pull/12826)) [@shwina](https://github.com/shwina) -- Optimize set-like operations ([#12769](https://github.com/rapidsai/cudf/pull/12769)) [@ttnghia](https://github.com/ttnghia) -- [REVIEW] Upgrade to `arrow-11` ([#12757](https://github.com/rapidsai/cudf/pull/12757)) [@galipremsagar](https://github.com/galipremsagar) -- Add empty test files for test reorganization ([#12288](https://github.com/rapidsai/cudf/pull/12288)) [@shwina](https://github.com/shwina) +- Drop extraneous dependencies from cudf conda recipe. ([#13406](https://github.com/NVIDIA/cudf/pull/13406)) [@bdice](https://github.com/bdice) +- Handle some corner-cases in indexing with boolean masks ([#13402](https://github.com/NVIDIA/cudf/pull/13402)) [@wence-](https://github.com/wence-) +- Add cudf::stable_distinct public API, tests, and benchmarks. ([#13392](https://github.com/NVIDIA/cudf/pull/13392)) [@bdice](https://github.com/bdice) +- [JNI] Pass this ColumnVector to the onClosed event handler ([#13386](https://github.com/NVIDIA/cudf/pull/13386)) [@abellina](https://github.com/abellina) +- Fix JNI method with mismatched parameter list ([#13384](https://github.com/NVIDIA/cudf/pull/13384)) [@ttnghia](https://github.com/ttnghia) +- Split up experimental_row_operator_tests.cu to improve its compile time ([#13382](https://github.com/NVIDIA/cudf/pull/13382)) [@davidwendt](https://github.com/davidwendt) +- Deprecate cudf::strings::slice_strings APIs that accept delimiters ([#13373](https://github.com/NVIDIA/cudf/pull/13373)) [@davidwendt](https://github.com/davidwendt) +- Remove UNKNOWN_NULL_COUNT ([#13372](https://github.com/NVIDIA/cudf/pull/13372)) [@vyasr](https://github.com/vyasr) +- Move some nvtext benchmarks to nvbench ([#13368](https://github.com/NVIDIA/cudf/pull/13368)) [@davidwendt](https://github.com/davidwendt) +- run docs nightly too ([#13366](https://github.com/NVIDIA/cudf/pull/13366)) [@AyodeAwe](https://github.com/AyodeAwe) +- Add warning for default `dtype` parameter in `get_dummies` ([#13365](https://github.com/NVIDIA/cudf/pull/13365)) [@galipremsagar](https://github.com/galipremsagar) +- Add log messages about kvikIO compatibility mode ([#13363](https://github.com/NVIDIA/cudf/pull/13363)) [@vuule](https://github.com/vuule) +- Switch back to using primary shared-action-workflows branch ([#13362](https://github.com/NVIDIA/cudf/pull/13362)) [@vyasr](https://github.com/vyasr) +- Deprecate `StringIndex` and use `Index` instead ([#13361](https://github.com/NVIDIA/cudf/pull/13361)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure columns have valid null counts in CUDF JNI. ([#13355](https://github.com/NVIDIA/cudf/pull/13355)) [@mythrocks](https://github.com/mythrocks) +- Expunge most uses of `TypeVar(bound="Foo")` ([#13346](https://github.com/NVIDIA/cudf/pull/13346)) [@wence-](https://github.com/wence-) +- Remove all references to UNKNOWN_NULL_COUNT in Python ([#13345](https://github.com/NVIDIA/cudf/pull/13345)) [@vyasr](https://github.com/vyasr) +- Improve `distinct_count` with `cuco::static_set` ([#13343](https://github.com/NVIDIA/cudf/pull/13343)) [@PointKernel](https://github.com/PointKernel) +- Fix `contiguous_split` performance ([#13342](https://github.com/NVIDIA/cudf/pull/13342)) [@ttnghia](https://github.com/ttnghia) +- Remove default UNKNOWN_NULL_COUNT from cudf::column member functions ([#13341](https://github.com/NVIDIA/cudf/pull/13341)) [@davidwendt](https://github.com/davidwendt) +- Update mypy to 1.3 ([#13340](https://github.com/NVIDIA/cudf/pull/13340)) [@wence-](https://github.com/wence-) +- [Java] Purge non-empty nulls when setting validity ([#13335](https://github.com/NVIDIA/cudf/pull/13335)) [@razajafri](https://github.com/razajafri) +- Add row-wise filtering step to `read_parquet` ([#13334](https://github.com/NVIDIA/cudf/pull/13334)) [@rjzamora](https://github.com/rjzamora) +- Performance improvement for nvtext::minhash ([#13333](https://github.com/NVIDIA/cudf/pull/13333)) [@davidwendt](https://github.com/davidwendt) +- Fix some libcudf functions to set the null count on returning columns ([#13331](https://github.com/NVIDIA/cudf/pull/13331)) [@davidwendt](https://github.com/davidwendt) +- Change cudf::detail::concatenate_masks to return null-count ([#13330](https://github.com/NVIDIA/cudf/pull/13330)) [@davidwendt](https://github.com/davidwendt) +- Move `meta` calculation in `dask_cudf.read_parquet` ([#13327](https://github.com/NVIDIA/cudf/pull/13327)) [@rjzamora](https://github.com/rjzamora) +- Changes to support Numpy >= 1.24 ([#13325](https://github.com/NVIDIA/cudf/pull/13325)) [@shwina](https://github.com/shwina) +- Use std::overflow_error when output would exceed column size limit ([#13323](https://github.com/NVIDIA/cudf/pull/13323)) [@davidwendt](https://github.com/davidwendt) +- Clean up `distinct_count` benchmark ([#13321](https://github.com/NVIDIA/cudf/pull/13321)) [@PointKernel](https://github.com/PointKernel) +- Fix gtest pinning to 1.13.0. ([#13319](https://github.com/NVIDIA/cudf/pull/13319)) [@bdice](https://github.com/bdice) +- Remove null mask and null count from column_view constructors ([#13311](https://github.com/NVIDIA/cudf/pull/13311)) [@vyasr](https://github.com/vyasr) +- Address feedback from 13289 ([#13306](https://github.com/NVIDIA/cudf/pull/13306)) [@vyasr](https://github.com/vyasr) +- Change default value of the `observed=` argument in groupby to `True` to reflect the actual behaviour ([#13296](https://github.com/NVIDIA/cudf/pull/13296)) [@shwina](https://github.com/shwina) +- First check for `BaseDtype` when infering the data type of an arbitrary object ([#13295](https://github.com/NVIDIA/cudf/pull/13295)) [@shwina](https://github.com/shwina) +- Throw error if UNINITIALIZED is passed to cudf::state_null_count ([#13292](https://github.com/NVIDIA/cudf/pull/13292)) [@davidwendt](https://github.com/davidwendt) +- Support CUDA 12.0 for pip wheels ([#13289](https://github.com/NVIDIA/cudf/pull/13289)) [@divyegala](https://github.com/divyegala) +- Refactor `transform_lists_of_structs` in `row_operators.cu` ([#13288](https://github.com/NVIDIA/cudf/pull/13288)) [@ttnghia](https://github.com/ttnghia) +- Branch 23.06 merge 23.04 ([#13286](https://github.com/NVIDIA/cudf/pull/13286)) [@vyasr](https://github.com/vyasr) +- Update cupy dependency ([#13284](https://github.com/NVIDIA/cudf/pull/13284)) [@vyasr](https://github.com/vyasr) +- Performance improvement in cudf::strings::join_strings for long strings ([#13283](https://github.com/NVIDIA/cudf/pull/13283)) [@davidwendt](https://github.com/davidwendt) +- Fix unused variables and functions ([#13275](https://github.com/NVIDIA/cudf/pull/13275)) [@karthikeyann](https://github.com/karthikeyann) +- Fix integer overflow in `partition` `scatter_map` construction ([#13272](https://github.com/NVIDIA/cudf/pull/13272)) [@wence-](https://github.com/wence-) +- Numba 0.57 compatibility fixes ([#13271](https://github.com/NVIDIA/cudf/pull/13271)) [@gmarkall](https://github.com/gmarkall) +- Performance improvement in cudf::strings::all_characters_of_type ([#13259](https://github.com/NVIDIA/cudf/pull/13259)) [@davidwendt](https://github.com/davidwendt) +- Remove default null-count parameter from some libcudf factory functions ([#13258](https://github.com/NVIDIA/cudf/pull/13258)) [@davidwendt](https://github.com/davidwendt) +- Roll our own generate_string() because mimesis' has gone away ([#13257](https://github.com/NVIDIA/cudf/pull/13257)) [@shwina](https://github.com/shwina) +- Build wheels using new single image workflow ([#13249](https://github.com/NVIDIA/cudf/pull/13249)) [@vyasr](https://github.com/vyasr) +- Enable sccache hits from local builds ([#13248](https://github.com/NVIDIA/cudf/pull/13248)) [@AyodeAwe](https://github.com/AyodeAwe) +- Revert to branch-23.06 for shared-action-workflows ([#13247](https://github.com/NVIDIA/cudf/pull/13247)) [@shwina](https://github.com/shwina) +- Introduce `pandas_compatible` option in `cudf` ([#13241](https://github.com/NVIDIA/cudf/pull/13241)) [@galipremsagar](https://github.com/galipremsagar) +- Add metadata_builder helper class ([#13232](https://github.com/NVIDIA/cudf/pull/13232)) [@abellina](https://github.com/abellina) +- Use libkvikio conda packages in libcudf, add explicit libcufile dependency. ([#13231](https://github.com/NVIDIA/cudf/pull/13231)) [@bdice](https://github.com/bdice) +- Remove default null-count parameter from cudf::make_strings_column factory ([#13227](https://github.com/NVIDIA/cudf/pull/13227)) [@davidwendt](https://github.com/davidwendt) +- Performance improvement in cudf::strings::find/rfind for long strings ([#13226](https://github.com/NVIDIA/cudf/pull/13226)) [@davidwendt](https://github.com/davidwendt) +- Add chunked reader benchmark ([#13223](https://github.com/NVIDIA/cudf/pull/13223)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Set the null count in output columns in the CSV reader ([#13221](https://github.com/NVIDIA/cudf/pull/13221)) [@vuule](https://github.com/vuule) +- Skip Non-Empty nulls tests for the nightly build just like we skip CuFileTest and CudaFatalTest ([#13213](https://github.com/NVIDIA/cudf/pull/13213)) [@razajafri](https://github.com/razajafri) +- Fix string_scalar stream usage in write_json.cu ([#13212](https://github.com/NVIDIA/cudf/pull/13212)) [@davidwendt](https://github.com/davidwendt) +- Use canonicalized name for dlopen'd libraries (libcufile) ([#13210](https://github.com/NVIDIA/cudf/pull/13210)) [@shwina](https://github.com/shwina) +- Refactor pinned memory vector and ORC+Parquet writers ([#13206](https://github.com/NVIDIA/cudf/pull/13206)) [@ttnghia](https://github.com/ttnghia) +- Remove UNKNOWN_NULL_COUNT where it can be easily computed ([#13205](https://github.com/NVIDIA/cudf/pull/13205)) [@vyasr](https://github.com/vyasr) +- Optimization to decoding of parquet level streams ([#13203](https://github.com/NVIDIA/cudf/pull/13203)) [@nvdbaranec](https://github.com/nvdbaranec) +- Clean up and simplify `gpuDecideCompression` ([#13202](https://github.com/NVIDIA/cudf/pull/13202)) [@vuule](https://github.com/vuule) +- Use std::array for a statically sized vector in `create_serialized_trie` ([#13201](https://github.com/NVIDIA/cudf/pull/13201)) [@vuule](https://github.com/vuule) +- Update minimum Python version to Python 3.9 ([#13196](https://github.com/NVIDIA/cudf/pull/13196)) [@shwina](https://github.com/shwina) +- Refactor contiguous_split API into contiguous_split.hpp ([#13186](https://github.com/NVIDIA/cudf/pull/13186)) [@abellina](https://github.com/abellina) +- Remove usage of rapids-get-rapids-version-from-git ([#13184](https://github.com/NVIDIA/cudf/pull/13184)) [@jjacobelli](https://github.com/jjacobelli) +- Enable mixed-dtype decimal/scalar binary operations ([#13171](https://github.com/NVIDIA/cudf/pull/13171)) [@shwina](https://github.com/shwina) +- Split up unique_count.cu to improve build time ([#13169](https://github.com/NVIDIA/cudf/pull/13169)) [@davidwendt](https://github.com/davidwendt) +- Use nvtx3 includes in string examples. ([#13165](https://github.com/NVIDIA/cudf/pull/13165)) [@bdice](https://github.com/bdice) +- Change some .cu gtest files to .cpp ([#13155](https://github.com/NVIDIA/cudf/pull/13155)) [@davidwendt](https://github.com/davidwendt) +- Remove wheel pytest verbosity ([#13151](https://github.com/NVIDIA/cudf/pull/13151)) [@sevagh](https://github.com/sevagh) +- Fix libcudf to always pass null-count to set_null_mask ([#13149](https://github.com/NVIDIA/cudf/pull/13149)) [@davidwendt](https://github.com/davidwendt) +- Fix gtests to always pass null-count to set_null_mask calls ([#13148](https://github.com/NVIDIA/cudf/pull/13148)) [@davidwendt](https://github.com/davidwendt) +- Optimize JSON writer ([#13144](https://github.com/NVIDIA/cudf/pull/13144)) [@karthikeyann](https://github.com/karthikeyann) +- Performance improvement for libcudf upper/lower conversion for long strings ([#13142](https://github.com/NVIDIA/cudf/pull/13142)) [@davidwendt](https://github.com/davidwendt) +- [REVIEW] Deprecate `pad` and `backfill` methods ([#13140](https://github.com/NVIDIA/cudf/pull/13140)) [@galipremsagar](https://github.com/galipremsagar) +- Use CTAD instead of functions in ProtobufReader ([#13135](https://github.com/NVIDIA/cudf/pull/13135)) [@vuule](https://github.com/vuule) +- Remove more instances of `UNKNOWN_NULL_COUNT` ([#13134](https://github.com/NVIDIA/cudf/pull/13134)) [@vyasr](https://github.com/vyasr) +- Update clang-format to 16.0.1. ([#13133](https://github.com/NVIDIA/cudf/pull/13133)) [@bdice](https://github.com/bdice) +- Add log messages about cuIO's nvCOMP and cuFile use ([#13132](https://github.com/NVIDIA/cudf/pull/13132)) [@vuule](https://github.com/vuule) +- Branch 23.06 merge 23.04 ([#13131](https://github.com/NVIDIA/cudf/pull/13131)) [@vyasr](https://github.com/vyasr) +- Compute null-count in cudf::detail::slice ([#13124](https://github.com/NVIDIA/cudf/pull/13124)) [@davidwendt](https://github.com/davidwendt) +- Use ARC V2 self-hosted runners for GPU jobs ([#13123](https://github.com/NVIDIA/cudf/pull/13123)) [@jjacobelli](https://github.com/jjacobelli) +- Set null-count in linked_column_view conversion operator ([#13121](https://github.com/NVIDIA/cudf/pull/13121)) [@davidwendt](https://github.com/davidwendt) +- Adding ifdefs around nvcc-specific pragmas ([#13110](https://github.com/NVIDIA/cudf/pull/13110)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add null-count parameter to json experimental parse_data utility ([#13107](https://github.com/NVIDIA/cudf/pull/13107)) [@davidwendt](https://github.com/davidwendt) +- Remove uses-setup-env-vars ([#13105](https://github.com/NVIDIA/cudf/pull/13105)) [@vyasr](https://github.com/vyasr) +- Explicitly compute null count in concatenate APIs ([#13104](https://github.com/NVIDIA/cudf/pull/13104)) [@vyasr](https://github.com/vyasr) +- Replace unnecessary uses of `UNKNOWN_NULL_COUNT` ([#13102](https://github.com/NVIDIA/cudf/pull/13102)) [@vyasr](https://github.com/vyasr) +- Performance improvement for cudf::string_view::find functions ([#13100](https://github.com/NVIDIA/cudf/pull/13100)) [@davidwendt](https://github.com/davidwendt) +- Use `.element()` instead of `.data()` for window range calculations ([#13095](https://github.com/NVIDIA/cudf/pull/13095)) [@mythrocks](https://github.com/mythrocks) +- Cleanup Parquet chunked writer ([#13094](https://github.com/NVIDIA/cudf/pull/13094)) [@ttnghia](https://github.com/ttnghia) +- Fix unused variable error/warning in page_data.cu ([#13093](https://github.com/NVIDIA/cudf/pull/13093)) [@davidwendt](https://github.com/davidwendt) +- Cleanup ORC chunked writer ([#13091](https://github.com/NVIDIA/cudf/pull/13091)) [@ttnghia](https://github.com/ttnghia) +- Remove using namespace cudf; from libcudf gtests source ([#13089](https://github.com/NVIDIA/cudf/pull/13089)) [@davidwendt](https://github.com/davidwendt) +- Change cudf::test::make_null_mask to also return null-count ([#13081](https://github.com/NVIDIA/cudf/pull/13081)) [@davidwendt](https://github.com/davidwendt) +- Resolved automerger from `branch-23.04` to `branch-23.06` ([#13080](https://github.com/NVIDIA/cudf/pull/13080)) [@galipremsagar](https://github.com/galipremsagar) +- Assert for non-empty nulls ([#13071](https://github.com/NVIDIA/cudf/pull/13071)) [@razajafri](https://github.com/razajafri) +- Remove deprecated regex functions from libcudf ([#13067](https://github.com/NVIDIA/cudf/pull/13067)) [@davidwendt](https://github.com/davidwendt) +- Refactor `cudf::detail::sorted_order` ([#13062](https://github.com/NVIDIA/cudf/pull/13062)) [@ttnghia](https://github.com/ttnghia) +- Improve performance of slice_strings for long strings ([#13057](https://github.com/NVIDIA/cudf/pull/13057)) [@davidwendt](https://github.com/davidwendt) +- Reduce shared memory usage in gpuComputePageSizes by 50% ([#13047](https://github.com/NVIDIA/cudf/pull/13047)) [@nvdbaranec](https://github.com/nvdbaranec) +- [REVIEW] Add notes to performance comparisons notebook ([#13044](https://github.com/NVIDIA/cudf/pull/13044)) [@galipremsagar](https://github.com/galipremsagar) +- Enable binary operations between scalars and columns of differing decimal types ([#13034](https://github.com/NVIDIA/cudf/pull/13034)) [@shwina](https://github.com/shwina) +- Remove console output from some libcudf gtests ([#13027](https://github.com/NVIDIA/cudf/pull/13027)) [@davidwendt](https://github.com/davidwendt) +- Remove underscore in build string. ([#13025](https://github.com/NVIDIA/cudf/pull/13025)) [@bdice](https://github.com/bdice) +- Bump up JNI version 23.06.0-SNAPSHOT ([#13021](https://github.com/NVIDIA/cudf/pull/13021)) [@pxLi](https://github.com/pxLi) +- Fix auto merger from `branch-23.04` to `branch-23.06` ([#13009](https://github.com/NVIDIA/cudf/pull/13009)) [@galipremsagar](https://github.com/galipremsagar) +- Reduce peak memory use when writing compressed ORC files. ([#12963](https://github.com/NVIDIA/cudf/pull/12963)) [@vuule](https://github.com/vuule) +- Add nvtx annotatations to groupby methods ([#12941](https://github.com/NVIDIA/cudf/pull/12941)) [@wence-](https://github.com/wence-) +- Compute column sizes in Parquet preprocess with single kernel ([#12931](https://github.com/NVIDIA/cudf/pull/12931)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Add Python bindings for time zone data (TZiF) reader ([#12826](https://github.com/NVIDIA/cudf/pull/12826)) [@shwina](https://github.com/shwina) +- Optimize set-like operations ([#12769](https://github.com/NVIDIA/cudf/pull/12769)) [@ttnghia](https://github.com/ttnghia) +- [REVIEW] Upgrade to `arrow-11` ([#12757](https://github.com/NVIDIA/cudf/pull/12757)) [@galipremsagar](https://github.com/galipremsagar) +- Add empty test files for test reorganization ([#12288](https://github.com/NVIDIA/cudf/pull/12288)) [@shwina](https://github.com/shwina) # cuDF 23.04.00 (6 Apr 2023) ## 🚨 Breaking Changes -- Pin `dask` and `distributed` for release ([#13070](https://github.com/rapidsai/cudf/pull/13070)) [@galipremsagar](https://github.com/galipremsagar) -- Declare a different name for nan_equality.UNEQUAL to prevent Cython warnings. ([#12947](https://github.com/rapidsai/cudf/pull/12947)) [@bdice](https://github.com/bdice) -- Update minimum `pandas` and `numpy` pinnings ([#12887](https://github.com/rapidsai/cudf/pull/12887)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `names` & `dtype` in `Index.copy` ([#12825](https://github.com/rapidsai/cudf/pull/12825)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `Index.is_*` methods ([#12820](https://github.com/rapidsai/cudf/pull/12820)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `datetime_is_numeric` from `describe` ([#12818](https://github.com/rapidsai/cudf/pull/12818)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `na_sentinel` in `factorize` ([#12817](https://github.com/rapidsai/cudf/pull/12817)) [@galipremsagar](https://github.com/galipremsagar) -- Make string methods return a Series with a useful Index ([#12814](https://github.com/rapidsai/cudf/pull/12814)) [@shwina](https://github.com/shwina) -- Produce useful guidance on overflow error in `to_csv` ([#12705](https://github.com/rapidsai/cudf/pull/12705)) [@wence-](https://github.com/wence-) -- Move `strings_udf` code into cuDF ([#12669](https://github.com/rapidsai/cudf/pull/12669)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings ([#12609](https://github.com/rapidsai/cudf/pull/12609)) [@davidwendt](https://github.com/davidwendt) -- Replace message parsing with throwing more specific exceptions ([#12426](https://github.com/rapidsai/cudf/pull/12426)) [@vyasr](https://github.com/vyasr) +- Pin `dask` and `distributed` for release ([#13070](https://github.com/NVIDIA/cudf/pull/13070)) [@galipremsagar](https://github.com/galipremsagar) +- Declare a different name for nan_equality.UNEQUAL to prevent Cython warnings. ([#12947](https://github.com/NVIDIA/cudf/pull/12947)) [@bdice](https://github.com/bdice) +- Update minimum `pandas` and `numpy` pinnings ([#12887](https://github.com/NVIDIA/cudf/pull/12887)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `names` & `dtype` in `Index.copy` ([#12825](https://github.com/NVIDIA/cudf/pull/12825)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `Index.is_*` methods ([#12820](https://github.com/NVIDIA/cudf/pull/12820)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `datetime_is_numeric` from `describe` ([#12818](https://github.com/NVIDIA/cudf/pull/12818)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `na_sentinel` in `factorize` ([#12817](https://github.com/NVIDIA/cudf/pull/12817)) [@galipremsagar](https://github.com/galipremsagar) +- Make string methods return a Series with a useful Index ([#12814](https://github.com/NVIDIA/cudf/pull/12814)) [@shwina](https://github.com/shwina) +- Produce useful guidance on overflow error in `to_csv` ([#12705](https://github.com/NVIDIA/cudf/pull/12705)) [@wence-](https://github.com/wence-) +- Move `strings_udf` code into cuDF ([#12669](https://github.com/NVIDIA/cudf/pull/12669)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings ([#12609](https://github.com/NVIDIA/cudf/pull/12609)) [@davidwendt](https://github.com/davidwendt) +- Replace message parsing with throwing more specific exceptions ([#12426](https://github.com/NVIDIA/cudf/pull/12426)) [@vyasr](https://github.com/vyasr) ## 🐛 Bug Fixes -- Fix memcheck script to execute only _TEST files found in bin/gtests/libcudf ([#13006](https://github.com/rapidsai/cudf/pull/13006)) [@davidwendt](https://github.com/davidwendt) -- Fix `DataFrame` constructor to broadcast scalar inputs properly ([#12997](https://github.com/rapidsai/cudf/pull/12997)) [@galipremsagar](https://github.com/galipremsagar) -- Drop `force_nullable_schema` from chunked parquet writer ([#12996](https://github.com/rapidsai/cudf/pull/12996)) [@galipremsagar](https://github.com/galipremsagar) -- Fix gtest column utility comparator diff reporting ([#12995](https://github.com/rapidsai/cudf/pull/12995)) [@davidwendt](https://github.com/davidwendt) -- Handle index names while performing `groupby` ([#12992](https://github.com/rapidsai/cudf/pull/12992)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `__setitem__` on string columns when the scalar value ends in a null byte ([#12991](https://github.com/rapidsai/cudf/pull/12991)) [@wence-](https://github.com/wence-) -- Fix `sort_values` when column is all empty strings ([#12988](https://github.com/rapidsai/cudf/pull/12988)) [@eriknw](https://github.com/eriknw) -- Remove unused variable and fix memory issue in ORC writer ([#12984](https://github.com/rapidsai/cudf/pull/12984)) [@ttnghia](https://github.com/ttnghia) -- Pre-emptive fix for upstream `dask.dataframe.read_parquet` changes ([#12983](https://github.com/rapidsai/cudf/pull/12983)) [@rjzamora](https://github.com/rjzamora) -- Remove MANIFEST.in use auto-generated one for sdists and package_data for wheels ([#12960](https://github.com/rapidsai/cudf/pull/12960)) [@vyasr](https://github.com/vyasr) -- Update to use rapids-export(COMPONENTS) feature. ([#12959](https://github.com/rapidsai/cudf/pull/12959)) [@robertmaynard](https://github.com/robertmaynard) -- cudftestutil supports static gtest dependencies ([#12957](https://github.com/rapidsai/cudf/pull/12957)) [@robertmaynard](https://github.com/robertmaynard) -- Include gtest in build environment. ([#12956](https://github.com/rapidsai/cudf/pull/12956)) [@vyasr](https://github.com/vyasr) -- Correctly handle scalar indices in `Index.__getitem__` ([#12955](https://github.com/rapidsai/cudf/pull/12955)) [@wence-](https://github.com/wence-) -- Avoid building cython twice ([#12945](https://github.com/rapidsai/cudf/pull/12945)) [@galipremsagar](https://github.com/galipremsagar) -- Fix set index error for Series rolling window operations ([#12942](https://github.com/rapidsai/cudf/pull/12942)) [@galipremsagar](https://github.com/galipremsagar) -- Fix calculation of null counts for Parquet statistics ([#12938](https://github.com/rapidsai/cudf/pull/12938)) [@etseidl](https://github.com/etseidl) -- Preserve integer dtype of hive-partitioned column containing nulls ([#12930](https://github.com/rapidsai/cudf/pull/12930)) [@rjzamora](https://github.com/rjzamora) -- Use get_current_device_resource for intermediate allocations in COLLECT_LIST window code ([#12927](https://github.com/rapidsai/cudf/pull/12927)) [@karthikeyann](https://github.com/karthikeyann) -- Mark dlpack tensor deleter as noexcept to match PyCapsule_Destructor signature. ([#12921](https://github.com/rapidsai/cudf/pull/12921)) [@bdice](https://github.com/bdice) -- Fix conda recipe post-link.sh typo ([#12916](https://github.com/rapidsai/cudf/pull/12916)) [@pentschev](https://github.com/pentschev) -- min_rows and num_rows are swapped in ComputePageSizes declaration in Parquet reader ([#12886](https://github.com/rapidsai/cudf/pull/12886)) [@etseidl](https://github.com/etseidl) -- Expect cupy to now support bool arrays for dlpack. ([#12883](https://github.com/rapidsai/cudf/pull/12883)) [@vyasr](https://github.com/vyasr) -- Use python -m pytest for nightly wheel tests ([#12871](https://github.com/rapidsai/cudf/pull/12871)) [@bdice](https://github.com/bdice) -- Parquet writer column_size() should return a size_t ([#12870](https://github.com/rapidsai/cudf/pull/12870)) [@etseidl](https://github.com/etseidl) -- Fix cudf::hash_partition kernel launch error with decimal128 types ([#12863](https://github.com/rapidsai/cudf/pull/12863)) [@davidwendt](https://github.com/davidwendt) -- Fix an issue with parquet chunked reader undercounting string lengths. ([#12859](https://github.com/rapidsai/cudf/pull/12859)) [@nvdbaranec](https://github.com/nvdbaranec) -- Remove tokenizers pre-install pinning. ([#12854](https://github.com/rapidsai/cudf/pull/12854)) [@vyasr](https://github.com/vyasr) -- Fix parquet `RangeIndex` bug ([#12838](https://github.com/rapidsai/cudf/pull/12838)) [@rjzamora](https://github.com/rjzamora) -- Remove KAFKA_HOST_TEST from compute-sanitizer check ([#12831](https://github.com/rapidsai/cudf/pull/12831)) [@davidwendt](https://github.com/davidwendt) -- Make string methods return a Series with a useful Index ([#12814](https://github.com/rapidsai/cudf/pull/12814)) [@shwina](https://github.com/shwina) -- Tell cudf_kafka to use header-only fmt ([#12796](https://github.com/rapidsai/cudf/pull/12796)) [@vyasr](https://github.com/vyasr) -- Add `GroupBy.dtypes` ([#12783](https://github.com/rapidsai/cudf/pull/12783)) [@galipremsagar](https://github.com/galipremsagar) -- Fix a leak in a test and clarify some test names ([#12781](https://github.com/rapidsai/cudf/pull/12781)) [@revans2](https://github.com/revans2) -- Fix bug in all-null list due to join_list_elements special handling ([#12767](https://github.com/rapidsai/cudf/pull/12767)) [@karthikeyann](https://github.com/karthikeyann) -- Add try/except for expected null-schema error in read_parquet ([#12756](https://github.com/rapidsai/cudf/pull/12756)) [@rjzamora](https://github.com/rjzamora) -- Throw an exception if an unsupported page encoding is detected in Parquet reader ([#12754](https://github.com/rapidsai/cudf/pull/12754)) [@etseidl](https://github.com/etseidl) -- Fix a bug with `num_keys` in `_scatter_by_slice` ([#12749](https://github.com/rapidsai/cudf/pull/12749)) [@thomcom](https://github.com/thomcom) -- Bump pinned rapids wheel deps to 23.4 ([#12735](https://github.com/rapidsai/cudf/pull/12735)) [@sevagh](https://github.com/sevagh) -- Rework logic in cudf::strings::split_record to improve performance ([#12729](https://github.com/rapidsai/cudf/pull/12729)) [@davidwendt](https://github.com/davidwendt) -- Add `always_nullable` flag to Dremel encoding ([#12727](https://github.com/rapidsai/cudf/pull/12727)) [@divyegala](https://github.com/divyegala) -- Fix memcheck read error in compound segmented reduce ([#12722](https://github.com/rapidsai/cudf/pull/12722)) [@davidwendt](https://github.com/davidwendt) -- Fix faulty conditional logic in JIT `GroupBy.apply` ([#12706](https://github.com/rapidsai/cudf/pull/12706)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Produce useful guidance on overflow error in `to_csv` ([#12705](https://github.com/rapidsai/cudf/pull/12705)) [@wence-](https://github.com/wence-) -- Handle parquet list data corner case ([#12698](https://github.com/rapidsai/cudf/pull/12698)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix missing trailing comma in json writer ([#12688](https://github.com/rapidsai/cudf/pull/12688)) [@karthikeyann](https://github.com/karthikeyann) -- Remove child fom newCudaAsyncMemoryResource ([#12681](https://github.com/rapidsai/cudf/pull/12681)) [@abellina](https://github.com/abellina) -- Handle bool types in `round` API ([#12670](https://github.com/rapidsai/cudf/pull/12670)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure all of device bitmask is initialized in from_arrow ([#12668](https://github.com/rapidsai/cudf/pull/12668)) [@wence-](https://github.com/wence-) -- Fix `from_arrow` to load a sliced arrow table ([#12665](https://github.com/rapidsai/cudf/pull/12665)) [@galipremsagar](https://github.com/galipremsagar) -- Fix dask-cudf read_parquet bug for multi-file aggregation ([#12663](https://github.com/rapidsai/cudf/pull/12663)) [@rjzamora](https://github.com/rjzamora) -- Fix AllocateLikeTest gtests reading uninitialized null-mask ([#12643](https://github.com/rapidsai/cudf/pull/12643)) [@davidwendt](https://github.com/davidwendt) -- Fix `find_common_dtype` and `values` to handle complex dtypes ([#12537](https://github.com/rapidsai/cudf/pull/12537)) [@galipremsagar](https://github.com/galipremsagar) -- Fix fetching of MultiIndex values when a label is passed ([#12521](https://github.com/rapidsai/cudf/pull/12521)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `Series` comparison vs scalars ([#12519](https://github.com/rapidsai/cudf/pull/12519)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Allow casting from `UDFString` back to `StringView` to call methods in `strings_udf` ([#12363](https://github.com/rapidsai/cudf/pull/12363)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix memcheck script to execute only _TEST files found in bin/gtests/libcudf ([#13006](https://github.com/NVIDIA/cudf/pull/13006)) [@davidwendt](https://github.com/davidwendt) +- Fix `DataFrame` constructor to broadcast scalar inputs properly ([#12997](https://github.com/NVIDIA/cudf/pull/12997)) [@galipremsagar](https://github.com/galipremsagar) +- Drop `force_nullable_schema` from chunked parquet writer ([#12996](https://github.com/NVIDIA/cudf/pull/12996)) [@galipremsagar](https://github.com/galipremsagar) +- Fix gtest column utility comparator diff reporting ([#12995](https://github.com/NVIDIA/cudf/pull/12995)) [@davidwendt](https://github.com/davidwendt) +- Handle index names while performing `groupby` ([#12992](https://github.com/NVIDIA/cudf/pull/12992)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `__setitem__` on string columns when the scalar value ends in a null byte ([#12991](https://github.com/NVIDIA/cudf/pull/12991)) [@wence-](https://github.com/wence-) +- Fix `sort_values` when column is all empty strings ([#12988](https://github.com/NVIDIA/cudf/pull/12988)) [@eriknw](https://github.com/eriknw) +- Remove unused variable and fix memory issue in ORC writer ([#12984](https://github.com/NVIDIA/cudf/pull/12984)) [@ttnghia](https://github.com/ttnghia) +- Pre-emptive fix for upstream `dask.dataframe.read_parquet` changes ([#12983](https://github.com/NVIDIA/cudf/pull/12983)) [@rjzamora](https://github.com/rjzamora) +- Remove MANIFEST.in use auto-generated one for sdists and package_data for wheels ([#12960](https://github.com/NVIDIA/cudf/pull/12960)) [@vyasr](https://github.com/vyasr) +- Update to use rapids-export(COMPONENTS) feature. ([#12959](https://github.com/NVIDIA/cudf/pull/12959)) [@robertmaynard](https://github.com/robertmaynard) +- cudftestutil supports static gtest dependencies ([#12957](https://github.com/NVIDIA/cudf/pull/12957)) [@robertmaynard](https://github.com/robertmaynard) +- Include gtest in build environment. ([#12956](https://github.com/NVIDIA/cudf/pull/12956)) [@vyasr](https://github.com/vyasr) +- Correctly handle scalar indices in `Index.__getitem__` ([#12955](https://github.com/NVIDIA/cudf/pull/12955)) [@wence-](https://github.com/wence-) +- Avoid building cython twice ([#12945](https://github.com/NVIDIA/cudf/pull/12945)) [@galipremsagar](https://github.com/galipremsagar) +- Fix set index error for Series rolling window operations ([#12942](https://github.com/NVIDIA/cudf/pull/12942)) [@galipremsagar](https://github.com/galipremsagar) +- Fix calculation of null counts for Parquet statistics ([#12938](https://github.com/NVIDIA/cudf/pull/12938)) [@etseidl](https://github.com/etseidl) +- Preserve integer dtype of hive-partitioned column containing nulls ([#12930](https://github.com/NVIDIA/cudf/pull/12930)) [@rjzamora](https://github.com/rjzamora) +- Use get_current_device_resource for intermediate allocations in COLLECT_LIST window code ([#12927](https://github.com/NVIDIA/cudf/pull/12927)) [@karthikeyann](https://github.com/karthikeyann) +- Mark dlpack tensor deleter as noexcept to match PyCapsule_Destructor signature. ([#12921](https://github.com/NVIDIA/cudf/pull/12921)) [@bdice](https://github.com/bdice) +- Fix conda recipe post-link.sh typo ([#12916](https://github.com/NVIDIA/cudf/pull/12916)) [@pentschev](https://github.com/pentschev) +- min_rows and num_rows are swapped in ComputePageSizes declaration in Parquet reader ([#12886](https://github.com/NVIDIA/cudf/pull/12886)) [@etseidl](https://github.com/etseidl) +- Expect cupy to now support bool arrays for dlpack. ([#12883](https://github.com/NVIDIA/cudf/pull/12883)) [@vyasr](https://github.com/vyasr) +- Use python -m pytest for nightly wheel tests ([#12871](https://github.com/NVIDIA/cudf/pull/12871)) [@bdice](https://github.com/bdice) +- Parquet writer column_size() should return a size_t ([#12870](https://github.com/NVIDIA/cudf/pull/12870)) [@etseidl](https://github.com/etseidl) +- Fix cudf::hash_partition kernel launch error with decimal128 types ([#12863](https://github.com/NVIDIA/cudf/pull/12863)) [@davidwendt](https://github.com/davidwendt) +- Fix an issue with parquet chunked reader undercounting string lengths. ([#12859](https://github.com/NVIDIA/cudf/pull/12859)) [@nvdbaranec](https://github.com/nvdbaranec) +- Remove tokenizers pre-install pinning. ([#12854](https://github.com/NVIDIA/cudf/pull/12854)) [@vyasr](https://github.com/vyasr) +- Fix parquet `RangeIndex` bug ([#12838](https://github.com/NVIDIA/cudf/pull/12838)) [@rjzamora](https://github.com/rjzamora) +- Remove KAFKA_HOST_TEST from compute-sanitizer check ([#12831](https://github.com/NVIDIA/cudf/pull/12831)) [@davidwendt](https://github.com/davidwendt) +- Make string methods return a Series with a useful Index ([#12814](https://github.com/NVIDIA/cudf/pull/12814)) [@shwina](https://github.com/shwina) +- Tell cudf_kafka to use header-only fmt ([#12796](https://github.com/NVIDIA/cudf/pull/12796)) [@vyasr](https://github.com/vyasr) +- Add `GroupBy.dtypes` ([#12783](https://github.com/NVIDIA/cudf/pull/12783)) [@galipremsagar](https://github.com/galipremsagar) +- Fix a leak in a test and clarify some test names ([#12781](https://github.com/NVIDIA/cudf/pull/12781)) [@revans2](https://github.com/revans2) +- Fix bug in all-null list due to join_list_elements special handling ([#12767](https://github.com/NVIDIA/cudf/pull/12767)) [@karthikeyann](https://github.com/karthikeyann) +- Add try/except for expected null-schema error in read_parquet ([#12756](https://github.com/NVIDIA/cudf/pull/12756)) [@rjzamora](https://github.com/rjzamora) +- Throw an exception if an unsupported page encoding is detected in Parquet reader ([#12754](https://github.com/NVIDIA/cudf/pull/12754)) [@etseidl](https://github.com/etseidl) +- Fix a bug with `num_keys` in `_scatter_by_slice` ([#12749](https://github.com/NVIDIA/cudf/pull/12749)) [@thomcom](https://github.com/thomcom) +- Bump pinned rapids wheel deps to 23.4 ([#12735](https://github.com/NVIDIA/cudf/pull/12735)) [@sevagh](https://github.com/sevagh) +- Rework logic in cudf::strings::split_record to improve performance ([#12729](https://github.com/NVIDIA/cudf/pull/12729)) [@davidwendt](https://github.com/davidwendt) +- Add `always_nullable` flag to Dremel encoding ([#12727](https://github.com/NVIDIA/cudf/pull/12727)) [@divyegala](https://github.com/divyegala) +- Fix memcheck read error in compound segmented reduce ([#12722](https://github.com/NVIDIA/cudf/pull/12722)) [@davidwendt](https://github.com/davidwendt) +- Fix faulty conditional logic in JIT `GroupBy.apply` ([#12706](https://github.com/NVIDIA/cudf/pull/12706)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Produce useful guidance on overflow error in `to_csv` ([#12705](https://github.com/NVIDIA/cudf/pull/12705)) [@wence-](https://github.com/wence-) +- Handle parquet list data corner case ([#12698](https://github.com/NVIDIA/cudf/pull/12698)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix missing trailing comma in json writer ([#12688](https://github.com/NVIDIA/cudf/pull/12688)) [@karthikeyann](https://github.com/karthikeyann) +- Remove child fom newCudaAsyncMemoryResource ([#12681](https://github.com/NVIDIA/cudf/pull/12681)) [@abellina](https://github.com/abellina) +- Handle bool types in `round` API ([#12670](https://github.com/NVIDIA/cudf/pull/12670)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure all of device bitmask is initialized in from_arrow ([#12668](https://github.com/NVIDIA/cudf/pull/12668)) [@wence-](https://github.com/wence-) +- Fix `from_arrow` to load a sliced arrow table ([#12665](https://github.com/NVIDIA/cudf/pull/12665)) [@galipremsagar](https://github.com/galipremsagar) +- Fix dask-cudf read_parquet bug for multi-file aggregation ([#12663](https://github.com/NVIDIA/cudf/pull/12663)) [@rjzamora](https://github.com/rjzamora) +- Fix AllocateLikeTest gtests reading uninitialized null-mask ([#12643](https://github.com/NVIDIA/cudf/pull/12643)) [@davidwendt](https://github.com/davidwendt) +- Fix `find_common_dtype` and `values` to handle complex dtypes ([#12537](https://github.com/NVIDIA/cudf/pull/12537)) [@galipremsagar](https://github.com/galipremsagar) +- Fix fetching of MultiIndex values when a label is passed ([#12521](https://github.com/NVIDIA/cudf/pull/12521)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `Series` comparison vs scalars ([#12519](https://github.com/NVIDIA/cudf/pull/12519)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Allow casting from `UDFString` back to `StringView` to call methods in `strings_udf` ([#12363](https://github.com/NVIDIA/cudf/pull/12363)) [@brandon-b-miller](https://github.com/brandon-b-miller) ## 📖 Documentation -- Fix `GroupBy.apply` doc examples rendering ([#12994](https://github.com/rapidsai/cudf/pull/12994)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- add sphinx building and s3 uploading for dask-cudf docs ([#12982](https://github.com/rapidsai/cudf/pull/12982)) [@quasiben](https://github.com/quasiben) -- Add developer documentation forbidding default parameters in detail APIs ([#12978](https://github.com/rapidsai/cudf/pull/12978)) [@vyasr](https://github.com/vyasr) -- Add README symlink for dask-cudf. ([#12946](https://github.com/rapidsai/cudf/pull/12946)) [@bdice](https://github.com/bdice) -- Remove return type from [@return doxygen tags ([#12908](https://github.com/rapidsai/cudf/pull/12908)) @davidwendt](https://github.com/return doxygen tags ([#12908](https://github.com/rapidsai/cudf/pull/12908)) @davidwendt) -- Fix docs build to be `pydata-sphinx-theme=0.13.0` compatible ([#12874](https://github.com/rapidsai/cudf/pull/12874)) [@galipremsagar](https://github.com/galipremsagar) -- Add skeleton API and prose documentation for dask-cudf ([#12725](https://github.com/rapidsai/cudf/pull/12725)) [@wence-](https://github.com/wence-) -- Enable doctests for GroupBy methods ([#12658](https://github.com/rapidsai/cudf/pull/12658)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add comment about CUB patch for SegmentedSortInt.Bool gtest ([#12611](https://github.com/rapidsai/cudf/pull/12611)) [@davidwendt](https://github.com/davidwendt) +- Fix `GroupBy.apply` doc examples rendering ([#12994](https://github.com/NVIDIA/cudf/pull/12994)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- add sphinx building and s3 uploading for dask-cudf docs ([#12982](https://github.com/NVIDIA/cudf/pull/12982)) [@quasiben](https://github.com/quasiben) +- Add developer documentation forbidding default parameters in detail APIs ([#12978](https://github.com/NVIDIA/cudf/pull/12978)) [@vyasr](https://github.com/vyasr) +- Add README symlink for dask-cudf. ([#12946](https://github.com/NVIDIA/cudf/pull/12946)) [@bdice](https://github.com/bdice) +- Remove return type from [@return doxygen tags ([#12908](https://github.com/NVIDIA/cudf/pull/12908)) @davidwendt](https://github.com/return doxygen tags ([#12908](https://github.com/NVIDIA/cudf/pull/12908)) @davidwendt) +- Fix docs build to be `pydata-sphinx-theme=0.13.0` compatible ([#12874](https://github.com/NVIDIA/cudf/pull/12874)) [@galipremsagar](https://github.com/galipremsagar) +- Add skeleton API and prose documentation for dask-cudf ([#12725](https://github.com/NVIDIA/cudf/pull/12725)) [@wence-](https://github.com/wence-) +- Enable doctests for GroupBy methods ([#12658](https://github.com/NVIDIA/cudf/pull/12658)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add comment about CUB patch for SegmentedSortInt.Bool gtest ([#12611](https://github.com/NVIDIA/cudf/pull/12611)) [@davidwendt](https://github.com/davidwendt) ## 🚀 New Features -- Add JNI method for strings::replace multi variety ([#12979](https://github.com/rapidsai/cudf/pull/12979)) [@NVnavkumar](https://github.com/NVnavkumar) -- Add nunique aggregation support for cudf::segmented_reduce ([#12972](https://github.com/rapidsai/cudf/pull/12972)) [@davidwendt](https://github.com/davidwendt) -- Refactor orc chunked writer ([#12949](https://github.com/rapidsai/cudf/pull/12949)) [@ttnghia](https://github.com/ttnghia) -- Make Parquet writer `nullable` option application to single table writes ([#12933](https://github.com/rapidsai/cudf/pull/12933)) [@vuule](https://github.com/vuule) -- Refactor `io::orc::ProtobufWriter` ([#12877](https://github.com/rapidsai/cudf/pull/12877)) [@ttnghia](https://github.com/ttnghia) -- Make timezone table independent from ORC ([#12805](https://github.com/rapidsai/cudf/pull/12805)) [@vuule](https://github.com/vuule) -- Cache JIT `GroupBy.apply` functions ([#12802](https://github.com/rapidsai/cudf/pull/12802)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Implement initial support for avro logical types ([#6482) (#12788](https://github.com/rapidsai/cudf/pull/6482) (#12788)) [@tpn](https://github.com/tpn) -- Update `tests/column_utilities` to use `experimental::equality` row comparator ([#12777](https://github.com/rapidsai/cudf/pull/12777)) [@divyegala](https://github.com/divyegala) -- Update `distinct/unique_count` to `experimental::row` hasher/comparator ([#12776](https://github.com/rapidsai/cudf/pull/12776)) [@divyegala](https://github.com/divyegala) -- Update `hash_partition` to use `experimental::row::row_hasher` ([#12761](https://github.com/rapidsai/cudf/pull/12761)) [@divyegala](https://github.com/divyegala) -- Update `is_sorted` to use `experimental::row::lexicographic` ([#12752](https://github.com/rapidsai/cudf/pull/12752)) [@divyegala](https://github.com/divyegala) -- Update default data source in cuio reader benchmarks ([#12740](https://github.com/rapidsai/cudf/pull/12740)) [@PointKernel](https://github.com/PointKernel) -- Reenable stream identification library in CI ([#12714](https://github.com/rapidsai/cudf/pull/12714)) [@vyasr](https://github.com/vyasr) -- Add `regex_program` strings splitting java APIs and tests ([#12713](https://github.com/rapidsai/cudf/pull/12713)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Add `regex_program` strings replacing java APIs and tests ([#12701](https://github.com/rapidsai/cudf/pull/12701)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Add `regex_program` strings extract java APIs and tests ([#12699](https://github.com/rapidsai/cudf/pull/12699)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Variable fragment sizes for Parquet writer ([#12685](https://github.com/rapidsai/cudf/pull/12685)) [@etseidl](https://github.com/etseidl) -- Add segmented reduction support for fixed-point types ([#12680](https://github.com/rapidsai/cudf/pull/12680)) [@davidwendt](https://github.com/davidwendt) -- Move `strings_udf` code into cuDF ([#12669](https://github.com/rapidsai/cudf/pull/12669)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add `regex_program` searching APIs and related java classes ([#12666](https://github.com/rapidsai/cudf/pull/12666)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Add logging to libcudf ([#12637](https://github.com/rapidsai/cudf/pull/12637)) [@vuule](https://github.com/vuule) -- Add compound aggregations to cudf::segmented_reduce ([#12573](https://github.com/rapidsai/cudf/pull/12573)) [@davidwendt](https://github.com/davidwendt) -- Convert `rank` to use to experimental row comparators ([#12481](https://github.com/rapidsai/cudf/pull/12481)) [@divyegala](https://github.com/divyegala) -- Use rapids-cmake parallel testing feature ([#12451](https://github.com/rapidsai/cudf/pull/12451)) [@robertmaynard](https://github.com/robertmaynard) -- Enable detection of undesired stream usage ([#12089](https://github.com/rapidsai/cudf/pull/12089)) [@vyasr](https://github.com/vyasr) +- Add JNI method for strings::replace multi variety ([#12979](https://github.com/NVIDIA/cudf/pull/12979)) [@NVnavkumar](https://github.com/NVnavkumar) +- Add nunique aggregation support for cudf::segmented_reduce ([#12972](https://github.com/NVIDIA/cudf/pull/12972)) [@davidwendt](https://github.com/davidwendt) +- Refactor orc chunked writer ([#12949](https://github.com/NVIDIA/cudf/pull/12949)) [@ttnghia](https://github.com/ttnghia) +- Make Parquet writer `nullable` option application to single table writes ([#12933](https://github.com/NVIDIA/cudf/pull/12933)) [@vuule](https://github.com/vuule) +- Refactor `io::orc::ProtobufWriter` ([#12877](https://github.com/NVIDIA/cudf/pull/12877)) [@ttnghia](https://github.com/ttnghia) +- Make timezone table independent from ORC ([#12805](https://github.com/NVIDIA/cudf/pull/12805)) [@vuule](https://github.com/vuule) +- Cache JIT `GroupBy.apply` functions ([#12802](https://github.com/NVIDIA/cudf/pull/12802)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Implement initial support for avro logical types ([#6482) (#12788](https://github.com/NVIDIA/cudf/pull/6482) (#12788)) [@tpn](https://github.com/tpn) +- Update `tests/column_utilities` to use `experimental::equality` row comparator ([#12777](https://github.com/NVIDIA/cudf/pull/12777)) [@divyegala](https://github.com/divyegala) +- Update `distinct/unique_count` to `experimental::row` hasher/comparator ([#12776](https://github.com/NVIDIA/cudf/pull/12776)) [@divyegala](https://github.com/divyegala) +- Update `hash_partition` to use `experimental::row::row_hasher` ([#12761](https://github.com/NVIDIA/cudf/pull/12761)) [@divyegala](https://github.com/divyegala) +- Update `is_sorted` to use `experimental::row::lexicographic` ([#12752](https://github.com/NVIDIA/cudf/pull/12752)) [@divyegala](https://github.com/divyegala) +- Update default data source in cuio reader benchmarks ([#12740](https://github.com/NVIDIA/cudf/pull/12740)) [@PointKernel](https://github.com/PointKernel) +- Reenable stream identification library in CI ([#12714](https://github.com/NVIDIA/cudf/pull/12714)) [@vyasr](https://github.com/vyasr) +- Add `regex_program` strings splitting java APIs and tests ([#12713](https://github.com/NVIDIA/cudf/pull/12713)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Add `regex_program` strings replacing java APIs and tests ([#12701](https://github.com/NVIDIA/cudf/pull/12701)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Add `regex_program` strings extract java APIs and tests ([#12699](https://github.com/NVIDIA/cudf/pull/12699)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Variable fragment sizes for Parquet writer ([#12685](https://github.com/NVIDIA/cudf/pull/12685)) [@etseidl](https://github.com/etseidl) +- Add segmented reduction support for fixed-point types ([#12680](https://github.com/NVIDIA/cudf/pull/12680)) [@davidwendt](https://github.com/davidwendt) +- Move `strings_udf` code into cuDF ([#12669](https://github.com/NVIDIA/cudf/pull/12669)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add `regex_program` searching APIs and related java classes ([#12666](https://github.com/NVIDIA/cudf/pull/12666)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Add logging to libcudf ([#12637](https://github.com/NVIDIA/cudf/pull/12637)) [@vuule](https://github.com/vuule) +- Add compound aggregations to cudf::segmented_reduce ([#12573](https://github.com/NVIDIA/cudf/pull/12573)) [@davidwendt](https://github.com/davidwendt) +- Convert `rank` to use to experimental row comparators ([#12481](https://github.com/NVIDIA/cudf/pull/12481)) [@divyegala](https://github.com/divyegala) +- Use rapids-cmake parallel testing feature ([#12451](https://github.com/NVIDIA/cudf/pull/12451)) [@robertmaynard](https://github.com/robertmaynard) +- Enable detection of undesired stream usage ([#12089](https://github.com/NVIDIA/cudf/pull/12089)) [@vyasr](https://github.com/vyasr) ## 🛠️ Improvements -- Pin `dask` and `distributed` for release ([#13070](https://github.com/rapidsai/cudf/pull/13070)) [@galipremsagar](https://github.com/galipremsagar) -- Pin cupy in wheel tests to supported versions ([#13041](https://github.com/rapidsai/cudf/pull/13041)) [@vyasr](https://github.com/vyasr) -- Pin numba version ([#13001](https://github.com/rapidsai/cudf/pull/13001)) [@vyasr](https://github.com/vyasr) -- Rework gtests SequenceTest to remove using namepace cudf ([#12985](https://github.com/rapidsai/cudf/pull/12985)) [@davidwendt](https://github.com/davidwendt) -- Stop setting package version attribute in wheels ([#12977](https://github.com/rapidsai/cudf/pull/12977)) [@vyasr](https://github.com/vyasr) -- Move detail reduction functions to cudf::reduction::detail namespace ([#12971](https://github.com/rapidsai/cudf/pull/12971)) [@davidwendt](https://github.com/davidwendt) -- Remove default detail mrs: part7 ([#12970](https://github.com/rapidsai/cudf/pull/12970)) [@vyasr](https://github.com/vyasr) -- Remove default detail mrs: part6 ([#12969](https://github.com/rapidsai/cudf/pull/12969)) [@vyasr](https://github.com/vyasr) -- Remove default detail mrs: part5 ([#12968](https://github.com/rapidsai/cudf/pull/12968)) [@vyasr](https://github.com/vyasr) -- Remove default detail mrs: part4 ([#12967](https://github.com/rapidsai/cudf/pull/12967)) [@vyasr](https://github.com/vyasr) -- Remove default detail mrs: part3 ([#12966](https://github.com/rapidsai/cudf/pull/12966)) [@vyasr](https://github.com/vyasr) -- Remove default detail mrs: part2 ([#12965](https://github.com/rapidsai/cudf/pull/12965)) [@vyasr](https://github.com/vyasr) -- Remove default detail mrs: part1 ([#12964](https://github.com/rapidsai/cudf/pull/12964)) [@vyasr](https://github.com/vyasr) -- Add `force_nullable_schema` parameter to Parquet writer. ([#12952](https://github.com/rapidsai/cudf/pull/12952)) [@galipremsagar](https://github.com/galipremsagar) -- Declare a different name for nan_equality.UNEQUAL to prevent Cython warnings. ([#12947](https://github.com/rapidsai/cudf/pull/12947)) [@bdice](https://github.com/bdice) -- Remove remaining default stream parameters ([#12943](https://github.com/rapidsai/cudf/pull/12943)) [@vyasr](https://github.com/vyasr) -- Fix cudf::segmented_reduce gtest for ANY aggregation ([#12940](https://github.com/rapidsai/cudf/pull/12940)) [@davidwendt](https://github.com/davidwendt) -- Implement `groupby.head` and `groupby.tail` ([#12939](https://github.com/rapidsai/cudf/pull/12939)) [@wence-](https://github.com/wence-) -- Fix libcudf gtests to pass null-count=0 for empty validity masks ([#12923](https://github.com/rapidsai/cudf/pull/12923)) [@davidwendt](https://github.com/davidwendt) -- Migrate parquet encoding to use experimental row operators ([#12918](https://github.com/rapidsai/cudf/pull/12918)) [@PointKernel](https://github.com/PointKernel) -- Fix benchmarks coded in namespace cudf and using namespace cudf ([#12915](https://github.com/rapidsai/cudf/pull/12915)) [@karthikeyann](https://github.com/karthikeyann) -- Fix io/text gtests coded in namespace cudf::test ([#12914](https://github.com/rapidsai/cudf/pull/12914)) [@karthikeyann](https://github.com/karthikeyann) -- Pass `SCCACHE_S3_USE_SSL` to conda builds ([#12910](https://github.com/rapidsai/cudf/pull/12910)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix FST, JSON gtests & benchmarks coded in namespace cudf::test ([#12907](https://github.com/rapidsai/cudf/pull/12907)) [@karthikeyann](https://github.com/karthikeyann) -- Generate pyproject dependencies using dfg ([#12906](https://github.com/rapidsai/cudf/pull/12906)) [@vyasr](https://github.com/vyasr) -- Update libcudf counting functions to specify cudf::size_type ([#12904](https://github.com/rapidsai/cudf/pull/12904)) [@davidwendt](https://github.com/davidwendt) -- Fix `moto` env vars & pass `AWS_SESSION_TOKEN` to conda builds ([#12902](https://github.com/rapidsai/cudf/pull/12902)) [@ajschmidt8](https://github.com/ajschmidt8) -- Rewrite CSV writer benchmark with nvbench ([#12901](https://github.com/rapidsai/cudf/pull/12901)) [@PointKernel](https://github.com/PointKernel) -- Rework some code logic to reduce iterator and comparator inlining to improve compile time ([#12900](https://github.com/rapidsai/cudf/pull/12900)) [@davidwendt](https://github.com/davidwendt) -- Deprecate `line_terminator` in favor of `lineterminator` in `to_csv` ([#12896](https://github.com/rapidsai/cudf/pull/12896)) [@wence-](https://github.com/wence-) -- Add `stream` and `mr` parameters for `structs::detail::flatten_nested_columns` ([#12892](https://github.com/rapidsai/cudf/pull/12892)) [@ttnghia](https://github.com/ttnghia) -- Deprecate libcudf regex APIs accepting pattern strings directly ([#12891](https://github.com/rapidsai/cudf/pull/12891)) [@davidwendt](https://github.com/davidwendt) -- Remove default parameters from detail headers in include ([#12888](https://github.com/rapidsai/cudf/pull/12888)) [@vyasr](https://github.com/vyasr) -- Update minimum `pandas` and `numpy` pinnings ([#12887](https://github.com/rapidsai/cudf/pull/12887)) [@galipremsagar](https://github.com/galipremsagar) -- Implement `groupby.sample` ([#12882](https://github.com/rapidsai/cudf/pull/12882)) [@wence-](https://github.com/wence-) -- Update JNI build ENV default to gcc 11 ([#12881](https://github.com/rapidsai/cudf/pull/12881)) [@pxLi](https://github.com/pxLi) -- Change return type of `cudf::structs::detail::flatten_nested_columns` to smart pointer ([#12878](https://github.com/rapidsai/cudf/pull/12878)) [@ttnghia](https://github.com/ttnghia) -- Fix passing seed parameter to MurmurHash3_32 in cudf::hash() function ([#12875](https://github.com/rapidsai/cudf/pull/12875)) [@davidwendt](https://github.com/davidwendt) -- Remove manual artifact upload step in CI ([#12869](https://github.com/rapidsai/cudf/pull/12869)) [@ajschmidt8](https://github.com/ajschmidt8) -- Update to GCC 11 ([#12868](https://github.com/rapidsai/cudf/pull/12868)) [@bdice](https://github.com/bdice) -- Fix null hive-partition behavior in dask-cudf parquet ([#12866](https://github.com/rapidsai/cudf/pull/12866)) [@rjzamora](https://github.com/rjzamora) -- Update to protobuf>=4.21.6,<4.22. ([#12864](https://github.com/rapidsai/cudf/pull/12864)) [@bdice](https://github.com/bdice) -- Update RMM allocators ([#12861](https://github.com/rapidsai/cudf/pull/12861)) [@pentschev](https://github.com/pentschev) -- Improve performance for replace-multi for long strings ([#12858](https://github.com/rapidsai/cudf/pull/12858)) [@davidwendt](https://github.com/davidwendt) -- Drop Python 3.7 handling for pickle protocol 4 ([#12857](https://github.com/rapidsai/cudf/pull/12857)) [@jakirkham](https://github.com/jakirkham) -- Migrate as much as possible to pyproject.toml ([#12850](https://github.com/rapidsai/cudf/pull/12850)) [@vyasr](https://github.com/vyasr) -- Enable nbqa pre-commit hooks for isort and black. ([#12848](https://github.com/rapidsai/cudf/pull/12848)) [@bdice](https://github.com/bdice) -- Setting a threshold for KvikIO IO ([#12841](https://github.com/rapidsai/cudf/pull/12841)) [@madsbk](https://github.com/madsbk) -- Update datasets download URL ([#12840](https://github.com/rapidsai/cudf/pull/12840)) [@jjacobelli](https://github.com/jjacobelli) -- Make docs builds less verbose ([#12836](https://github.com/rapidsai/cudf/pull/12836)) [@AyodeAwe](https://github.com/AyodeAwe) -- Consolidate linter configs into pyproject.toml ([#12834](https://github.com/rapidsai/cudf/pull/12834)) [@vyasr](https://github.com/vyasr) -- Deprecate `names` & `dtype` in `Index.copy` ([#12825](https://github.com/rapidsai/cudf/pull/12825)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `inplace` parameters in categorical methods ([#12824](https://github.com/rapidsai/cudf/pull/12824)) [@galipremsagar](https://github.com/galipremsagar) -- Add optional text file support to ninja-log utility ([#12823](https://github.com/rapidsai/cudf/pull/12823)) [@davidwendt](https://github.com/davidwendt) -- Deprecate `Index.is_*` methods ([#12820](https://github.com/rapidsai/cudf/pull/12820)) [@galipremsagar](https://github.com/galipremsagar) -- Add dfg as a pre-commit hook ([#12819](https://github.com/rapidsai/cudf/pull/12819)) [@vyasr](https://github.com/vyasr) -- Deprecate `datetime_is_numeric` from `describe` ([#12818](https://github.com/rapidsai/cudf/pull/12818)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `na_sentinel` in `factorize` ([#12817](https://github.com/rapidsai/cudf/pull/12817)) [@galipremsagar](https://github.com/galipremsagar) -- Shuffling read into a sub function in parquet read ([#12809](https://github.com/rapidsai/cudf/pull/12809)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fixing parquet coalescing of reads ([#12808](https://github.com/rapidsai/cudf/pull/12808)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- CI: Remove specification of manual stage for check_style.sh script. ([#12803](https://github.com/rapidsai/cudf/pull/12803)) [@csadorf](https://github.com/csadorf) -- Add compute-sanitizer github workflow action to nightly tests ([#12800](https://github.com/rapidsai/cudf/pull/12800)) [@davidwendt](https://github.com/davidwendt) -- Enable groupby std and variance aggregation types in libcudf Debug build ([#12799](https://github.com/rapidsai/cudf/pull/12799)) [@davidwendt](https://github.com/davidwendt) -- Expose seed argument to hash_values ([#12795](https://github.com/rapidsai/cudf/pull/12795)) [@ayushdg](https://github.com/ayushdg) -- Fix groupby gtests coded in namespace cudf::test ([#12784](https://github.com/rapidsai/cudf/pull/12784)) [@davidwendt](https://github.com/davidwendt) -- Improve performance for cudf::strings::count_characters for long strings ([#12779](https://github.com/rapidsai/cudf/pull/12779)) [@davidwendt](https://github.com/davidwendt) -- Deallocate encoded data in ORC writer immediately after compression ([#12770](https://github.com/rapidsai/cudf/pull/12770)) [@vuule](https://github.com/vuule) -- Stop force pulling fmt in nvbench. ([#12768](https://github.com/rapidsai/cudf/pull/12768)) [@vyasr](https://github.com/vyasr) -- Remove now redundant cuda initialization ([#12758](https://github.com/rapidsai/cudf/pull/12758)) [@vyasr](https://github.com/vyasr) -- Adds JSON reader, writer io benchmark ([#12753](https://github.com/rapidsai/cudf/pull/12753)) [@karthikeyann](https://github.com/karthikeyann) -- Use test paths relative to package directory. ([#12751](https://github.com/rapidsai/cudf/pull/12751)) [@bdice](https://github.com/bdice) -- Add build metrics report as artifact to cpp-build workflow ([#12750](https://github.com/rapidsai/cudf/pull/12750)) [@davidwendt](https://github.com/davidwendt) -- Add JNI methods for detecting and purging non-empty nulls from LIST and STRUCT ([#12742](https://github.com/rapidsai/cudf/pull/12742)) [@razajafri](https://github.com/razajafri) -- Stop using versioneer to manage versions ([#12741](https://github.com/rapidsai/cudf/pull/12741)) [@vyasr](https://github.com/vyasr) -- Reduce error handling verbosity in CI tests scripts ([#12738](https://github.com/rapidsai/cudf/pull/12738)) [@AjayThorve](https://github.com/AjayThorve) -- Reduce the number of test cases in multibyte_split benchmark ([#12737](https://github.com/rapidsai/cudf/pull/12737)) [@PointKernel](https://github.com/PointKernel) -- Update shared workflow branches ([#12733](https://github.com/rapidsai/cudf/pull/12733)) [@ajschmidt8](https://github.com/ajschmidt8) -- JNI switches to nested JSON reader ([#12732](https://github.com/rapidsai/cudf/pull/12732)) [@res-life](https://github.com/res-life) -- Changing `cudf::io::source_info` to use `cudf::host_span<std::byte>` in a non-breaking form ([#12730](https://github.com/rapidsai/cudf/pull/12730)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add nvbench environment class for initializing RMM in benchmarks ([#12728](https://github.com/rapidsai/cudf/pull/12728)) [@davidwendt](https://github.com/davidwendt) -- Split C++ and Python build dependencies into separate lists. ([#12724](https://github.com/rapidsai/cudf/pull/12724)) [@bdice](https://github.com/bdice) -- Add build dependencies to Java tests. ([#12723](https://github.com/rapidsai/cudf/pull/12723)) [@bdice](https://github.com/bdice) -- Allow setting the seed argument for hash partition ([#12715](https://github.com/rapidsai/cudf/pull/12715)) [@firestarman](https://github.com/firestarman) -- Remove gpuCI scripts. ([#12712](https://github.com/rapidsai/cudf/pull/12712)) [@bdice](https://github.com/bdice) -- Unpin `dask` and `distributed` for development ([#12710](https://github.com/rapidsai/cudf/pull/12710)) [@galipremsagar](https://github.com/galipremsagar) -- `partition_by_hash()`: use `_split()` ([#12704](https://github.com/rapidsai/cudf/pull/12704)) [@madsbk](https://github.com/madsbk) -- Remove DataFrame.quantiles from docs. ([#12684](https://github.com/rapidsai/cudf/pull/12684)) [@bdice](https://github.com/bdice) -- Fast path for `experimental::row::equality` ([#12676](https://github.com/rapidsai/cudf/pull/12676)) [@divyegala](https://github.com/divyegala) -- Move date to build string in `conda` recipe ([#12661](https://github.com/rapidsai/cudf/pull/12661)) [@ajschmidt8](https://github.com/ajschmidt8) -- Refactor reduction logic for fixed-point types ([#12652](https://github.com/rapidsai/cudf/pull/12652)) [@davidwendt](https://github.com/davidwendt) -- Pay off some JNI RMM API tech debt ([#12632](https://github.com/rapidsai/cudf/pull/12632)) [@revans2](https://github.com/revans2) -- Merge `copy-on-write` feature branch into `branch-23.04` ([#12619](https://github.com/rapidsai/cudf/pull/12619)) [@galipremsagar](https://github.com/galipremsagar) -- Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings ([#12609](https://github.com/rapidsai/cudf/pull/12609)) [@davidwendt](https://github.com/davidwendt) -- Pin cuda-nvrtc. ([#12606](https://github.com/rapidsai/cudf/pull/12606)) [@bdice](https://github.com/bdice) -- Remove cudf::test::print calls from libcudf gtests ([#12604](https://github.com/rapidsai/cudf/pull/12604)) [@davidwendt](https://github.com/davidwendt) -- Init JNI version 23.04.0-SNAPSHOT ([#12599](https://github.com/rapidsai/cudf/pull/12599)) [@pxLi](https://github.com/pxLi) -- Add performance benchmarks to user facing docs ([#12595](https://github.com/rapidsai/cudf/pull/12595)) [@galipremsagar](https://github.com/galipremsagar) -- Add docs build job ([#12592](https://github.com/rapidsai/cudf/pull/12592)) [@AyodeAwe](https://github.com/AyodeAwe) -- Replace message parsing with throwing more specific exceptions ([#12426](https://github.com/rapidsai/cudf/pull/12426)) [@vyasr](https://github.com/vyasr) -- Support conversion to/from cudf in dask.dataframe.core.to_backend ([#12380](https://github.com/rapidsai/cudf/pull/12380)) [@rjzamora](https://github.com/rjzamora) +- Pin `dask` and `distributed` for release ([#13070](https://github.com/NVIDIA/cudf/pull/13070)) [@galipremsagar](https://github.com/galipremsagar) +- Pin cupy in wheel tests to supported versions ([#13041](https://github.com/NVIDIA/cudf/pull/13041)) [@vyasr](https://github.com/vyasr) +- Pin numba version ([#13001](https://github.com/NVIDIA/cudf/pull/13001)) [@vyasr](https://github.com/vyasr) +- Rework gtests SequenceTest to remove using namepace cudf ([#12985](https://github.com/NVIDIA/cudf/pull/12985)) [@davidwendt](https://github.com/davidwendt) +- Stop setting package version attribute in wheels ([#12977](https://github.com/NVIDIA/cudf/pull/12977)) [@vyasr](https://github.com/vyasr) +- Move detail reduction functions to cudf::reduction::detail namespace ([#12971](https://github.com/NVIDIA/cudf/pull/12971)) [@davidwendt](https://github.com/davidwendt) +- Remove default detail mrs: part7 ([#12970](https://github.com/NVIDIA/cudf/pull/12970)) [@vyasr](https://github.com/vyasr) +- Remove default detail mrs: part6 ([#12969](https://github.com/NVIDIA/cudf/pull/12969)) [@vyasr](https://github.com/vyasr) +- Remove default detail mrs: part5 ([#12968](https://github.com/NVIDIA/cudf/pull/12968)) [@vyasr](https://github.com/vyasr) +- Remove default detail mrs: part4 ([#12967](https://github.com/NVIDIA/cudf/pull/12967)) [@vyasr](https://github.com/vyasr) +- Remove default detail mrs: part3 ([#12966](https://github.com/NVIDIA/cudf/pull/12966)) [@vyasr](https://github.com/vyasr) +- Remove default detail mrs: part2 ([#12965](https://github.com/NVIDIA/cudf/pull/12965)) [@vyasr](https://github.com/vyasr) +- Remove default detail mrs: part1 ([#12964](https://github.com/NVIDIA/cudf/pull/12964)) [@vyasr](https://github.com/vyasr) +- Add `force_nullable_schema` parameter to Parquet writer. ([#12952](https://github.com/NVIDIA/cudf/pull/12952)) [@galipremsagar](https://github.com/galipremsagar) +- Declare a different name for nan_equality.UNEQUAL to prevent Cython warnings. ([#12947](https://github.com/NVIDIA/cudf/pull/12947)) [@bdice](https://github.com/bdice) +- Remove remaining default stream parameters ([#12943](https://github.com/NVIDIA/cudf/pull/12943)) [@vyasr](https://github.com/vyasr) +- Fix cudf::segmented_reduce gtest for ANY aggregation ([#12940](https://github.com/NVIDIA/cudf/pull/12940)) [@davidwendt](https://github.com/davidwendt) +- Implement `groupby.head` and `groupby.tail` ([#12939](https://github.com/NVIDIA/cudf/pull/12939)) [@wence-](https://github.com/wence-) +- Fix libcudf gtests to pass null-count=0 for empty validity masks ([#12923](https://github.com/NVIDIA/cudf/pull/12923)) [@davidwendt](https://github.com/davidwendt) +- Migrate parquet encoding to use experimental row operators ([#12918](https://github.com/NVIDIA/cudf/pull/12918)) [@PointKernel](https://github.com/PointKernel) +- Fix benchmarks coded in namespace cudf and using namespace cudf ([#12915](https://github.com/NVIDIA/cudf/pull/12915)) [@karthikeyann](https://github.com/karthikeyann) +- Fix io/text gtests coded in namespace cudf::test ([#12914](https://github.com/NVIDIA/cudf/pull/12914)) [@karthikeyann](https://github.com/karthikeyann) +- Pass `SCCACHE_S3_USE_SSL` to conda builds ([#12910](https://github.com/NVIDIA/cudf/pull/12910)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix FST, JSON gtests & benchmarks coded in namespace cudf::test ([#12907](https://github.com/NVIDIA/cudf/pull/12907)) [@karthikeyann](https://github.com/karthikeyann) +- Generate pyproject dependencies using dfg ([#12906](https://github.com/NVIDIA/cudf/pull/12906)) [@vyasr](https://github.com/vyasr) +- Update libcudf counting functions to specify cudf::size_type ([#12904](https://github.com/NVIDIA/cudf/pull/12904)) [@davidwendt](https://github.com/davidwendt) +- Fix `moto` env vars & pass `AWS_SESSION_TOKEN` to conda builds ([#12902](https://github.com/NVIDIA/cudf/pull/12902)) [@ajschmidt8](https://github.com/ajschmidt8) +- Rewrite CSV writer benchmark with nvbench ([#12901](https://github.com/NVIDIA/cudf/pull/12901)) [@PointKernel](https://github.com/PointKernel) +- Rework some code logic to reduce iterator and comparator inlining to improve compile time ([#12900](https://github.com/NVIDIA/cudf/pull/12900)) [@davidwendt](https://github.com/davidwendt) +- Deprecate `line_terminator` in favor of `lineterminator` in `to_csv` ([#12896](https://github.com/NVIDIA/cudf/pull/12896)) [@wence-](https://github.com/wence-) +- Add `stream` and `mr` parameters for `structs::detail::flatten_nested_columns` ([#12892](https://github.com/NVIDIA/cudf/pull/12892)) [@ttnghia](https://github.com/ttnghia) +- Deprecate libcudf regex APIs accepting pattern strings directly ([#12891](https://github.com/NVIDIA/cudf/pull/12891)) [@davidwendt](https://github.com/davidwendt) +- Remove default parameters from detail headers in include ([#12888](https://github.com/NVIDIA/cudf/pull/12888)) [@vyasr](https://github.com/vyasr) +- Update minimum `pandas` and `numpy` pinnings ([#12887](https://github.com/NVIDIA/cudf/pull/12887)) [@galipremsagar](https://github.com/galipremsagar) +- Implement `groupby.sample` ([#12882](https://github.com/NVIDIA/cudf/pull/12882)) [@wence-](https://github.com/wence-) +- Update JNI build ENV default to gcc 11 ([#12881](https://github.com/NVIDIA/cudf/pull/12881)) [@pxLi](https://github.com/pxLi) +- Change return type of `cudf::structs::detail::flatten_nested_columns` to smart pointer ([#12878](https://github.com/NVIDIA/cudf/pull/12878)) [@ttnghia](https://github.com/ttnghia) +- Fix passing seed parameter to MurmurHash3_32 in cudf::hash() function ([#12875](https://github.com/NVIDIA/cudf/pull/12875)) [@davidwendt](https://github.com/davidwendt) +- Remove manual artifact upload step in CI ([#12869](https://github.com/NVIDIA/cudf/pull/12869)) [@ajschmidt8](https://github.com/ajschmidt8) +- Update to GCC 11 ([#12868](https://github.com/NVIDIA/cudf/pull/12868)) [@bdice](https://github.com/bdice) +- Fix null hive-partition behavior in dask-cudf parquet ([#12866](https://github.com/NVIDIA/cudf/pull/12866)) [@rjzamora](https://github.com/rjzamora) +- Update to protobuf>=4.21.6,<4.22. ([#12864](https://github.com/NVIDIA/cudf/pull/12864)) [@bdice](https://github.com/bdice) +- Update RMM allocators ([#12861](https://github.com/NVIDIA/cudf/pull/12861)) [@pentschev](https://github.com/pentschev) +- Improve performance for replace-multi for long strings ([#12858](https://github.com/NVIDIA/cudf/pull/12858)) [@davidwendt](https://github.com/davidwendt) +- Drop Python 3.7 handling for pickle protocol 4 ([#12857](https://github.com/NVIDIA/cudf/pull/12857)) [@jakirkham](https://github.com/jakirkham) +- Migrate as much as possible to pyproject.toml ([#12850](https://github.com/NVIDIA/cudf/pull/12850)) [@vyasr](https://github.com/vyasr) +- Enable nbqa pre-commit hooks for isort and black. ([#12848](https://github.com/NVIDIA/cudf/pull/12848)) [@bdice](https://github.com/bdice) +- Setting a threshold for KvikIO IO ([#12841](https://github.com/NVIDIA/cudf/pull/12841)) [@madsbk](https://github.com/madsbk) +- Update datasets download URL ([#12840](https://github.com/NVIDIA/cudf/pull/12840)) [@jjacobelli](https://github.com/jjacobelli) +- Make docs builds less verbose ([#12836](https://github.com/NVIDIA/cudf/pull/12836)) [@AyodeAwe](https://github.com/AyodeAwe) +- Consolidate linter configs into pyproject.toml ([#12834](https://github.com/NVIDIA/cudf/pull/12834)) [@vyasr](https://github.com/vyasr) +- Deprecate `names` & `dtype` in `Index.copy` ([#12825](https://github.com/NVIDIA/cudf/pull/12825)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `inplace` parameters in categorical methods ([#12824](https://github.com/NVIDIA/cudf/pull/12824)) [@galipremsagar](https://github.com/galipremsagar) +- Add optional text file support to ninja-log utility ([#12823](https://github.com/NVIDIA/cudf/pull/12823)) [@davidwendt](https://github.com/davidwendt) +- Deprecate `Index.is_*` methods ([#12820](https://github.com/NVIDIA/cudf/pull/12820)) [@galipremsagar](https://github.com/galipremsagar) +- Add dfg as a pre-commit hook ([#12819](https://github.com/NVIDIA/cudf/pull/12819)) [@vyasr](https://github.com/vyasr) +- Deprecate `datetime_is_numeric` from `describe` ([#12818](https://github.com/NVIDIA/cudf/pull/12818)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `na_sentinel` in `factorize` ([#12817](https://github.com/NVIDIA/cudf/pull/12817)) [@galipremsagar](https://github.com/galipremsagar) +- Shuffling read into a sub function in parquet read ([#12809](https://github.com/NVIDIA/cudf/pull/12809)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fixing parquet coalescing of reads ([#12808](https://github.com/NVIDIA/cudf/pull/12808)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- CI: Remove specification of manual stage for check_style.sh script. ([#12803](https://github.com/NVIDIA/cudf/pull/12803)) [@csadorf](https://github.com/csadorf) +- Add compute-sanitizer github workflow action to nightly tests ([#12800](https://github.com/NVIDIA/cudf/pull/12800)) [@davidwendt](https://github.com/davidwendt) +- Enable groupby std and variance aggregation types in libcudf Debug build ([#12799](https://github.com/NVIDIA/cudf/pull/12799)) [@davidwendt](https://github.com/davidwendt) +- Expose seed argument to hash_values ([#12795](https://github.com/NVIDIA/cudf/pull/12795)) [@ayushdg](https://github.com/ayushdg) +- Fix groupby gtests coded in namespace cudf::test ([#12784](https://github.com/NVIDIA/cudf/pull/12784)) [@davidwendt](https://github.com/davidwendt) +- Improve performance for cudf::strings::count_characters for long strings ([#12779](https://github.com/NVIDIA/cudf/pull/12779)) [@davidwendt](https://github.com/davidwendt) +- Deallocate encoded data in ORC writer immediately after compression ([#12770](https://github.com/NVIDIA/cudf/pull/12770)) [@vuule](https://github.com/vuule) +- Stop force pulling fmt in nvbench. ([#12768](https://github.com/NVIDIA/cudf/pull/12768)) [@vyasr](https://github.com/vyasr) +- Remove now redundant cuda initialization ([#12758](https://github.com/NVIDIA/cudf/pull/12758)) [@vyasr](https://github.com/vyasr) +- Adds JSON reader, writer io benchmark ([#12753](https://github.com/NVIDIA/cudf/pull/12753)) [@karthikeyann](https://github.com/karthikeyann) +- Use test paths relative to package directory. ([#12751](https://github.com/NVIDIA/cudf/pull/12751)) [@bdice](https://github.com/bdice) +- Add build metrics report as artifact to cpp-build workflow ([#12750](https://github.com/NVIDIA/cudf/pull/12750)) [@davidwendt](https://github.com/davidwendt) +- Add JNI methods for detecting and purging non-empty nulls from LIST and STRUCT ([#12742](https://github.com/NVIDIA/cudf/pull/12742)) [@razajafri](https://github.com/razajafri) +- Stop using versioneer to manage versions ([#12741](https://github.com/NVIDIA/cudf/pull/12741)) [@vyasr](https://github.com/vyasr) +- Reduce error handling verbosity in CI tests scripts ([#12738](https://github.com/NVIDIA/cudf/pull/12738)) [@AjayThorve](https://github.com/AjayThorve) +- Reduce the number of test cases in multibyte_split benchmark ([#12737](https://github.com/NVIDIA/cudf/pull/12737)) [@PointKernel](https://github.com/PointKernel) +- Update shared workflow branches ([#12733](https://github.com/NVIDIA/cudf/pull/12733)) [@ajschmidt8](https://github.com/ajschmidt8) +- JNI switches to nested JSON reader ([#12732](https://github.com/NVIDIA/cudf/pull/12732)) [@res-life](https://github.com/res-life) +- Changing `cudf::io::source_info` to use `cudf::host_span<std::byte>` in a non-breaking form ([#12730](https://github.com/NVIDIA/cudf/pull/12730)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add nvbench environment class for initializing RMM in benchmarks ([#12728](https://github.com/NVIDIA/cudf/pull/12728)) [@davidwendt](https://github.com/davidwendt) +- Split C++ and Python build dependencies into separate lists. ([#12724](https://github.com/NVIDIA/cudf/pull/12724)) [@bdice](https://github.com/bdice) +- Add build dependencies to Java tests. ([#12723](https://github.com/NVIDIA/cudf/pull/12723)) [@bdice](https://github.com/bdice) +- Allow setting the seed argument for hash partition ([#12715](https://github.com/NVIDIA/cudf/pull/12715)) [@firestarman](https://github.com/firestarman) +- Remove gpuCI scripts. ([#12712](https://github.com/NVIDIA/cudf/pull/12712)) [@bdice](https://github.com/bdice) +- Unpin `dask` and `distributed` for development ([#12710](https://github.com/NVIDIA/cudf/pull/12710)) [@galipremsagar](https://github.com/galipremsagar) +- `partition_by_hash()`: use `_split()` ([#12704](https://github.com/NVIDIA/cudf/pull/12704)) [@madsbk](https://github.com/madsbk) +- Remove DataFrame.quantiles from docs. ([#12684](https://github.com/NVIDIA/cudf/pull/12684)) [@bdice](https://github.com/bdice) +- Fast path for `experimental::row::equality` ([#12676](https://github.com/NVIDIA/cudf/pull/12676)) [@divyegala](https://github.com/divyegala) +- Move date to build string in `conda` recipe ([#12661](https://github.com/NVIDIA/cudf/pull/12661)) [@ajschmidt8](https://github.com/ajschmidt8) +- Refactor reduction logic for fixed-point types ([#12652](https://github.com/NVIDIA/cudf/pull/12652)) [@davidwendt](https://github.com/davidwendt) +- Pay off some JNI RMM API tech debt ([#12632](https://github.com/NVIDIA/cudf/pull/12632)) [@revans2](https://github.com/revans2) +- Merge `copy-on-write` feature branch into `branch-23.04` ([#12619](https://github.com/NVIDIA/cudf/pull/12619)) [@galipremsagar](https://github.com/galipremsagar) +- Remove cudf::strings::repeat_strings_output_sizes and optional parameter from cudf::strings::repeat_strings ([#12609](https://github.com/NVIDIA/cudf/pull/12609)) [@davidwendt](https://github.com/davidwendt) +- Pin cuda-nvrtc. ([#12606](https://github.com/NVIDIA/cudf/pull/12606)) [@bdice](https://github.com/bdice) +- Remove cudf::test::print calls from libcudf gtests ([#12604](https://github.com/NVIDIA/cudf/pull/12604)) [@davidwendt](https://github.com/davidwendt) +- Init JNI version 23.04.0-SNAPSHOT ([#12599](https://github.com/NVIDIA/cudf/pull/12599)) [@pxLi](https://github.com/pxLi) +- Add performance benchmarks to user facing docs ([#12595](https://github.com/NVIDIA/cudf/pull/12595)) [@galipremsagar](https://github.com/galipremsagar) +- Add docs build job ([#12592](https://github.com/NVIDIA/cudf/pull/12592)) [@AyodeAwe](https://github.com/AyodeAwe) +- Replace message parsing with throwing more specific exceptions ([#12426](https://github.com/NVIDIA/cudf/pull/12426)) [@vyasr](https://github.com/vyasr) +- Support conversion to/from cudf in dask.dataframe.core.to_backend ([#12380](https://github.com/NVIDIA/cudf/pull/12380)) [@rjzamora](https://github.com/rjzamora) # cuDF 23.02.00 (9 Feb 2023) ## 🚨 Breaking Changes -- Pin `dask` and `distributed` for release ([#12695](https://github.com/rapidsai/cudf/pull/12695)) [@galipremsagar](https://github.com/galipremsagar) -- Change ways to access `ptr` in `Buffer` ([#12587](https://github.com/rapidsai/cudf/pull/12587)) [@galipremsagar](https://github.com/galipremsagar) -- Remove column names ([#12578](https://github.com/rapidsai/cudf/pull/12578)) [@vuule](https://github.com/vuule) -- Default `cudf::io::read_json` to nested JSON parser ([#12544](https://github.com/rapidsai/cudf/pull/12544)) [@vuule](https://github.com/vuule) -- Switch `engine=cudf` to the new `JSON` reader ([#12509](https://github.com/rapidsai/cudf/pull/12509)) [@galipremsagar](https://github.com/galipremsagar) -- Add trailing comma support for nested JSON reader ([#12448](https://github.com/rapidsai/cudf/pull/12448)) [@karthikeyann](https://github.com/karthikeyann) -- Upgrade to `arrow-10.0.1` ([#12327](https://github.com/rapidsai/cudf/pull/12327)) [@galipremsagar](https://github.com/galipremsagar) -- Fail loudly to avoid data corruption with unsupported input in `read_orc` ([#12325](https://github.com/rapidsai/cudf/pull/12325)) [@vuule](https://github.com/vuule) -- CSV, JSON reader to infer integer column with nulls as int64 instead of float64 ([#12309](https://github.com/rapidsai/cudf/pull/12309)) [@karthikeyann](https://github.com/karthikeyann) -- Remove deprecated code for 23.02 ([#12281](https://github.com/rapidsai/cudf/pull/12281)) [@vyasr](https://github.com/vyasr) -- Null element for parsing error in numeric types in JSON, CSV reader ([#12272](https://github.com/rapidsai/cudf/pull/12272)) [@karthikeyann](https://github.com/karthikeyann) -- Purge non-empty nulls for `superimpose_nulls` and `push_down_nulls` ([#12239](https://github.com/rapidsai/cudf/pull/12239)) [@ttnghia](https://github.com/ttnghia) -- Rename `cudf::structs::detail::superimpose_parent_nulls` APIs ([#12230](https://github.com/rapidsai/cudf/pull/12230)) [@ttnghia](https://github.com/ttnghia) -- Remove JIT type names, refactor id_to_type. ([#12158](https://github.com/rapidsai/cudf/pull/12158)) [@bdice](https://github.com/bdice) -- Floor division uses integer division for integral arguments ([#12131](https://github.com/rapidsai/cudf/pull/12131)) [@wence-](https://github.com/wence-) +- Pin `dask` and `distributed` for release ([#12695](https://github.com/NVIDIA/cudf/pull/12695)) [@galipremsagar](https://github.com/galipremsagar) +- Change ways to access `ptr` in `Buffer` ([#12587](https://github.com/NVIDIA/cudf/pull/12587)) [@galipremsagar](https://github.com/galipremsagar) +- Remove column names ([#12578](https://github.com/NVIDIA/cudf/pull/12578)) [@vuule](https://github.com/vuule) +- Default `cudf::io::read_json` to nested JSON parser ([#12544](https://github.com/NVIDIA/cudf/pull/12544)) [@vuule](https://github.com/vuule) +- Switch `engine=cudf` to the new `JSON` reader ([#12509](https://github.com/NVIDIA/cudf/pull/12509)) [@galipremsagar](https://github.com/galipremsagar) +- Add trailing comma support for nested JSON reader ([#12448](https://github.com/NVIDIA/cudf/pull/12448)) [@karthikeyann](https://github.com/karthikeyann) +- Upgrade to `arrow-10.0.1` ([#12327](https://github.com/NVIDIA/cudf/pull/12327)) [@galipremsagar](https://github.com/galipremsagar) +- Fail loudly to avoid data corruption with unsupported input in `read_orc` ([#12325](https://github.com/NVIDIA/cudf/pull/12325)) [@vuule](https://github.com/vuule) +- CSV, JSON reader to infer integer column with nulls as int64 instead of float64 ([#12309](https://github.com/NVIDIA/cudf/pull/12309)) [@karthikeyann](https://github.com/karthikeyann) +- Remove deprecated code for 23.02 ([#12281](https://github.com/NVIDIA/cudf/pull/12281)) [@vyasr](https://github.com/vyasr) +- Null element for parsing error in numeric types in JSON, CSV reader ([#12272](https://github.com/NVIDIA/cudf/pull/12272)) [@karthikeyann](https://github.com/karthikeyann) +- Purge non-empty nulls for `superimpose_nulls` and `push_down_nulls` ([#12239](https://github.com/NVIDIA/cudf/pull/12239)) [@ttnghia](https://github.com/ttnghia) +- Rename `cudf::structs::detail::superimpose_parent_nulls` APIs ([#12230](https://github.com/NVIDIA/cudf/pull/12230)) [@ttnghia](https://github.com/ttnghia) +- Remove JIT type names, refactor id_to_type. ([#12158](https://github.com/NVIDIA/cudf/pull/12158)) [@bdice](https://github.com/bdice) +- Floor division uses integer division for integral arguments ([#12131](https://github.com/NVIDIA/cudf/pull/12131)) [@wence-](https://github.com/wence-) ## 🐛 Bug Fixes -- Fix a mask data corruption in UDF ([#12647](https://github.com/rapidsai/cudf/pull/12647)) [@galipremsagar](https://github.com/galipremsagar) -- pre-commit: Update isort version to 5.12.0 ([#12645](https://github.com/rapidsai/cudf/pull/12645)) [@wence-](https://github.com/wence-) -- tests: Skip cuInit tests if cuda-gdb is not found or not working ([#12644](https://github.com/rapidsai/cudf/pull/12644)) [@wence-](https://github.com/wence-) -- Revert regex program java APIs and tests ([#12639](https://github.com/rapidsai/cudf/pull/12639)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Fix leaks in ColumnVectorTest ([#12625](https://github.com/rapidsai/cudf/pull/12625)) [@jlowe](https://github.com/jlowe) -- Handle when spillable buffers own each other ([#12607](https://github.com/rapidsai/cudf/pull/12607)) [@madsbk](https://github.com/madsbk) -- Fix incorrect null counts for sliced columns in JCudfSerialization ([#12589](https://github.com/rapidsai/cudf/pull/12589)) [@jlowe](https://github.com/jlowe) -- lists: Transfer dtypes correctly through list.get ([#12586](https://github.com/rapidsai/cudf/pull/12586)) [@wence-](https://github.com/wence-) -- timedelta: Don't go via float intermediates for floordiv ([#12585](https://github.com/rapidsai/cudf/pull/12585)) [@wence-](https://github.com/wence-) -- Fixing BUG, `get_next_chunk()` should use the blocking function `device_read()` ([#12584](https://github.com/rapidsai/cudf/pull/12584)) [@madsbk](https://github.com/madsbk) -- Make JNI QuoteStyle accessible outside ai.rapids.cudf ([#12572](https://github.com/rapidsai/cudf/pull/12572)) [@mythrocks](https://github.com/mythrocks) -- `partition_by_hash()`: support index ([#12554](https://github.com/rapidsai/cudf/pull/12554)) [@madsbk](https://github.com/madsbk) -- Mixed Join benchmark bug due to wrong conditional column ([#12553](https://github.com/rapidsai/cudf/pull/12553)) [@divyegala](https://github.com/divyegala) -- Update List Lexicographical Comparator ([#12538](https://github.com/rapidsai/cudf/pull/12538)) [@divyegala](https://github.com/divyegala) -- Dynamically read PTX version ([#12534](https://github.com/rapidsai/cudf/pull/12534)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- build.sh switch to use `RAPIDS` magic value ([#12525](https://github.com/rapidsai/cudf/pull/12525)) [@robertmaynard](https://github.com/robertmaynard) -- Loosen runtime arrow pinning ([#12522](https://github.com/rapidsai/cudf/pull/12522)) [@vyasr](https://github.com/vyasr) -- Enable metadata transfer for complex types in transpose ([#12491](https://github.com/rapidsai/cudf/pull/12491)) [@galipremsagar](https://github.com/galipremsagar) -- Fix issues with parquet chunked reader ([#12488](https://github.com/rapidsai/cudf/pull/12488)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix missing metadata transfer in concat for `ListColumn` ([#12487](https://github.com/rapidsai/cudf/pull/12487)) [@galipremsagar](https://github.com/galipremsagar) -- Rename libcudf substring source files to slice ([#12484](https://github.com/rapidsai/cudf/pull/12484)) [@davidwendt](https://github.com/davidwendt) -- Fix compile issue with arrow 10 ([#12465](https://github.com/rapidsai/cudf/pull/12465)) [@ttnghia](https://github.com/ttnghia) -- Fix List offsets bug in mixed type list column in nested JSON reader ([#12447](https://github.com/rapidsai/cudf/pull/12447)) [@karthikeyann](https://github.com/karthikeyann) -- Fix xfail incompatibilities ([#12423](https://github.com/rapidsai/cudf/pull/12423)) [@vyasr](https://github.com/vyasr) -- Fix bug in Parquet column index encoding ([#12404](https://github.com/rapidsai/cudf/pull/12404)) [@etseidl](https://github.com/etseidl) -- When building Arrow shared look for a shared OpenSSL ([#12396](https://github.com/rapidsai/cudf/pull/12396)) [@robertmaynard](https://github.com/robertmaynard) -- Fix get_json_object to return empty column on empty input ([#12384](https://github.com/rapidsai/cudf/pull/12384)) [@davidwendt](https://github.com/davidwendt) -- Pin arrow 9 in testing dependencies to prevent conda solve issues ([#12377](https://github.com/rapidsai/cudf/pull/12377)) [@vyasr](https://github.com/vyasr) -- Fix reductions any/all return value for empty input ([#12374](https://github.com/rapidsai/cudf/pull/12374)) [@davidwendt](https://github.com/davidwendt) -- Fix debug compile errors in parquet.hpp ([#12372](https://github.com/rapidsai/cudf/pull/12372)) [@davidwendt](https://github.com/davidwendt) -- Purge non-empty nulls in `cudf::make_lists_column` ([#12370](https://github.com/rapidsai/cudf/pull/12370)) [@ttnghia](https://github.com/ttnghia) -- Use correct memory resource in io::make_column ([#12364](https://github.com/rapidsai/cudf/pull/12364)) [@vyasr](https://github.com/vyasr) -- Add code to detect possible malformed page data in parquet files. ([#12360](https://github.com/rapidsai/cudf/pull/12360)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fail loudly to avoid data corruption with unsupported input in `read_orc` ([#12325](https://github.com/rapidsai/cudf/pull/12325)) [@vuule](https://github.com/vuule) -- Fix NumericPairIteratorTest for float values ([#12306](https://github.com/rapidsai/cudf/pull/12306)) [@davidwendt](https://github.com/davidwendt) -- Fixes memory allocation in nested JSON tokenizer ([#12300](https://github.com/rapidsai/cudf/pull/12300)) [@elstehle](https://github.com/elstehle) -- Reconstruct dtypes correctly for list aggs of struct columns ([#12290](https://github.com/rapidsai/cudf/pull/12290)) [@wence-](https://github.com/wence-) -- Fix regex \A and \Z to strictly match string begin/end ([#12282](https://github.com/rapidsai/cudf/pull/12282)) [@davidwendt](https://github.com/davidwendt) -- Fix compile issue in `json_chunked_reader.cpp` ([#12280](https://github.com/rapidsai/cudf/pull/12280)) [@ttnghia](https://github.com/ttnghia) -- Change reductions any/all to return valid values for empty input ([#12279](https://github.com/rapidsai/cudf/pull/12279)) [@davidwendt](https://github.com/davidwendt) -- Only exclude join keys that are indices from key columns ([#12271](https://github.com/rapidsai/cudf/pull/12271)) [@wence-](https://github.com/wence-) -- Fix spill to device limit ([#12252](https://github.com/rapidsai/cudf/pull/12252)) [@madsbk](https://github.com/madsbk) -- Correct behaviour of sort in `concat` for singleton concatenations ([#12247](https://github.com/rapidsai/cudf/pull/12247)) [@wence-](https://github.com/wence-) -- Purge non-empty nulls for `superimpose_nulls` and `push_down_nulls` ([#12239](https://github.com/rapidsai/cudf/pull/12239)) [@ttnghia](https://github.com/ttnghia) -- Patch CUB DeviceSegmentedSort and remove workaround ([#12234](https://github.com/rapidsai/cudf/pull/12234)) [@davidwendt](https://github.com/davidwendt) -- Fix memory leak in udf_string::assign(&&) function ([#12206](https://github.com/rapidsai/cudf/pull/12206)) [@davidwendt](https://github.com/davidwendt) -- Workaround thrust-copy-if limit in json get_tree_representation ([#12190](https://github.com/rapidsai/cudf/pull/12190)) [@davidwendt](https://github.com/davidwendt) -- Fix page size calculation in Parquet writer ([#12182](https://github.com/rapidsai/cudf/pull/12182)) [@etseidl](https://github.com/etseidl) -- Add cudf::detail::sizes_to_offsets_iterator to allow checking overflow in offsets ([#12180](https://github.com/rapidsai/cudf/pull/12180)) [@davidwendt](https://github.com/davidwendt) -- Workaround thrust-copy-if limit in wordpiece-tokenizer ([#12168](https://github.com/rapidsai/cudf/pull/12168)) [@davidwendt](https://github.com/davidwendt) -- Floor division uses integer division for integral arguments ([#12131](https://github.com/rapidsai/cudf/pull/12131)) [@wence-](https://github.com/wence-) +- Fix a mask data corruption in UDF ([#12647](https://github.com/NVIDIA/cudf/pull/12647)) [@galipremsagar](https://github.com/galipremsagar) +- pre-commit: Update isort version to 5.12.0 ([#12645](https://github.com/NVIDIA/cudf/pull/12645)) [@wence-](https://github.com/wence-) +- tests: Skip cuInit tests if cuda-gdb is not found or not working ([#12644](https://github.com/NVIDIA/cudf/pull/12644)) [@wence-](https://github.com/wence-) +- Revert regex program java APIs and tests ([#12639](https://github.com/NVIDIA/cudf/pull/12639)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Fix leaks in ColumnVectorTest ([#12625](https://github.com/NVIDIA/cudf/pull/12625)) [@jlowe](https://github.com/jlowe) +- Handle when spillable buffers own each other ([#12607](https://github.com/NVIDIA/cudf/pull/12607)) [@madsbk](https://github.com/madsbk) +- Fix incorrect null counts for sliced columns in JCudfSerialization ([#12589](https://github.com/NVIDIA/cudf/pull/12589)) [@jlowe](https://github.com/jlowe) +- lists: Transfer dtypes correctly through list.get ([#12586](https://github.com/NVIDIA/cudf/pull/12586)) [@wence-](https://github.com/wence-) +- timedelta: Don't go via float intermediates for floordiv ([#12585](https://github.com/NVIDIA/cudf/pull/12585)) [@wence-](https://github.com/wence-) +- Fixing BUG, `get_next_chunk()` should use the blocking function `device_read()` ([#12584](https://github.com/NVIDIA/cudf/pull/12584)) [@madsbk](https://github.com/madsbk) +- Make JNI QuoteStyle accessible outside ai.rapids.cudf ([#12572](https://github.com/NVIDIA/cudf/pull/12572)) [@mythrocks](https://github.com/mythrocks) +- `partition_by_hash()`: support index ([#12554](https://github.com/NVIDIA/cudf/pull/12554)) [@madsbk](https://github.com/madsbk) +- Mixed Join benchmark bug due to wrong conditional column ([#12553](https://github.com/NVIDIA/cudf/pull/12553)) [@divyegala](https://github.com/divyegala) +- Update List Lexicographical Comparator ([#12538](https://github.com/NVIDIA/cudf/pull/12538)) [@divyegala](https://github.com/divyegala) +- Dynamically read PTX version ([#12534](https://github.com/NVIDIA/cudf/pull/12534)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- build.sh switch to use `RAPIDS` magic value ([#12525](https://github.com/NVIDIA/cudf/pull/12525)) [@robertmaynard](https://github.com/robertmaynard) +- Loosen runtime arrow pinning ([#12522](https://github.com/NVIDIA/cudf/pull/12522)) [@vyasr](https://github.com/vyasr) +- Enable metadata transfer for complex types in transpose ([#12491](https://github.com/NVIDIA/cudf/pull/12491)) [@galipremsagar](https://github.com/galipremsagar) +- Fix issues with parquet chunked reader ([#12488](https://github.com/NVIDIA/cudf/pull/12488)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix missing metadata transfer in concat for `ListColumn` ([#12487](https://github.com/NVIDIA/cudf/pull/12487)) [@galipremsagar](https://github.com/galipremsagar) +- Rename libcudf substring source files to slice ([#12484](https://github.com/NVIDIA/cudf/pull/12484)) [@davidwendt](https://github.com/davidwendt) +- Fix compile issue with arrow 10 ([#12465](https://github.com/NVIDIA/cudf/pull/12465)) [@ttnghia](https://github.com/ttnghia) +- Fix List offsets bug in mixed type list column in nested JSON reader ([#12447](https://github.com/NVIDIA/cudf/pull/12447)) [@karthikeyann](https://github.com/karthikeyann) +- Fix xfail incompatibilities ([#12423](https://github.com/NVIDIA/cudf/pull/12423)) [@vyasr](https://github.com/vyasr) +- Fix bug in Parquet column index encoding ([#12404](https://github.com/NVIDIA/cudf/pull/12404)) [@etseidl](https://github.com/etseidl) +- When building Arrow shared look for a shared OpenSSL ([#12396](https://github.com/NVIDIA/cudf/pull/12396)) [@robertmaynard](https://github.com/robertmaynard) +- Fix get_json_object to return empty column on empty input ([#12384](https://github.com/NVIDIA/cudf/pull/12384)) [@davidwendt](https://github.com/davidwendt) +- Pin arrow 9 in testing dependencies to prevent conda solve issues ([#12377](https://github.com/NVIDIA/cudf/pull/12377)) [@vyasr](https://github.com/vyasr) +- Fix reductions any/all return value for empty input ([#12374](https://github.com/NVIDIA/cudf/pull/12374)) [@davidwendt](https://github.com/davidwendt) +- Fix debug compile errors in parquet.hpp ([#12372](https://github.com/NVIDIA/cudf/pull/12372)) [@davidwendt](https://github.com/davidwendt) +- Purge non-empty nulls in `cudf::make_lists_column` ([#12370](https://github.com/NVIDIA/cudf/pull/12370)) [@ttnghia](https://github.com/ttnghia) +- Use correct memory resource in io::make_column ([#12364](https://github.com/NVIDIA/cudf/pull/12364)) [@vyasr](https://github.com/vyasr) +- Add code to detect possible malformed page data in parquet files. ([#12360](https://github.com/NVIDIA/cudf/pull/12360)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fail loudly to avoid data corruption with unsupported input in `read_orc` ([#12325](https://github.com/NVIDIA/cudf/pull/12325)) [@vuule](https://github.com/vuule) +- Fix NumericPairIteratorTest for float values ([#12306](https://github.com/NVIDIA/cudf/pull/12306)) [@davidwendt](https://github.com/davidwendt) +- Fixes memory allocation in nested JSON tokenizer ([#12300](https://github.com/NVIDIA/cudf/pull/12300)) [@elstehle](https://github.com/elstehle) +- Reconstruct dtypes correctly for list aggs of struct columns ([#12290](https://github.com/NVIDIA/cudf/pull/12290)) [@wence-](https://github.com/wence-) +- Fix regex \A and \Z to strictly match string begin/end ([#12282](https://github.com/NVIDIA/cudf/pull/12282)) [@davidwendt](https://github.com/davidwendt) +- Fix compile issue in `json_chunked_reader.cpp` ([#12280](https://github.com/NVIDIA/cudf/pull/12280)) [@ttnghia](https://github.com/ttnghia) +- Change reductions any/all to return valid values for empty input ([#12279](https://github.com/NVIDIA/cudf/pull/12279)) [@davidwendt](https://github.com/davidwendt) +- Only exclude join keys that are indices from key columns ([#12271](https://github.com/NVIDIA/cudf/pull/12271)) [@wence-](https://github.com/wence-) +- Fix spill to device limit ([#12252](https://github.com/NVIDIA/cudf/pull/12252)) [@madsbk](https://github.com/madsbk) +- Correct behaviour of sort in `concat` for singleton concatenations ([#12247](https://github.com/NVIDIA/cudf/pull/12247)) [@wence-](https://github.com/wence-) +- Purge non-empty nulls for `superimpose_nulls` and `push_down_nulls` ([#12239](https://github.com/NVIDIA/cudf/pull/12239)) [@ttnghia](https://github.com/ttnghia) +- Patch CUB DeviceSegmentedSort and remove workaround ([#12234](https://github.com/NVIDIA/cudf/pull/12234)) [@davidwendt](https://github.com/davidwendt) +- Fix memory leak in udf_string::assign(&&) function ([#12206](https://github.com/NVIDIA/cudf/pull/12206)) [@davidwendt](https://github.com/davidwendt) +- Workaround thrust-copy-if limit in json get_tree_representation ([#12190](https://github.com/NVIDIA/cudf/pull/12190)) [@davidwendt](https://github.com/davidwendt) +- Fix page size calculation in Parquet writer ([#12182](https://github.com/NVIDIA/cudf/pull/12182)) [@etseidl](https://github.com/etseidl) +- Add cudf::detail::sizes_to_offsets_iterator to allow checking overflow in offsets ([#12180](https://github.com/NVIDIA/cudf/pull/12180)) [@davidwendt](https://github.com/davidwendt) +- Workaround thrust-copy-if limit in wordpiece-tokenizer ([#12168](https://github.com/NVIDIA/cudf/pull/12168)) [@davidwendt](https://github.com/davidwendt) +- Floor division uses integer division for integral arguments ([#12131](https://github.com/NVIDIA/cudf/pull/12131)) [@wence-](https://github.com/wence-) ## 📖 Documentation -- Fix link to NVTX ([#12598](https://github.com/rapidsai/cudf/pull/12598)) [@sameerz](https://github.com/sameerz) -- Include missing groupby functions in documentation ([#12580](https://github.com/rapidsai/cudf/pull/12580)) [@quasiben](https://github.com/quasiben) -- Fix documentation author ([#12527](https://github.com/rapidsai/cudf/pull/12527)) [@bdice](https://github.com/bdice) -- Update libcudf reduction docs for casting output types ([#12526](https://github.com/rapidsai/cudf/pull/12526)) [@davidwendt](https://github.com/davidwendt) -- Add JSON reader page in user guide ([#12499](https://github.com/rapidsai/cudf/pull/12499)) [@GregoryKimball](https://github.com/GregoryKimball) -- Link unsupported iteration API docstrings ([#12482](https://github.com/rapidsai/cudf/pull/12482)) [@galipremsagar](https://github.com/galipremsagar) -- `strings_udf` doc update ([#12469](https://github.com/rapidsai/cudf/pull/12469)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Update cudf_assert docs with correct NDEBUG behavior ([#12464](https://github.com/rapidsai/cudf/pull/12464)) [@robertmaynard](https://github.com/robertmaynard) -- Update pre-commit hooks guide ([#12395](https://github.com/rapidsai/cudf/pull/12395)) [@bdice](https://github.com/bdice) -- Update test docs to not use detail comparison utilities ([#12332](https://github.com/rapidsai/cudf/pull/12332)) [@PointKernel](https://github.com/PointKernel) -- Fix doxygen description for regex_program::compute_working_memory_size ([#12329](https://github.com/rapidsai/cudf/pull/12329)) [@davidwendt](https://github.com/davidwendt) -- Add eval to docs. ([#12322](https://github.com/rapidsai/cudf/pull/12322)) [@vyasr](https://github.com/vyasr) -- Turn on xfail_strict=true ([#12244](https://github.com/rapidsai/cudf/pull/12244)) [@wence-](https://github.com/wence-) -- Update 10 minutes to cuDF ([#12114](https://github.com/rapidsai/cudf/pull/12114)) [@wence-](https://github.com/wence-) +- Fix link to NVTX ([#12598](https://github.com/NVIDIA/cudf/pull/12598)) [@sameerz](https://github.com/sameerz) +- Include missing groupby functions in documentation ([#12580](https://github.com/NVIDIA/cudf/pull/12580)) [@quasiben](https://github.com/quasiben) +- Fix documentation author ([#12527](https://github.com/NVIDIA/cudf/pull/12527)) [@bdice](https://github.com/bdice) +- Update libcudf reduction docs for casting output types ([#12526](https://github.com/NVIDIA/cudf/pull/12526)) [@davidwendt](https://github.com/davidwendt) +- Add JSON reader page in user guide ([#12499](https://github.com/NVIDIA/cudf/pull/12499)) [@GregoryKimball](https://github.com/GregoryKimball) +- Link unsupported iteration API docstrings ([#12482](https://github.com/NVIDIA/cudf/pull/12482)) [@galipremsagar](https://github.com/galipremsagar) +- `strings_udf` doc update ([#12469](https://github.com/NVIDIA/cudf/pull/12469)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Update cudf_assert docs with correct NDEBUG behavior ([#12464](https://github.com/NVIDIA/cudf/pull/12464)) [@robertmaynard](https://github.com/robertmaynard) +- Update pre-commit hooks guide ([#12395](https://github.com/NVIDIA/cudf/pull/12395)) [@bdice](https://github.com/bdice) +- Update test docs to not use detail comparison utilities ([#12332](https://github.com/NVIDIA/cudf/pull/12332)) [@PointKernel](https://github.com/PointKernel) +- Fix doxygen description for regex_program::compute_working_memory_size ([#12329](https://github.com/NVIDIA/cudf/pull/12329)) [@davidwendt](https://github.com/davidwendt) +- Add eval to docs. ([#12322](https://github.com/NVIDIA/cudf/pull/12322)) [@vyasr](https://github.com/vyasr) +- Turn on xfail_strict=true ([#12244](https://github.com/NVIDIA/cudf/pull/12244)) [@wence-](https://github.com/wence-) +- Update 10 minutes to cuDF ([#12114](https://github.com/NVIDIA/cudf/pull/12114)) [@wence-](https://github.com/wence-) ## 🚀 New Features -- Use kvikIO as the default IO backend ([#12574](https://github.com/rapidsai/cudf/pull/12574)) [@vuule](https://github.com/vuule) -- Use `has_nonempty_nulls` instead of `may_contain_non_empty_nulls` in `superimpose_nulls` and `push_down_nulls` ([#12560](https://github.com/rapidsai/cudf/pull/12560)) [@ttnghia](https://github.com/ttnghia) -- Add strings methods removeprefix and removesuffix ([#12557](https://github.com/rapidsai/cudf/pull/12557)) [@davidwendt](https://github.com/davidwendt) -- Add `regex_program` java APIs and unit tests ([#12548](https://github.com/rapidsai/cudf/pull/12548)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Default `cudf::io::read_json` to nested JSON parser ([#12544](https://github.com/rapidsai/cudf/pull/12544)) [@vuule](https://github.com/vuule) -- Make string quoting optional on CSV write ([#12539](https://github.com/rapidsai/cudf/pull/12539)) [@mythrocks](https://github.com/mythrocks) -- Use new nvCOMP API to optimize the compression temp memory size ([#12533](https://github.com/rapidsai/cudf/pull/12533)) [@vuule](https://github.com/vuule) -- Support "values" orient (array of arrays) in Nested JSON reader ([#12498](https://github.com/rapidsai/cudf/pull/12498)) [@karthikeyann](https://github.com/karthikeyann) -- `one_hot_encode` to use experimental row comparators ([#12478](https://github.com/rapidsai/cudf/pull/12478)) [@divyegala](https://github.com/divyegala) -- Support %W and %w format specifiers in cudf::strings::to_timestamps ([#12475](https://github.com/rapidsai/cudf/pull/12475)) [@davidwendt](https://github.com/davidwendt) -- Add JSON Writer ([#12474](https://github.com/rapidsai/cudf/pull/12474)) [@karthikeyann](https://github.com/karthikeyann) -- Refactor `thrust_copy_if` into `cudf::detail::copy_if_safe` ([#12455](https://github.com/rapidsai/cudf/pull/12455)) [@ttnghia](https://github.com/ttnghia) -- Add trailing comma support for nested JSON reader ([#12448](https://github.com/rapidsai/cudf/pull/12448)) [@karthikeyann](https://github.com/karthikeyann) -- Extract `tokenize_json.hpp` detail header from `src/io/json/nested_json.hpp` ([#12432](https://github.com/rapidsai/cudf/pull/12432)) [@ttnghia](https://github.com/ttnghia) -- JNI bindings to write CSV ([#12425](https://github.com/rapidsai/cudf/pull/12425)) [@mythrocks](https://github.com/mythrocks) -- Nested JSON depth benchmark ([#12371](https://github.com/rapidsai/cudf/pull/12371)) [@karthikeyann](https://github.com/karthikeyann) -- Implement `lists::reverse` ([#12336](https://github.com/rapidsai/cudf/pull/12336)) [@ttnghia](https://github.com/ttnghia) -- Use `device_read` in experimental `read_json` ([#12314](https://github.com/rapidsai/cudf/pull/12314)) [@vuule](https://github.com/vuule) -- Implement JNI for `strings::reverse` ([#12283](https://github.com/rapidsai/cudf/pull/12283)) [@ttnghia](https://github.com/ttnghia) -- Null element for parsing error in numeric types in JSON, CSV reader ([#12272](https://github.com/rapidsai/cudf/pull/12272)) [@karthikeyann](https://github.com/karthikeyann) -- Add cudf::strings:like function with multiple patterns ([#12269](https://github.com/rapidsai/cudf/pull/12269)) [@davidwendt](https://github.com/davidwendt) -- Add environment variable to control host memory allocation in `hostdevice_vector` ([#12251](https://github.com/rapidsai/cudf/pull/12251)) [@vuule](https://github.com/vuule) -- Add cudf::strings::reverse function ([#12227](https://github.com/rapidsai/cudf/pull/12227)) [@davidwendt](https://github.com/davidwendt) -- Selectively use dictionary encoding in Parquet writer ([#12211](https://github.com/rapidsai/cudf/pull/12211)) [@etseidl](https://github.com/etseidl) -- Support `replace` in `strings_udf` ([#12207](https://github.com/rapidsai/cudf/pull/12207)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add support to read binary encoded decimals in parquet ([#12205](https://github.com/rapidsai/cudf/pull/12205)) [@PointKernel](https://github.com/PointKernel) -- Support regex EOL where the string ends with a new-line character ([#12181](https://github.com/rapidsai/cudf/pull/12181)) [@davidwendt](https://github.com/davidwendt) -- Updating `stream_compaction/unique` to use new row comparators ([#12159](https://github.com/rapidsai/cudf/pull/12159)) [@divyegala](https://github.com/divyegala) -- Add device buffer datasource ([#12024](https://github.com/rapidsai/cudf/pull/12024)) [@PointKernel](https://github.com/PointKernel) -- Implement groupby apply with JIT ([#11452](https://github.com/rapidsai/cudf/pull/11452)) [@bwyogatama](https://github.com/bwyogatama) +- Use kvikIO as the default IO backend ([#12574](https://github.com/NVIDIA/cudf/pull/12574)) [@vuule](https://github.com/vuule) +- Use `has_nonempty_nulls` instead of `may_contain_non_empty_nulls` in `superimpose_nulls` and `push_down_nulls` ([#12560](https://github.com/NVIDIA/cudf/pull/12560)) [@ttnghia](https://github.com/ttnghia) +- Add strings methods removeprefix and removesuffix ([#12557](https://github.com/NVIDIA/cudf/pull/12557)) [@davidwendt](https://github.com/davidwendt) +- Add `regex_program` java APIs and unit tests ([#12548](https://github.com/NVIDIA/cudf/pull/12548)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Default `cudf::io::read_json` to nested JSON parser ([#12544](https://github.com/NVIDIA/cudf/pull/12544)) [@vuule](https://github.com/vuule) +- Make string quoting optional on CSV write ([#12539](https://github.com/NVIDIA/cudf/pull/12539)) [@mythrocks](https://github.com/mythrocks) +- Use new nvCOMP API to optimize the compression temp memory size ([#12533](https://github.com/NVIDIA/cudf/pull/12533)) [@vuule](https://github.com/vuule) +- Support "values" orient (array of arrays) in Nested JSON reader ([#12498](https://github.com/NVIDIA/cudf/pull/12498)) [@karthikeyann](https://github.com/karthikeyann) +- `one_hot_encode` to use experimental row comparators ([#12478](https://github.com/NVIDIA/cudf/pull/12478)) [@divyegala](https://github.com/divyegala) +- Support %W and %w format specifiers in cudf::strings::to_timestamps ([#12475](https://github.com/NVIDIA/cudf/pull/12475)) [@davidwendt](https://github.com/davidwendt) +- Add JSON Writer ([#12474](https://github.com/NVIDIA/cudf/pull/12474)) [@karthikeyann](https://github.com/karthikeyann) +- Refactor `thrust_copy_if` into `cudf::detail::copy_if_safe` ([#12455](https://github.com/NVIDIA/cudf/pull/12455)) [@ttnghia](https://github.com/ttnghia) +- Add trailing comma support for nested JSON reader ([#12448](https://github.com/NVIDIA/cudf/pull/12448)) [@karthikeyann](https://github.com/karthikeyann) +- Extract `tokenize_json.hpp` detail header from `src/io/json/nested_json.hpp` ([#12432](https://github.com/NVIDIA/cudf/pull/12432)) [@ttnghia](https://github.com/ttnghia) +- JNI bindings to write CSV ([#12425](https://github.com/NVIDIA/cudf/pull/12425)) [@mythrocks](https://github.com/mythrocks) +- Nested JSON depth benchmark ([#12371](https://github.com/NVIDIA/cudf/pull/12371)) [@karthikeyann](https://github.com/karthikeyann) +- Implement `lists::reverse` ([#12336](https://github.com/NVIDIA/cudf/pull/12336)) [@ttnghia](https://github.com/ttnghia) +- Use `device_read` in experimental `read_json` ([#12314](https://github.com/NVIDIA/cudf/pull/12314)) [@vuule](https://github.com/vuule) +- Implement JNI for `strings::reverse` ([#12283](https://github.com/NVIDIA/cudf/pull/12283)) [@ttnghia](https://github.com/ttnghia) +- Null element for parsing error in numeric types in JSON, CSV reader ([#12272](https://github.com/NVIDIA/cudf/pull/12272)) [@karthikeyann](https://github.com/karthikeyann) +- Add cudf::strings:like function with multiple patterns ([#12269](https://github.com/NVIDIA/cudf/pull/12269)) [@davidwendt](https://github.com/davidwendt) +- Add environment variable to control host memory allocation in `hostdevice_vector` ([#12251](https://github.com/NVIDIA/cudf/pull/12251)) [@vuule](https://github.com/vuule) +- Add cudf::strings::reverse function ([#12227](https://github.com/NVIDIA/cudf/pull/12227)) [@davidwendt](https://github.com/davidwendt) +- Selectively use dictionary encoding in Parquet writer ([#12211](https://github.com/NVIDIA/cudf/pull/12211)) [@etseidl](https://github.com/etseidl) +- Support `replace` in `strings_udf` ([#12207](https://github.com/NVIDIA/cudf/pull/12207)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add support to read binary encoded decimals in parquet ([#12205](https://github.com/NVIDIA/cudf/pull/12205)) [@PointKernel](https://github.com/PointKernel) +- Support regex EOL where the string ends with a new-line character ([#12181](https://github.com/NVIDIA/cudf/pull/12181)) [@davidwendt](https://github.com/davidwendt) +- Updating `stream_compaction/unique` to use new row comparators ([#12159](https://github.com/NVIDIA/cudf/pull/12159)) [@divyegala](https://github.com/divyegala) +- Add device buffer datasource ([#12024](https://github.com/NVIDIA/cudf/pull/12024)) [@PointKernel](https://github.com/PointKernel) +- Implement groupby apply with JIT ([#11452](https://github.com/NVIDIA/cudf/pull/11452)) [@bwyogatama](https://github.com/bwyogatama) ## 🛠️ Improvements -- Update shared workflow branches ([#12696](https://github.com/rapidsai/cudf/pull/12696)) [@ajschmidt8](https://github.com/ajschmidt8) -- Pin `dask` and `distributed` for release ([#12695](https://github.com/rapidsai/cudf/pull/12695)) [@galipremsagar](https://github.com/galipremsagar) -- Don't upload `libcudf-example` to Anaconda.org ([#12671](https://github.com/rapidsai/cudf/pull/12671)) [@ajschmidt8](https://github.com/ajschmidt8) -- Pin wheel dependencies to same RAPIDS release ([#12659](https://github.com/rapidsai/cudf/pull/12659)) [@sevagh](https://github.com/sevagh) -- Use CTK 118/cp310 branch of wheel workflows ([#12602](https://github.com/rapidsai/cudf/pull/12602)) [@sevagh](https://github.com/sevagh) -- Change ways to access `ptr` in `Buffer` ([#12587](https://github.com/rapidsai/cudf/pull/12587)) [@galipremsagar](https://github.com/galipremsagar) -- Version a parquet writer xfail ([#12579](https://github.com/rapidsai/cudf/pull/12579)) [@galipremsagar](https://github.com/galipremsagar) -- Remove column names ([#12578](https://github.com/rapidsai/cudf/pull/12578)) [@vuule](https://github.com/vuule) -- Parquet reader optimization to address V100 regression. ([#12577](https://github.com/rapidsai/cudf/pull/12577)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add support for `category` dtypes in CSV reader ([#12571](https://github.com/rapidsai/cudf/pull/12571)) [@galipremsagar](https://github.com/galipremsagar) -- Remove `spill_lock` parameter from `SpillableBuffer.get_ptr()` ([#12564](https://github.com/rapidsai/cudf/pull/12564)) [@madsbk](https://github.com/madsbk) -- Optimize `cudf::make_lists_column` ([#12547](https://github.com/rapidsai/cudf/pull/12547)) [@ttnghia](https://github.com/ttnghia) -- Remove `cudf::strings::repeat_strings_output_sizes` from Java and JNI ([#12546](https://github.com/rapidsai/cudf/pull/12546)) [@ttnghia](https://github.com/ttnghia) -- Test that cuInit is not called when RAPIDS_NO_INITIALIZE is set ([#12545](https://github.com/rapidsai/cudf/pull/12545)) [@wence-](https://github.com/wence-) -- Rework repeat_strings to use sizes-to-offsets utility ([#12543](https://github.com/rapidsai/cudf/pull/12543)) [@davidwendt](https://github.com/davidwendt) -- Replace exclusive_scan with sizes_to_offsets in cudf::lists::sequences ([#12541](https://github.com/rapidsai/cudf/pull/12541)) [@davidwendt](https://github.com/davidwendt) -- Rework nvtext::ngrams_tokenize to use sizes-to-offsets utility ([#12540](https://github.com/rapidsai/cudf/pull/12540)) [@davidwendt](https://github.com/davidwendt) -- Fix binary-ops gtests coded in namespace cudf::test ([#12536](https://github.com/rapidsai/cudf/pull/12536)) [@davidwendt](https://github.com/davidwendt) -- More `[@acquire_spill_lock()` and `as_buffer(..., exposed=False)` ([#12535](https://github.com/rapidsai/cudf/pull/12535)) @madsbk](https://github.com/acquire_spill_lock()` and `as_buffer(..., exposed=False)` ([#12535](https://github.com/rapidsai/cudf/pull/12535)) @madsbk) -- Guard CUDA runtime APIs with error checking ([#12531](https://github.com/rapidsai/cudf/pull/12531)) [@PointKernel](https://github.com/PointKernel) -- Update TODOs from issue 10432. ([#12528](https://github.com/rapidsai/cudf/pull/12528)) [@bdice](https://github.com/bdice) -- Update rapids-cmake definitions version in GitHub Actions style checks. ([#12511](https://github.com/rapidsai/cudf/pull/12511)) [@bdice](https://github.com/bdice) -- Switch `engine=cudf` to the new `JSON` reader ([#12509](https://github.com/rapidsai/cudf/pull/12509)) [@galipremsagar](https://github.com/galipremsagar) -- Fix SUM/MEAN aggregation type support. ([#12503](https://github.com/rapidsai/cudf/pull/12503)) [@bdice](https://github.com/bdice) -- Stop using pandas._testing ([#12492](https://github.com/rapidsai/cudf/pull/12492)) [@vyasr](https://github.com/vyasr) -- Fix ROLLING_TEST gtests coded in namespace cudf::test ([#12490](https://github.com/rapidsai/cudf/pull/12490)) [@davidwendt](https://github.com/davidwendt) -- Fix erroneously skipped ORC ZSTD test ([#12486](https://github.com/rapidsai/cudf/pull/12486)) [@vuule](https://github.com/vuule) -- Rework nvtext::generate_character_ngrams to use make_strings_children ([#12480](https://github.com/rapidsai/cudf/pull/12480)) [@davidwendt](https://github.com/davidwendt) -- Raise warnings as errors in the test suite ([#12468](https://github.com/rapidsai/cudf/pull/12468)) [@vyasr](https://github.com/vyasr) -- Remove `int32` hard-coding in python ([#12467](https://github.com/rapidsai/cudf/pull/12467)) [@galipremsagar](https://github.com/galipremsagar) -- Use cudaMemcpyDefault. ([#12466](https://github.com/rapidsai/cudf/pull/12466)) [@bdice](https://github.com/bdice) -- Update workflows for nightly tests ([#12462](https://github.com/rapidsai/cudf/pull/12462)) [@ajschmidt8](https://github.com/ajschmidt8) -- Build CUDA `11.8` and Python `3.10` Packages ([#12457](https://github.com/rapidsai/cudf/pull/12457)) [@ajschmidt8](https://github.com/ajschmidt8) -- JNI build image default as cuda11.8 ([#12441](https://github.com/rapidsai/cudf/pull/12441)) [@pxLi](https://github.com/pxLi) -- Re-enable `Recently Updated` Check ([#12435](https://github.com/rapidsai/cudf/pull/12435)) [@ajschmidt8](https://github.com/ajschmidt8) -- Rework remaining cudf::strings::from_xyz functions to use make_strings_children ([#12434](https://github.com/rapidsai/cudf/pull/12434)) [@vuule](https://github.com/vuule) -- Build wheels alongside conda CI ([#12427](https://github.com/rapidsai/cudf/pull/12427)) [@sevagh](https://github.com/sevagh) -- Remove arguments for checking exception messages in Python ([#12424](https://github.com/rapidsai/cudf/pull/12424)) [@vyasr](https://github.com/vyasr) -- Clean up cuco usage ([#12421](https://github.com/rapidsai/cudf/pull/12421)) [@PointKernel](https://github.com/PointKernel) -- Fix warnings in remaining modules ([#12406](https://github.com/rapidsai/cudf/pull/12406)) [@vyasr](https://github.com/vyasr) -- Update `ops-bot.yaml` ([#12402](https://github.com/rapidsai/cudf/pull/12402)) [@ajschmidt8](https://github.com/ajschmidt8) -- Rework cudf::strings::integers_to_ipv4 to use make_strings_children utility ([#12401](https://github.com/rapidsai/cudf/pull/12401)) [@davidwendt](https://github.com/davidwendt) -- Use `numpy.empty()` instead of `bytearray` to allocate host memory for spilling ([#12399](https://github.com/rapidsai/cudf/pull/12399)) [@madsbk](https://github.com/madsbk) -- Deprecate chunksize from dask_cudf.read_csv ([#12394](https://github.com/rapidsai/cudf/pull/12394)) [@rjzamora](https://github.com/rjzamora) -- Expose the RMM pool size in JNI ([#12390](https://github.com/rapidsai/cudf/pull/12390)) [@revans2](https://github.com/revans2) -- Fix COPYING_TEST: gtests coded in namespace cudf::test ([#12387](https://github.com/rapidsai/cudf/pull/12387)) [@davidwendt](https://github.com/davidwendt) -- Rework cudf::strings::url_encode to use make_strings_children utility ([#12385](https://github.com/rapidsai/cudf/pull/12385)) [@davidwendt](https://github.com/davidwendt) -- Use make_strings_children in parse_data nested json reader ([#12382](https://github.com/rapidsai/cudf/pull/12382)) [@karthikeyann](https://github.com/karthikeyann) -- Fix warnings in test_datetime.py ([#12381](https://github.com/rapidsai/cudf/pull/12381)) [@vyasr](https://github.com/vyasr) -- Mixed Join Benchmarks ([#12375](https://github.com/rapidsai/cudf/pull/12375)) [@divyegala](https://github.com/divyegala) -- Fix warnings in dataframe.py ([#12369](https://github.com/rapidsai/cudf/pull/12369)) [@vyasr](https://github.com/vyasr) -- Update conda recipes. ([#12368](https://github.com/rapidsai/cudf/pull/12368)) [@bdice](https://github.com/bdice) -- Use gpu-latest-1 runner tag ([#12366](https://github.com/rapidsai/cudf/pull/12366)) [@bdice](https://github.com/bdice) -- Rework cudf::strings::from_booleans to use make_strings_children ([#12365](https://github.com/rapidsai/cudf/pull/12365)) [@vuule](https://github.com/vuule) -- Fix warnings in test modules up to test_dataframe.py ([#12355](https://github.com/rapidsai/cudf/pull/12355)) [@vyasr](https://github.com/vyasr) -- JSON column performance optimization - struct column nulls ([#12354](https://github.com/rapidsai/cudf/pull/12354)) [@karthikeyann](https://github.com/karthikeyann) -- Accelerate stable-segmented-sort with CUB segmented sort ([#12347](https://github.com/rapidsai/cudf/pull/12347)) [@davidwendt](https://github.com/davidwendt) -- Add size check to make_offsets_child_column utility ([#12345](https://github.com/rapidsai/cudf/pull/12345)) [@davidwendt](https://github.com/davidwendt) -- Enable max compression ratio small block optimization for ZSTD ([#12338](https://github.com/rapidsai/cudf/pull/12338)) [@vuule](https://github.com/vuule) -- Fix warnings in test_monotonic.py ([#12334](https://github.com/rapidsai/cudf/pull/12334)) [@vyasr](https://github.com/vyasr) -- Improve JSON column creation performance (list offsets) ([#12330](https://github.com/rapidsai/cudf/pull/12330)) [@karthikeyann](https://github.com/karthikeyann) -- Upgrade to `arrow-10.0.1` ([#12327](https://github.com/rapidsai/cudf/pull/12327)) [@galipremsagar](https://github.com/galipremsagar) -- Fix warnings in test_orc.py ([#12326](https://github.com/rapidsai/cudf/pull/12326)) [@vyasr](https://github.com/vyasr) -- Fix warnings in test_groupby.py ([#12324](https://github.com/rapidsai/cudf/pull/12324)) [@vyasr](https://github.com/vyasr) -- Fix `test_notebooks.sh` ([#12323](https://github.com/rapidsai/cudf/pull/12323)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix transform gtests coded in namespace cudf::test ([#12321](https://github.com/rapidsai/cudf/pull/12321)) [@davidwendt](https://github.com/davidwendt) -- Fix `check_style.sh` script ([#12320](https://github.com/rapidsai/cudf/pull/12320)) [@ajschmidt8](https://github.com/ajschmidt8) -- Rework cudf::strings::from_timestamps to use make_strings_children ([#12317](https://github.com/rapidsai/cudf/pull/12317)) [@davidwendt](https://github.com/davidwendt) -- Fix warnings in test_index.py ([#12313](https://github.com/rapidsai/cudf/pull/12313)) [@vyasr](https://github.com/vyasr) -- Fix warnings in test_multiindex.py ([#12310](https://github.com/rapidsai/cudf/pull/12310)) [@vyasr](https://github.com/vyasr) -- CSV, JSON reader to infer integer column with nulls as int64 instead of float64 ([#12309](https://github.com/rapidsai/cudf/pull/12309)) [@karthikeyann](https://github.com/karthikeyann) -- Fix warnings in test_indexing.py ([#12305](https://github.com/rapidsai/cudf/pull/12305)) [@vyasr](https://github.com/vyasr) -- Fix warnings in test_joining.py ([#12304](https://github.com/rapidsai/cudf/pull/12304)) [@vyasr](https://github.com/vyasr) -- Unpin `dask` and `distributed` for development ([#12302](https://github.com/rapidsai/cudf/pull/12302)) [@galipremsagar](https://github.com/galipremsagar) -- Re-enable `sccache` for Jenkins builds ([#12297](https://github.com/rapidsai/cudf/pull/12297)) [@ajschmidt8](https://github.com/ajschmidt8) -- Define needs for pr-builder workflow. ([#12296](https://github.com/rapidsai/cudf/pull/12296)) [@bdice](https://github.com/bdice) -- Forward merge 22.12 into 23.02 ([#12294](https://github.com/rapidsai/cudf/pull/12294)) [@vyasr](https://github.com/vyasr) -- Fix warnings in test_stats.py ([#12293](https://github.com/rapidsai/cudf/pull/12293)) [@vyasr](https://github.com/vyasr) -- Fix table gtests coded in namespace cudf::test ([#12292](https://github.com/rapidsai/cudf/pull/12292)) [@davidwendt](https://github.com/davidwendt) -- Change cython for regex calls to use cudf::strings::regex_program ([#12289](https://github.com/rapidsai/cudf/pull/12289)) [@davidwendt](https://github.com/davidwendt) -- Improved error reporting when reading multiple JSON files ([#12285](https://github.com/rapidsai/cudf/pull/12285)) [@vuule](https://github.com/vuule) -- Deprecate Frame.sum_of_squares ([#12284](https://github.com/rapidsai/cudf/pull/12284)) [@vyasr](https://github.com/vyasr) -- Remove deprecated code for 23.02 ([#12281](https://github.com/rapidsai/cudf/pull/12281)) [@vyasr](https://github.com/vyasr) -- Clean up handling of max_page_size_bytes in Parquet writer ([#12277](https://github.com/rapidsai/cudf/pull/12277)) [@etseidl](https://github.com/etseidl) -- Fix replace gtests coded in namespace cudf::test ([#12270](https://github.com/rapidsai/cudf/pull/12270)) [@davidwendt](https://github.com/davidwendt) -- Add pandas nullable type support in `Index.to_pandas` ([#12268](https://github.com/rapidsai/cudf/pull/12268)) [@galipremsagar](https://github.com/galipremsagar) -- Rework nvtext::detokenize to use indexalator for row indices ([#12267](https://github.com/rapidsai/cudf/pull/12267)) [@davidwendt](https://github.com/davidwendt) -- Fix reduction gtests coded in namespace cudf::test ([#12257](https://github.com/rapidsai/cudf/pull/12257)) [@davidwendt](https://github.com/davidwendt) -- Remove default parameters from cudf::detail::sort function declarations ([#12254](https://github.com/rapidsai/cudf/pull/12254)) [@davidwendt](https://github.com/davidwendt) -- Add `duplicated` support for `Series`, `DataFrame` and `Index` ([#12246](https://github.com/rapidsai/cudf/pull/12246)) [@galipremsagar](https://github.com/galipremsagar) -- Replace column/table test utilities with macros ([#12242](https://github.com/rapidsai/cudf/pull/12242)) [@PointKernel](https://github.com/PointKernel) -- Rework cudf::strings::pad and zfill to use make_strings_children ([#12238](https://github.com/rapidsai/cudf/pull/12238)) [@davidwendt](https://github.com/davidwendt) -- Fix sort gtests coded in namespace cudf::test ([#12237](https://github.com/rapidsai/cudf/pull/12237)) [@davidwendt](https://github.com/davidwendt) -- Wrapping concat and file writes in `[@acquire_spill_lock()` ([#12232](https://github.com/rapidsai/cudf/pull/12232)) @madsbk](https://github.com/acquire_spill_lock()` ([#12232](https://github.com/rapidsai/cudf/pull/12232)) @madsbk) -- Rename `cudf::structs::detail::superimpose_parent_nulls` APIs ([#12230](https://github.com/rapidsai/cudf/pull/12230)) [@ttnghia](https://github.com/ttnghia) -- Cover parsing to decimal types in `read_json` tests ([#12229](https://github.com/rapidsai/cudf/pull/12229)) [@vuule](https://github.com/vuule) -- Spill Statistics ([#12223](https://github.com/rapidsai/cudf/pull/12223)) [@madsbk](https://github.com/madsbk) -- Use CUDF_JNI_ENABLE_PROFILING to conditionally enable profiling support. ([#12221](https://github.com/rapidsai/cudf/pull/12221)) [@bdice](https://github.com/bdice) -- Clean up of `test_spilling.py` ([#12220](https://github.com/rapidsai/cudf/pull/12220)) [@madsbk](https://github.com/madsbk) -- Simplify repetitive boolean logic ([#12218](https://github.com/rapidsai/cudf/pull/12218)) [@vuule](https://github.com/vuule) -- Add `Series.hasnans` and `Index.hasnans` ([#12214](https://github.com/rapidsai/cudf/pull/12214)) [@galipremsagar](https://github.com/galipremsagar) -- Add cudf::strings:udf::replace function ([#12210](https://github.com/rapidsai/cudf/pull/12210)) [@davidwendt](https://github.com/davidwendt) -- Adds in new java APIs for appending byte arrays to host columnar data ([#12208](https://github.com/rapidsai/cudf/pull/12208)) [@revans2](https://github.com/revans2) -- Remove Python dependencies from Java CI. ([#12193](https://github.com/rapidsai/cudf/pull/12193)) [@bdice](https://github.com/bdice) -- Fix null order in sort-based groupby and improve groupby tests ([#12191](https://github.com/rapidsai/cudf/pull/12191)) [@divyegala](https://github.com/divyegala) -- Move strings children functions from cudf/strings/detail/utilities.cuh to new header ([#12185](https://github.com/rapidsai/cudf/pull/12185)) [@davidwendt](https://github.com/davidwendt) -- Clean up existing JNI scalar to column code ([#12173](https://github.com/rapidsai/cudf/pull/12173)) [@revans2](https://github.com/revans2) -- Remove JIT type names, refactor id_to_type. ([#12158](https://github.com/rapidsai/cudf/pull/12158)) [@bdice](https://github.com/bdice) -- Update JNI version to 23.02.0-SNAPSHOT ([#12129](https://github.com/rapidsai/cudf/pull/12129)) [@pxLi](https://github.com/pxLi) -- Minor refactor of cpp/src/io/parquet/page_data.cu ([#12126](https://github.com/rapidsai/cudf/pull/12126)) [@etseidl](https://github.com/etseidl) -- Add codespell as a linter ([#12097](https://github.com/rapidsai/cudf/pull/12097)) [@benfred](https://github.com/benfred) -- Enable specifying exceptions in error macros ([#12078](https://github.com/rapidsai/cudf/pull/12078)) [@vyasr](https://github.com/vyasr) -- Move `_label_encoding` from Series to Column ([#12040](https://github.com/rapidsai/cudf/pull/12040)) [@shwina](https://github.com/shwina) -- Add GitHub Actions Workflows ([#12002](https://github.com/rapidsai/cudf/pull/12002)) [@ajschmidt8](https://github.com/ajschmidt8) -- Consolidate dask-cudf `groupby_agg` calls in one place ([#10835](https://github.com/rapidsai/cudf/pull/10835)) [@charlesbluca](https://github.com/charlesbluca) +- Update shared workflow branches ([#12696](https://github.com/NVIDIA/cudf/pull/12696)) [@ajschmidt8](https://github.com/ajschmidt8) +- Pin `dask` and `distributed` for release ([#12695](https://github.com/NVIDIA/cudf/pull/12695)) [@galipremsagar](https://github.com/galipremsagar) +- Don't upload `libcudf-example` to Anaconda.org ([#12671](https://github.com/NVIDIA/cudf/pull/12671)) [@ajschmidt8](https://github.com/ajschmidt8) +- Pin wheel dependencies to same RAPIDS release ([#12659](https://github.com/NVIDIA/cudf/pull/12659)) [@sevagh](https://github.com/sevagh) +- Use CTK 118/cp310 branch of wheel workflows ([#12602](https://github.com/NVIDIA/cudf/pull/12602)) [@sevagh](https://github.com/sevagh) +- Change ways to access `ptr` in `Buffer` ([#12587](https://github.com/NVIDIA/cudf/pull/12587)) [@galipremsagar](https://github.com/galipremsagar) +- Version a parquet writer xfail ([#12579](https://github.com/NVIDIA/cudf/pull/12579)) [@galipremsagar](https://github.com/galipremsagar) +- Remove column names ([#12578](https://github.com/NVIDIA/cudf/pull/12578)) [@vuule](https://github.com/vuule) +- Parquet reader optimization to address V100 regression. ([#12577](https://github.com/NVIDIA/cudf/pull/12577)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add support for `category` dtypes in CSV reader ([#12571](https://github.com/NVIDIA/cudf/pull/12571)) [@galipremsagar](https://github.com/galipremsagar) +- Remove `spill_lock` parameter from `SpillableBuffer.get_ptr()` ([#12564](https://github.com/NVIDIA/cudf/pull/12564)) [@madsbk](https://github.com/madsbk) +- Optimize `cudf::make_lists_column` ([#12547](https://github.com/NVIDIA/cudf/pull/12547)) [@ttnghia](https://github.com/ttnghia) +- Remove `cudf::strings::repeat_strings_output_sizes` from Java and JNI ([#12546](https://github.com/NVIDIA/cudf/pull/12546)) [@ttnghia](https://github.com/ttnghia) +- Test that cuInit is not called when RAPIDS_NO_INITIALIZE is set ([#12545](https://github.com/NVIDIA/cudf/pull/12545)) [@wence-](https://github.com/wence-) +- Rework repeat_strings to use sizes-to-offsets utility ([#12543](https://github.com/NVIDIA/cudf/pull/12543)) [@davidwendt](https://github.com/davidwendt) +- Replace exclusive_scan with sizes_to_offsets in cudf::lists::sequences ([#12541](https://github.com/NVIDIA/cudf/pull/12541)) [@davidwendt](https://github.com/davidwendt) +- Rework nvtext::ngrams_tokenize to use sizes-to-offsets utility ([#12540](https://github.com/NVIDIA/cudf/pull/12540)) [@davidwendt](https://github.com/davidwendt) +- Fix binary-ops gtests coded in namespace cudf::test ([#12536](https://github.com/NVIDIA/cudf/pull/12536)) [@davidwendt](https://github.com/davidwendt) +- More `[@acquire_spill_lock()` and `as_buffer(..., exposed=False)` ([#12535](https://github.com/NVIDIA/cudf/pull/12535)) @madsbk](https://github.com/acquire_spill_lock()` and `as_buffer(..., exposed=False)` ([#12535](https://github.com/NVIDIA/cudf/pull/12535)) @madsbk) +- Guard CUDA runtime APIs with error checking ([#12531](https://github.com/NVIDIA/cudf/pull/12531)) [@PointKernel](https://github.com/PointKernel) +- Update TODOs from issue 10432. ([#12528](https://github.com/NVIDIA/cudf/pull/12528)) [@bdice](https://github.com/bdice) +- Update rapids-cmake definitions version in GitHub Actions style checks. ([#12511](https://github.com/NVIDIA/cudf/pull/12511)) [@bdice](https://github.com/bdice) +- Switch `engine=cudf` to the new `JSON` reader ([#12509](https://github.com/NVIDIA/cudf/pull/12509)) [@galipremsagar](https://github.com/galipremsagar) +- Fix SUM/MEAN aggregation type support. ([#12503](https://github.com/NVIDIA/cudf/pull/12503)) [@bdice](https://github.com/bdice) +- Stop using pandas._testing ([#12492](https://github.com/NVIDIA/cudf/pull/12492)) [@vyasr](https://github.com/vyasr) +- Fix ROLLING_TEST gtests coded in namespace cudf::test ([#12490](https://github.com/NVIDIA/cudf/pull/12490)) [@davidwendt](https://github.com/davidwendt) +- Fix erroneously skipped ORC ZSTD test ([#12486](https://github.com/NVIDIA/cudf/pull/12486)) [@vuule](https://github.com/vuule) +- Rework nvtext::generate_character_ngrams to use make_strings_children ([#12480](https://github.com/NVIDIA/cudf/pull/12480)) [@davidwendt](https://github.com/davidwendt) +- Raise warnings as errors in the test suite ([#12468](https://github.com/NVIDIA/cudf/pull/12468)) [@vyasr](https://github.com/vyasr) +- Remove `int32` hard-coding in python ([#12467](https://github.com/NVIDIA/cudf/pull/12467)) [@galipremsagar](https://github.com/galipremsagar) +- Use cudaMemcpyDefault. ([#12466](https://github.com/NVIDIA/cudf/pull/12466)) [@bdice](https://github.com/bdice) +- Update workflows for nightly tests ([#12462](https://github.com/NVIDIA/cudf/pull/12462)) [@ajschmidt8](https://github.com/ajschmidt8) +- Build CUDA `11.8` and Python `3.10` Packages ([#12457](https://github.com/NVIDIA/cudf/pull/12457)) [@ajschmidt8](https://github.com/ajschmidt8) +- JNI build image default as cuda11.8 ([#12441](https://github.com/NVIDIA/cudf/pull/12441)) [@pxLi](https://github.com/pxLi) +- Re-enable `Recently Updated` Check ([#12435](https://github.com/NVIDIA/cudf/pull/12435)) [@ajschmidt8](https://github.com/ajschmidt8) +- Rework remaining cudf::strings::from_xyz functions to use make_strings_children ([#12434](https://github.com/NVIDIA/cudf/pull/12434)) [@vuule](https://github.com/vuule) +- Build wheels alongside conda CI ([#12427](https://github.com/NVIDIA/cudf/pull/12427)) [@sevagh](https://github.com/sevagh) +- Remove arguments for checking exception messages in Python ([#12424](https://github.com/NVIDIA/cudf/pull/12424)) [@vyasr](https://github.com/vyasr) +- Clean up cuco usage ([#12421](https://github.com/NVIDIA/cudf/pull/12421)) [@PointKernel](https://github.com/PointKernel) +- Fix warnings in remaining modules ([#12406](https://github.com/NVIDIA/cudf/pull/12406)) [@vyasr](https://github.com/vyasr) +- Update `ops-bot.yaml` ([#12402](https://github.com/NVIDIA/cudf/pull/12402)) [@ajschmidt8](https://github.com/ajschmidt8) +- Rework cudf::strings::integers_to_ipv4 to use make_strings_children utility ([#12401](https://github.com/NVIDIA/cudf/pull/12401)) [@davidwendt](https://github.com/davidwendt) +- Use `numpy.empty()` instead of `bytearray` to allocate host memory for spilling ([#12399](https://github.com/NVIDIA/cudf/pull/12399)) [@madsbk](https://github.com/madsbk) +- Deprecate chunksize from dask_cudf.read_csv ([#12394](https://github.com/NVIDIA/cudf/pull/12394)) [@rjzamora](https://github.com/rjzamora) +- Expose the RMM pool size in JNI ([#12390](https://github.com/NVIDIA/cudf/pull/12390)) [@revans2](https://github.com/revans2) +- Fix COPYING_TEST: gtests coded in namespace cudf::test ([#12387](https://github.com/NVIDIA/cudf/pull/12387)) [@davidwendt](https://github.com/davidwendt) +- Rework cudf::strings::url_encode to use make_strings_children utility ([#12385](https://github.com/NVIDIA/cudf/pull/12385)) [@davidwendt](https://github.com/davidwendt) +- Use make_strings_children in parse_data nested json reader ([#12382](https://github.com/NVIDIA/cudf/pull/12382)) [@karthikeyann](https://github.com/karthikeyann) +- Fix warnings in test_datetime.py ([#12381](https://github.com/NVIDIA/cudf/pull/12381)) [@vyasr](https://github.com/vyasr) +- Mixed Join Benchmarks ([#12375](https://github.com/NVIDIA/cudf/pull/12375)) [@divyegala](https://github.com/divyegala) +- Fix warnings in dataframe.py ([#12369](https://github.com/NVIDIA/cudf/pull/12369)) [@vyasr](https://github.com/vyasr) +- Update conda recipes. ([#12368](https://github.com/NVIDIA/cudf/pull/12368)) [@bdice](https://github.com/bdice) +- Use gpu-latest-1 runner tag ([#12366](https://github.com/NVIDIA/cudf/pull/12366)) [@bdice](https://github.com/bdice) +- Rework cudf::strings::from_booleans to use make_strings_children ([#12365](https://github.com/NVIDIA/cudf/pull/12365)) [@vuule](https://github.com/vuule) +- Fix warnings in test modules up to test_dataframe.py ([#12355](https://github.com/NVIDIA/cudf/pull/12355)) [@vyasr](https://github.com/vyasr) +- JSON column performance optimization - struct column nulls ([#12354](https://github.com/NVIDIA/cudf/pull/12354)) [@karthikeyann](https://github.com/karthikeyann) +- Accelerate stable-segmented-sort with CUB segmented sort ([#12347](https://github.com/NVIDIA/cudf/pull/12347)) [@davidwendt](https://github.com/davidwendt) +- Add size check to make_offsets_child_column utility ([#12345](https://github.com/NVIDIA/cudf/pull/12345)) [@davidwendt](https://github.com/davidwendt) +- Enable max compression ratio small block optimization for ZSTD ([#12338](https://github.com/NVIDIA/cudf/pull/12338)) [@vuule](https://github.com/vuule) +- Fix warnings in test_monotonic.py ([#12334](https://github.com/NVIDIA/cudf/pull/12334)) [@vyasr](https://github.com/vyasr) +- Improve JSON column creation performance (list offsets) ([#12330](https://github.com/NVIDIA/cudf/pull/12330)) [@karthikeyann](https://github.com/karthikeyann) +- Upgrade to `arrow-10.0.1` ([#12327](https://github.com/NVIDIA/cudf/pull/12327)) [@galipremsagar](https://github.com/galipremsagar) +- Fix warnings in test_orc.py ([#12326](https://github.com/NVIDIA/cudf/pull/12326)) [@vyasr](https://github.com/vyasr) +- Fix warnings in test_groupby.py ([#12324](https://github.com/NVIDIA/cudf/pull/12324)) [@vyasr](https://github.com/vyasr) +- Fix `test_notebooks.sh` ([#12323](https://github.com/NVIDIA/cudf/pull/12323)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix transform gtests coded in namespace cudf::test ([#12321](https://github.com/NVIDIA/cudf/pull/12321)) [@davidwendt](https://github.com/davidwendt) +- Fix `check_style.sh` script ([#12320](https://github.com/NVIDIA/cudf/pull/12320)) [@ajschmidt8](https://github.com/ajschmidt8) +- Rework cudf::strings::from_timestamps to use make_strings_children ([#12317](https://github.com/NVIDIA/cudf/pull/12317)) [@davidwendt](https://github.com/davidwendt) +- Fix warnings in test_index.py ([#12313](https://github.com/NVIDIA/cudf/pull/12313)) [@vyasr](https://github.com/vyasr) +- Fix warnings in test_multiindex.py ([#12310](https://github.com/NVIDIA/cudf/pull/12310)) [@vyasr](https://github.com/vyasr) +- CSV, JSON reader to infer integer column with nulls as int64 instead of float64 ([#12309](https://github.com/NVIDIA/cudf/pull/12309)) [@karthikeyann](https://github.com/karthikeyann) +- Fix warnings in test_indexing.py ([#12305](https://github.com/NVIDIA/cudf/pull/12305)) [@vyasr](https://github.com/vyasr) +- Fix warnings in test_joining.py ([#12304](https://github.com/NVIDIA/cudf/pull/12304)) [@vyasr](https://github.com/vyasr) +- Unpin `dask` and `distributed` for development ([#12302](https://github.com/NVIDIA/cudf/pull/12302)) [@galipremsagar](https://github.com/galipremsagar) +- Re-enable `sccache` for Jenkins builds ([#12297](https://github.com/NVIDIA/cudf/pull/12297)) [@ajschmidt8](https://github.com/ajschmidt8) +- Define needs for pr-builder workflow. ([#12296](https://github.com/NVIDIA/cudf/pull/12296)) [@bdice](https://github.com/bdice) +- Forward merge 22.12 into 23.02 ([#12294](https://github.com/NVIDIA/cudf/pull/12294)) [@vyasr](https://github.com/vyasr) +- Fix warnings in test_stats.py ([#12293](https://github.com/NVIDIA/cudf/pull/12293)) [@vyasr](https://github.com/vyasr) +- Fix table gtests coded in namespace cudf::test ([#12292](https://github.com/NVIDIA/cudf/pull/12292)) [@davidwendt](https://github.com/davidwendt) +- Change cython for regex calls to use cudf::strings::regex_program ([#12289](https://github.com/NVIDIA/cudf/pull/12289)) [@davidwendt](https://github.com/davidwendt) +- Improved error reporting when reading multiple JSON files ([#12285](https://github.com/NVIDIA/cudf/pull/12285)) [@vuule](https://github.com/vuule) +- Deprecate Frame.sum_of_squares ([#12284](https://github.com/NVIDIA/cudf/pull/12284)) [@vyasr](https://github.com/vyasr) +- Remove deprecated code for 23.02 ([#12281](https://github.com/NVIDIA/cudf/pull/12281)) [@vyasr](https://github.com/vyasr) +- Clean up handling of max_page_size_bytes in Parquet writer ([#12277](https://github.com/NVIDIA/cudf/pull/12277)) [@etseidl](https://github.com/etseidl) +- Fix replace gtests coded in namespace cudf::test ([#12270](https://github.com/NVIDIA/cudf/pull/12270)) [@davidwendt](https://github.com/davidwendt) +- Add pandas nullable type support in `Index.to_pandas` ([#12268](https://github.com/NVIDIA/cudf/pull/12268)) [@galipremsagar](https://github.com/galipremsagar) +- Rework nvtext::detokenize to use indexalator for row indices ([#12267](https://github.com/NVIDIA/cudf/pull/12267)) [@davidwendt](https://github.com/davidwendt) +- Fix reduction gtests coded in namespace cudf::test ([#12257](https://github.com/NVIDIA/cudf/pull/12257)) [@davidwendt](https://github.com/davidwendt) +- Remove default parameters from cudf::detail::sort function declarations ([#12254](https://github.com/NVIDIA/cudf/pull/12254)) [@davidwendt](https://github.com/davidwendt) +- Add `duplicated` support for `Series`, `DataFrame` and `Index` ([#12246](https://github.com/NVIDIA/cudf/pull/12246)) [@galipremsagar](https://github.com/galipremsagar) +- Replace column/table test utilities with macros ([#12242](https://github.com/NVIDIA/cudf/pull/12242)) [@PointKernel](https://github.com/PointKernel) +- Rework cudf::strings::pad and zfill to use make_strings_children ([#12238](https://github.com/NVIDIA/cudf/pull/12238)) [@davidwendt](https://github.com/davidwendt) +- Fix sort gtests coded in namespace cudf::test ([#12237](https://github.com/NVIDIA/cudf/pull/12237)) [@davidwendt](https://github.com/davidwendt) +- Wrapping concat and file writes in `[@acquire_spill_lock()` ([#12232](https://github.com/NVIDIA/cudf/pull/12232)) @madsbk](https://github.com/acquire_spill_lock()` ([#12232](https://github.com/NVIDIA/cudf/pull/12232)) @madsbk) +- Rename `cudf::structs::detail::superimpose_parent_nulls` APIs ([#12230](https://github.com/NVIDIA/cudf/pull/12230)) [@ttnghia](https://github.com/ttnghia) +- Cover parsing to decimal types in `read_json` tests ([#12229](https://github.com/NVIDIA/cudf/pull/12229)) [@vuule](https://github.com/vuule) +- Spill Statistics ([#12223](https://github.com/NVIDIA/cudf/pull/12223)) [@madsbk](https://github.com/madsbk) +- Use CUDF_JNI_ENABLE_PROFILING to conditionally enable profiling support. ([#12221](https://github.com/NVIDIA/cudf/pull/12221)) [@bdice](https://github.com/bdice) +- Clean up of `test_spilling.py` ([#12220](https://github.com/NVIDIA/cudf/pull/12220)) [@madsbk](https://github.com/madsbk) +- Simplify repetitive boolean logic ([#12218](https://github.com/NVIDIA/cudf/pull/12218)) [@vuule](https://github.com/vuule) +- Add `Series.hasnans` and `Index.hasnans` ([#12214](https://github.com/NVIDIA/cudf/pull/12214)) [@galipremsagar](https://github.com/galipremsagar) +- Add cudf::strings:udf::replace function ([#12210](https://github.com/NVIDIA/cudf/pull/12210)) [@davidwendt](https://github.com/davidwendt) +- Adds in new java APIs for appending byte arrays to host columnar data ([#12208](https://github.com/NVIDIA/cudf/pull/12208)) [@revans2](https://github.com/revans2) +- Remove Python dependencies from Java CI. ([#12193](https://github.com/NVIDIA/cudf/pull/12193)) [@bdice](https://github.com/bdice) +- Fix null order in sort-based groupby and improve groupby tests ([#12191](https://github.com/NVIDIA/cudf/pull/12191)) [@divyegala](https://github.com/divyegala) +- Move strings children functions from cudf/strings/detail/utilities.cuh to new header ([#12185](https://github.com/NVIDIA/cudf/pull/12185)) [@davidwendt](https://github.com/davidwendt) +- Clean up existing JNI scalar to column code ([#12173](https://github.com/NVIDIA/cudf/pull/12173)) [@revans2](https://github.com/revans2) +- Remove JIT type names, refactor id_to_type. ([#12158](https://github.com/NVIDIA/cudf/pull/12158)) [@bdice](https://github.com/bdice) +- Update JNI version to 23.02.0-SNAPSHOT ([#12129](https://github.com/NVIDIA/cudf/pull/12129)) [@pxLi](https://github.com/pxLi) +- Minor refactor of cpp/src/io/parquet/page_data.cu ([#12126](https://github.com/NVIDIA/cudf/pull/12126)) [@etseidl](https://github.com/etseidl) +- Add codespell as a linter ([#12097](https://github.com/NVIDIA/cudf/pull/12097)) [@benfred](https://github.com/benfred) +- Enable specifying exceptions in error macros ([#12078](https://github.com/NVIDIA/cudf/pull/12078)) [@vyasr](https://github.com/vyasr) +- Move `_label_encoding` from Series to Column ([#12040](https://github.com/NVIDIA/cudf/pull/12040)) [@shwina](https://github.com/shwina) +- Add GitHub Actions Workflows ([#12002](https://github.com/NVIDIA/cudf/pull/12002)) [@ajschmidt8](https://github.com/ajschmidt8) +- Consolidate dask-cudf `groupby_agg` calls in one place ([#10835](https://github.com/NVIDIA/cudf/pull/10835)) [@charlesbluca](https://github.com/charlesbluca) # cuDF 22.12.00 (8 Dec 2022) ## 🚨 Breaking Changes -- Add JNI for `substring` without 'end' parameter. ([#12113](https://github.com/rapidsai/cudf/pull/12113)) [@firestarman](https://github.com/firestarman) -- Refactor `purge_nonempty_nulls` ([#12111](https://github.com/rapidsai/cudf/pull/12111)) [@ttnghia](https://github.com/ttnghia) -- Create an `int8` column in `read_csv` when all elements are missing ([#12110](https://github.com/rapidsai/cudf/pull/12110)) [@vuule](https://github.com/vuule) -- Throw an error when libcudf is built without cuFile and `LIBCUDF_CUFILE_POLICY` is set to `"ALWAYS"` ([#12080](https://github.com/rapidsai/cudf/pull/12080)) [@vuule](https://github.com/vuule) -- Fix type promotion edge cases in numerical binops ([#12074](https://github.com/rapidsai/cudf/pull/12074)) [@wence-](https://github.com/wence-) -- Reduce/Remove reliance on `**kwargs` and `*args` in `IO` readers & writers ([#12025](https://github.com/rapidsai/cudf/pull/12025)) [@galipremsagar](https://github.com/galipremsagar) -- Rollback of `DeviceBufferLike` ([#12009](https://github.com/rapidsai/cudf/pull/12009)) [@madsbk](https://github.com/madsbk) -- Remove unused `managed_allocator` ([#12005](https://github.com/rapidsai/cudf/pull/12005)) [@vyasr](https://github.com/vyasr) -- Pass column names to `write_csv` instead of `table_metadata` pointer ([#11972](https://github.com/rapidsai/cudf/pull/11972)) [@vuule](https://github.com/vuule) -- Accept const refs instead of const unique_ptr refs in reduce and scan APIs. ([#11960](https://github.com/rapidsai/cudf/pull/11960)) [@vyasr](https://github.com/vyasr) -- Default to equal NaNs in make_merge_sets_aggregation. ([#11952](https://github.com/rapidsai/cudf/pull/11952)) [@bdice](https://github.com/bdice) -- Remove validation that requires introspection ([#11938](https://github.com/rapidsai/cudf/pull/11938)) [@vyasr](https://github.com/vyasr) -- Trim quotes for non-string values in nested json parsing ([#11898](https://github.com/rapidsai/cudf/pull/11898)) [@karthikeyann](https://github.com/karthikeyann) -- Add tests ensuring that cudf's default stream is always used ([#11875](https://github.com/rapidsai/cudf/pull/11875)) [@vyasr](https://github.com/vyasr) -- Support nested types as groupby keys in libcudf ([#11792](https://github.com/rapidsai/cudf/pull/11792)) [@PointKernel](https://github.com/PointKernel) -- Default to equal NaNs in make_collect_set_aggregation. ([#11621](https://github.com/rapidsai/cudf/pull/11621)) [@bdice](https://github.com/bdice) -- Removing int8 column option from parquet byte_array writing ([#11539](https://github.com/rapidsai/cudf/pull/11539)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- part1: Simplify BaseIndex to an abstract class ([#10389](https://github.com/rapidsai/cudf/pull/10389)) [@skirui-source](https://github.com/skirui-source) +- Add JNI for `substring` without 'end' parameter. ([#12113](https://github.com/NVIDIA/cudf/pull/12113)) [@firestarman](https://github.com/firestarman) +- Refactor `purge_nonempty_nulls` ([#12111](https://github.com/NVIDIA/cudf/pull/12111)) [@ttnghia](https://github.com/ttnghia) +- Create an `int8` column in `read_csv` when all elements are missing ([#12110](https://github.com/NVIDIA/cudf/pull/12110)) [@vuule](https://github.com/vuule) +- Throw an error when libcudf is built without cuFile and `LIBCUDF_CUFILE_POLICY` is set to `"ALWAYS"` ([#12080](https://github.com/NVIDIA/cudf/pull/12080)) [@vuule](https://github.com/vuule) +- Fix type promotion edge cases in numerical binops ([#12074](https://github.com/NVIDIA/cudf/pull/12074)) [@wence-](https://github.com/wence-) +- Reduce/Remove reliance on `**kwargs` and `*args` in `IO` readers & writers ([#12025](https://github.com/NVIDIA/cudf/pull/12025)) [@galipremsagar](https://github.com/galipremsagar) +- Rollback of `DeviceBufferLike` ([#12009](https://github.com/NVIDIA/cudf/pull/12009)) [@madsbk](https://github.com/madsbk) +- Remove unused `managed_allocator` ([#12005](https://github.com/NVIDIA/cudf/pull/12005)) [@vyasr](https://github.com/vyasr) +- Pass column names to `write_csv` instead of `table_metadata` pointer ([#11972](https://github.com/NVIDIA/cudf/pull/11972)) [@vuule](https://github.com/vuule) +- Accept const refs instead of const unique_ptr refs in reduce and scan APIs. ([#11960](https://github.com/NVIDIA/cudf/pull/11960)) [@vyasr](https://github.com/vyasr) +- Default to equal NaNs in make_merge_sets_aggregation. ([#11952](https://github.com/NVIDIA/cudf/pull/11952)) [@bdice](https://github.com/bdice) +- Remove validation that requires introspection ([#11938](https://github.com/NVIDIA/cudf/pull/11938)) [@vyasr](https://github.com/vyasr) +- Trim quotes for non-string values in nested json parsing ([#11898](https://github.com/NVIDIA/cudf/pull/11898)) [@karthikeyann](https://github.com/karthikeyann) +- Add tests ensuring that cudf's default stream is always used ([#11875](https://github.com/NVIDIA/cudf/pull/11875)) [@vyasr](https://github.com/vyasr) +- Support nested types as groupby keys in libcudf ([#11792](https://github.com/NVIDIA/cudf/pull/11792)) [@PointKernel](https://github.com/PointKernel) +- Default to equal NaNs in make_collect_set_aggregation. ([#11621](https://github.com/NVIDIA/cudf/pull/11621)) [@bdice](https://github.com/bdice) +- Removing int8 column option from parquet byte_array writing ([#11539](https://github.com/NVIDIA/cudf/pull/11539)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- part1: Simplify BaseIndex to an abstract class ([#10389](https://github.com/NVIDIA/cudf/pull/10389)) [@skirui-source](https://github.com/skirui-source) ## 🐛 Bug Fixes -- Fix include line for IO Cython modules ([#12250](https://github.com/rapidsai/cudf/pull/12250)) [@vyasr](https://github.com/vyasr) -- Make dask pinning looser ([#12231](https://github.com/rapidsai/cudf/pull/12231)) [@vyasr](https://github.com/vyasr) -- Workaround for CUB segmented-sort bug with boolean keys ([#12217](https://github.com/rapidsai/cudf/pull/12217)) [@davidwendt](https://github.com/davidwendt) -- Fix `from_dict` backend dispatch to match upstream `dask` ([#12203](https://github.com/rapidsai/cudf/pull/12203)) [@galipremsagar](https://github.com/galipremsagar) -- Merge branch-22.10 into branch-22.12 ([#12198](https://github.com/rapidsai/cudf/pull/12198)) [@davidwendt](https://github.com/davidwendt) -- Fix compression in ORC writer ([#12194](https://github.com/rapidsai/cudf/pull/12194)) [@vuule](https://github.com/vuule) -- Don't use CMake 3.25.0 as it has a show stopping FindCUDAToolkit bug ([#12188](https://github.com/rapidsai/cudf/pull/12188)) [@robertmaynard](https://github.com/robertmaynard) -- Fix data corruption when reading ORC files with empty stripes ([#12160](https://github.com/rapidsai/cudf/pull/12160)) [@vuule](https://github.com/vuule) -- Fix decimal binary operations ([#12142](https://github.com/rapidsai/cudf/pull/12142)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure dlpack include is provided to cudf interop lib ([#12139](https://github.com/rapidsai/cudf/pull/12139)) [@robertmaynard](https://github.com/robertmaynard) -- Safely allocate `udf_string` pointers in `strings_udf` ([#12138](https://github.com/rapidsai/cudf/pull/12138)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix/disable jitify lto ([#12122](https://github.com/rapidsai/cudf/pull/12122)) [@robertmaynard](https://github.com/robertmaynard) -- Fix conditional_full_join benchmark ([#12121](https://github.com/rapidsai/cudf/pull/12121)) [@GregoryKimball](https://github.com/GregoryKimball) -- Fix regex working-memory-size refactor error ([#12119](https://github.com/rapidsai/cudf/pull/12119)) [@davidwendt](https://github.com/davidwendt) -- Add in negative size checks for columns ([#12118](https://github.com/rapidsai/cudf/pull/12118)) [@revans2](https://github.com/revans2) -- Add JNI for `substring` without 'end' parameter. ([#12113](https://github.com/rapidsai/cudf/pull/12113)) [@firestarman](https://github.com/firestarman) -- Fix reading of CSV files with blank second row ([#12098](https://github.com/rapidsai/cudf/pull/12098)) [@vuule](https://github.com/vuule) -- Fix an error in IO with `GzipFile` type ([#12085](https://github.com/rapidsai/cudf/pull/12085)) [@galipremsagar](https://github.com/galipremsagar) -- Workaround groupby aggregate thrust::copy_if overflow ([#12079](https://github.com/rapidsai/cudf/pull/12079)) [@davidwendt](https://github.com/davidwendt) -- Fix alignment of compressed blocks in ORC writer ([#12077](https://github.com/rapidsai/cudf/pull/12077)) [@vuule](https://github.com/vuule) -- Fix singleton-range `__setitem__` edge case ([#12075](https://github.com/rapidsai/cudf/pull/12075)) [@wence-](https://github.com/wence-) -- Fix type promotion edge cases in numerical binops ([#12074](https://github.com/rapidsai/cudf/pull/12074)) [@wence-](https://github.com/wence-) -- Force using old fmt in nvbench. ([#12067](https://github.com/rapidsai/cudf/pull/12067)) [@vyasr](https://github.com/vyasr) -- Fixes List offset bug in Nested JSON reader ([#12060](https://github.com/rapidsai/cudf/pull/12060)) [@karthikeyann](https://github.com/karthikeyann) -- Allow falling back to `shim_60.ptx` by default in `strings_udf` ([#12056](https://github.com/rapidsai/cudf/pull/12056)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Force black exclusions for pre-commit. ([#12036](https://github.com/rapidsai/cudf/pull/12036)) [@bdice](https://github.com/bdice) -- Add `memory_usage` & `items` implementation for `Struct` column & dtype ([#12033](https://github.com/rapidsai/cudf/pull/12033)) [@galipremsagar](https://github.com/galipremsagar) -- Reduce/Remove reliance on `**kwargs` and `*args` in `IO` readers & writers ([#12025](https://github.com/rapidsai/cudf/pull/12025)) [@galipremsagar](https://github.com/galipremsagar) -- Fixes bug in csv_reader_options construction in cython ([#12021](https://github.com/rapidsai/cudf/pull/12021)) [@karthikeyann](https://github.com/karthikeyann) -- Fix issues when both `usecols` and `names` options are used in `read_csv` ([#12018](https://github.com/rapidsai/cudf/pull/12018)) [@vuule](https://github.com/vuule) -- Port thrust's pinned_allocator to cudf, since Thrust 1.17 removes the type ([#12004](https://github.com/rapidsai/cudf/pull/12004)) [@robertmaynard](https://github.com/robertmaynard) -- Revert "Replace most of preprocessor usage in nvcomp adapter with `constexpr`" ([#11999](https://github.com/rapidsai/cudf/pull/11999)) [@vuule](https://github.com/vuule) -- Fix bug where `df.loc` resulting in single row could give wrong index ([#11998](https://github.com/rapidsai/cudf/pull/11998)) [@eriknw](https://github.com/eriknw) -- Switch to DISABLE_DEPRECATION_WARNINGS to match other RAPIDS projects ([#11989](https://github.com/rapidsai/cudf/pull/11989)) [@robertmaynard](https://github.com/robertmaynard) -- Fix maximum page size estimate in Parquet writer ([#11962](https://github.com/rapidsai/cudf/pull/11962)) [@vuule](https://github.com/vuule) -- Fix local offset handling in bgzip reader ([#11918](https://github.com/rapidsai/cudf/pull/11918)) [@upsj](https://github.com/upsj) -- Fix an issue reading struct-of-list types in Parquet. ([#11910](https://github.com/rapidsai/cudf/pull/11910)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix memcheck error in TypeInference.Timestamp gtest ([#11905](https://github.com/rapidsai/cudf/pull/11905)) [@davidwendt](https://github.com/davidwendt) -- Fix type casting in Series.__setitem__ ([#11904](https://github.com/rapidsai/cudf/pull/11904)) [@wence-](https://github.com/wence-) -- Fix memcheck error in get_dremel_data ([#11903](https://github.com/rapidsai/cudf/pull/11903)) [@davidwendt](https://github.com/davidwendt) -- Fixes Unsupported column type error due to empty list columns in Nested JSON reader ([#11897](https://github.com/rapidsai/cudf/pull/11897)) [@karthikeyann](https://github.com/karthikeyann) -- Fix segmented-sort to ignore indices outside the offsets ([#11888](https://github.com/rapidsai/cudf/pull/11888)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf::stable_sorted_order for NaN and -NaN in FLOAT64 columns ([#11874](https://github.com/rapidsai/cudf/pull/11874)) [@davidwendt](https://github.com/davidwendt) -- Fix writing of Parquet files with many fragments ([#11869](https://github.com/rapidsai/cudf/pull/11869)) [@etseidl](https://github.com/etseidl) -- Fix RangeIndex unary operators. ([#11868](https://github.com/rapidsai/cudf/pull/11868)) [@vyasr](https://github.com/vyasr) -- JNI Avoid NPE for reading host binary data ([#11865](https://github.com/rapidsai/cudf/pull/11865)) [@revans2](https://github.com/revans2) -- Fix decimal benchmark input data generation ([#11863](https://github.com/rapidsai/cudf/pull/11863)) [@karthikeyann](https://github.com/karthikeyann) -- Fix pre-commit copyright check ([#11860](https://github.com/rapidsai/cudf/pull/11860)) [@galipremsagar](https://github.com/galipremsagar) -- Fix Parquet support for seconds and milliseconds duration types ([#11854](https://github.com/rapidsai/cudf/pull/11854)) [@vuule](https://github.com/vuule) -- Ensure better compiler cache results between cudf cal-ver branches ([#11835](https://github.com/rapidsai/cudf/pull/11835)) [@robertmaynard](https://github.com/robertmaynard) -- Fix make_column_from_scalar for all-null strings column ([#11807](https://github.com/rapidsai/cudf/pull/11807)) [@davidwendt](https://github.com/davidwendt) -- Tell jitify_preprocess where to search for libnvrtc ([#11787](https://github.com/rapidsai/cudf/pull/11787)) [@robertmaynard](https://github.com/robertmaynard) -- add V2 page header support to parquet reader ([#11778](https://github.com/rapidsai/cudf/pull/11778)) [@etseidl](https://github.com/etseidl) -- Parquet reader: bug fix for a num_rows/skip_rows corner case, w/optimization for nested preprocessing ([#11752](https://github.com/rapidsai/cudf/pull/11752)) [@nvdbaranec](https://github.com/nvdbaranec) -- Determine if Arrow has S3 support at runtime in unit test. ([#11560](https://github.com/rapidsai/cudf/pull/11560)) [@bdice](https://github.com/bdice) +- Fix include line for IO Cython modules ([#12250](https://github.com/NVIDIA/cudf/pull/12250)) [@vyasr](https://github.com/vyasr) +- Make dask pinning looser ([#12231](https://github.com/NVIDIA/cudf/pull/12231)) [@vyasr](https://github.com/vyasr) +- Workaround for CUB segmented-sort bug with boolean keys ([#12217](https://github.com/NVIDIA/cudf/pull/12217)) [@davidwendt](https://github.com/davidwendt) +- Fix `from_dict` backend dispatch to match upstream `dask` ([#12203](https://github.com/NVIDIA/cudf/pull/12203)) [@galipremsagar](https://github.com/galipremsagar) +- Merge branch-22.10 into branch-22.12 ([#12198](https://github.com/NVIDIA/cudf/pull/12198)) [@davidwendt](https://github.com/davidwendt) +- Fix compression in ORC writer ([#12194](https://github.com/NVIDIA/cudf/pull/12194)) [@vuule](https://github.com/vuule) +- Don't use CMake 3.25.0 as it has a show stopping FindCUDAToolkit bug ([#12188](https://github.com/NVIDIA/cudf/pull/12188)) [@robertmaynard](https://github.com/robertmaynard) +- Fix data corruption when reading ORC files with empty stripes ([#12160](https://github.com/NVIDIA/cudf/pull/12160)) [@vuule](https://github.com/vuule) +- Fix decimal binary operations ([#12142](https://github.com/NVIDIA/cudf/pull/12142)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure dlpack include is provided to cudf interop lib ([#12139](https://github.com/NVIDIA/cudf/pull/12139)) [@robertmaynard](https://github.com/robertmaynard) +- Safely allocate `udf_string` pointers in `strings_udf` ([#12138](https://github.com/NVIDIA/cudf/pull/12138)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix/disable jitify lto ([#12122](https://github.com/NVIDIA/cudf/pull/12122)) [@robertmaynard](https://github.com/robertmaynard) +- Fix conditional_full_join benchmark ([#12121](https://github.com/NVIDIA/cudf/pull/12121)) [@GregoryKimball](https://github.com/GregoryKimball) +- Fix regex working-memory-size refactor error ([#12119](https://github.com/NVIDIA/cudf/pull/12119)) [@davidwendt](https://github.com/davidwendt) +- Add in negative size checks for columns ([#12118](https://github.com/NVIDIA/cudf/pull/12118)) [@revans2](https://github.com/revans2) +- Add JNI for `substring` without 'end' parameter. ([#12113](https://github.com/NVIDIA/cudf/pull/12113)) [@firestarman](https://github.com/firestarman) +- Fix reading of CSV files with blank second row ([#12098](https://github.com/NVIDIA/cudf/pull/12098)) [@vuule](https://github.com/vuule) +- Fix an error in IO with `GzipFile` type ([#12085](https://github.com/NVIDIA/cudf/pull/12085)) [@galipremsagar](https://github.com/galipremsagar) +- Workaround groupby aggregate thrust::copy_if overflow ([#12079](https://github.com/NVIDIA/cudf/pull/12079)) [@davidwendt](https://github.com/davidwendt) +- Fix alignment of compressed blocks in ORC writer ([#12077](https://github.com/NVIDIA/cudf/pull/12077)) [@vuule](https://github.com/vuule) +- Fix singleton-range `__setitem__` edge case ([#12075](https://github.com/NVIDIA/cudf/pull/12075)) [@wence-](https://github.com/wence-) +- Fix type promotion edge cases in numerical binops ([#12074](https://github.com/NVIDIA/cudf/pull/12074)) [@wence-](https://github.com/wence-) +- Force using old fmt in nvbench. ([#12067](https://github.com/NVIDIA/cudf/pull/12067)) [@vyasr](https://github.com/vyasr) +- Fixes List offset bug in Nested JSON reader ([#12060](https://github.com/NVIDIA/cudf/pull/12060)) [@karthikeyann](https://github.com/karthikeyann) +- Allow falling back to `shim_60.ptx` by default in `strings_udf` ([#12056](https://github.com/NVIDIA/cudf/pull/12056)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Force black exclusions for pre-commit. ([#12036](https://github.com/NVIDIA/cudf/pull/12036)) [@bdice](https://github.com/bdice) +- Add `memory_usage` & `items` implementation for `Struct` column & dtype ([#12033](https://github.com/NVIDIA/cudf/pull/12033)) [@galipremsagar](https://github.com/galipremsagar) +- Reduce/Remove reliance on `**kwargs` and `*args` in `IO` readers & writers ([#12025](https://github.com/NVIDIA/cudf/pull/12025)) [@galipremsagar](https://github.com/galipremsagar) +- Fixes bug in csv_reader_options construction in cython ([#12021](https://github.com/NVIDIA/cudf/pull/12021)) [@karthikeyann](https://github.com/karthikeyann) +- Fix issues when both `usecols` and `names` options are used in `read_csv` ([#12018](https://github.com/NVIDIA/cudf/pull/12018)) [@vuule](https://github.com/vuule) +- Port thrust's pinned_allocator to cudf, since Thrust 1.17 removes the type ([#12004](https://github.com/NVIDIA/cudf/pull/12004)) [@robertmaynard](https://github.com/robertmaynard) +- Revert "Replace most of preprocessor usage in nvcomp adapter with `constexpr`" ([#11999](https://github.com/NVIDIA/cudf/pull/11999)) [@vuule](https://github.com/vuule) +- Fix bug where `df.loc` resulting in single row could give wrong index ([#11998](https://github.com/NVIDIA/cudf/pull/11998)) [@eriknw](https://github.com/eriknw) +- Switch to DISABLE_DEPRECATION_WARNINGS to match other RAPIDS projects ([#11989](https://github.com/NVIDIA/cudf/pull/11989)) [@robertmaynard](https://github.com/robertmaynard) +- Fix maximum page size estimate in Parquet writer ([#11962](https://github.com/NVIDIA/cudf/pull/11962)) [@vuule](https://github.com/vuule) +- Fix local offset handling in bgzip reader ([#11918](https://github.com/NVIDIA/cudf/pull/11918)) [@upsj](https://github.com/upsj) +- Fix an issue reading struct-of-list types in Parquet. ([#11910](https://github.com/NVIDIA/cudf/pull/11910)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix memcheck error in TypeInference.Timestamp gtest ([#11905](https://github.com/NVIDIA/cudf/pull/11905)) [@davidwendt](https://github.com/davidwendt) +- Fix type casting in Series.__setitem__ ([#11904](https://github.com/NVIDIA/cudf/pull/11904)) [@wence-](https://github.com/wence-) +- Fix memcheck error in get_dremel_data ([#11903](https://github.com/NVIDIA/cudf/pull/11903)) [@davidwendt](https://github.com/davidwendt) +- Fixes Unsupported column type error due to empty list columns in Nested JSON reader ([#11897](https://github.com/NVIDIA/cudf/pull/11897)) [@karthikeyann](https://github.com/karthikeyann) +- Fix segmented-sort to ignore indices outside the offsets ([#11888](https://github.com/NVIDIA/cudf/pull/11888)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf::stable_sorted_order for NaN and -NaN in FLOAT64 columns ([#11874](https://github.com/NVIDIA/cudf/pull/11874)) [@davidwendt](https://github.com/davidwendt) +- Fix writing of Parquet files with many fragments ([#11869](https://github.com/NVIDIA/cudf/pull/11869)) [@etseidl](https://github.com/etseidl) +- Fix RangeIndex unary operators. ([#11868](https://github.com/NVIDIA/cudf/pull/11868)) [@vyasr](https://github.com/vyasr) +- JNI Avoid NPE for reading host binary data ([#11865](https://github.com/NVIDIA/cudf/pull/11865)) [@revans2](https://github.com/revans2) +- Fix decimal benchmark input data generation ([#11863](https://github.com/NVIDIA/cudf/pull/11863)) [@karthikeyann](https://github.com/karthikeyann) +- Fix pre-commit copyright check ([#11860](https://github.com/NVIDIA/cudf/pull/11860)) [@galipremsagar](https://github.com/galipremsagar) +- Fix Parquet support for seconds and milliseconds duration types ([#11854](https://github.com/NVIDIA/cudf/pull/11854)) [@vuule](https://github.com/vuule) +- Ensure better compiler cache results between cudf cal-ver branches ([#11835](https://github.com/NVIDIA/cudf/pull/11835)) [@robertmaynard](https://github.com/robertmaynard) +- Fix make_column_from_scalar for all-null strings column ([#11807](https://github.com/NVIDIA/cudf/pull/11807)) [@davidwendt](https://github.com/davidwendt) +- Tell jitify_preprocess where to search for libnvrtc ([#11787](https://github.com/NVIDIA/cudf/pull/11787)) [@robertmaynard](https://github.com/robertmaynard) +- add V2 page header support to parquet reader ([#11778](https://github.com/NVIDIA/cudf/pull/11778)) [@etseidl](https://github.com/etseidl) +- Parquet reader: bug fix for a num_rows/skip_rows corner case, w/optimization for nested preprocessing ([#11752](https://github.com/NVIDIA/cudf/pull/11752)) [@nvdbaranec](https://github.com/nvdbaranec) +- Determine if Arrow has S3 support at runtime in unit test. ([#11560](https://github.com/NVIDIA/cudf/pull/11560)) [@bdice](https://github.com/bdice) ## 📖 Documentation -- Use rapidsai CODE_OF_CONDUCT.md ([#12166](https://github.com/rapidsai/cudf/pull/12166)) [@bdice](https://github.com/bdice) -- Add symlinks to notebooks. ([#12128](https://github.com/rapidsai/cudf/pull/12128)) [@bdice](https://github.com/bdice) -- Add `truncate` API to python doc pages ([#12109](https://github.com/rapidsai/cudf/pull/12109)) [@galipremsagar](https://github.com/galipremsagar) -- Update Numba docs links. ([#12107](https://github.com/rapidsai/cudf/pull/12107)) [@bdice](https://github.com/bdice) -- Remove "Multi-GPU with Dask-cuDF" notebook. ([#12095](https://github.com/rapidsai/cudf/pull/12095)) [@bdice](https://github.com/bdice) -- Fix link to c++ developer guide from `CONTRIBUTING.md` ([#12084](https://github.com/rapidsai/cudf/pull/12084)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add pivot_table and crosstab to docs. ([#12014](https://github.com/rapidsai/cudf/pull/12014)) [@bdice](https://github.com/bdice) -- Fix doxygen text for cudf::dictionary::encode ([#11991](https://github.com/rapidsai/cudf/pull/11991)) [@davidwendt](https://github.com/davidwendt) -- Replace default_stream_value with get_default_stream in docs. ([#11985](https://github.com/rapidsai/cudf/pull/11985)) [@vyasr](https://github.com/vyasr) -- Add dtype docs pages and docstrings for `cudf` specific dtypes ([#11974](https://github.com/rapidsai/cudf/pull/11974)) [@galipremsagar](https://github.com/galipremsagar) -- Update Unit Testing in libcudf guidelines to code tests outside the cudf::test namespace ([#11959](https://github.com/rapidsai/cudf/pull/11959)) [@davidwendt](https://github.com/davidwendt) -- Rename libcudf++ to libcudf. ([#11953](https://github.com/rapidsai/cudf/pull/11953)) [@bdice](https://github.com/bdice) -- Fix documentation referring to removed as_gpu_matrix method. ([#11937](https://github.com/rapidsai/cudf/pull/11937)) [@bdice](https://github.com/bdice) -- Remove "experimental" warning for struct columns in ORC reader and writer ([#11880](https://github.com/rapidsai/cudf/pull/11880)) [@vuule](https://github.com/vuule) -- Initial draft of policies and guidelines for libcudf usage. ([#11853](https://github.com/rapidsai/cudf/pull/11853)) [@vyasr](https://github.com/vyasr) -- Add clear indication of non-GPU accelerated parameters in read_json docstring ([#11825](https://github.com/rapidsai/cudf/pull/11825)) [@GregoryKimball](https://github.com/GregoryKimball) -- Add developer docs for writing tests ([#11199](https://github.com/rapidsai/cudf/pull/11199)) [@vyasr](https://github.com/vyasr) +- Use rapidsai CODE_OF_CONDUCT.md ([#12166](https://github.com/NVIDIA/cudf/pull/12166)) [@bdice](https://github.com/bdice) +- Add symlinks to notebooks. ([#12128](https://github.com/NVIDIA/cudf/pull/12128)) [@bdice](https://github.com/bdice) +- Add `truncate` API to python doc pages ([#12109](https://github.com/NVIDIA/cudf/pull/12109)) [@galipremsagar](https://github.com/galipremsagar) +- Update Numba docs links. ([#12107](https://github.com/NVIDIA/cudf/pull/12107)) [@bdice](https://github.com/bdice) +- Remove "Multi-GPU with Dask-cuDF" notebook. ([#12095](https://github.com/NVIDIA/cudf/pull/12095)) [@bdice](https://github.com/bdice) +- Fix link to c++ developer guide from `CONTRIBUTING.md` ([#12084](https://github.com/NVIDIA/cudf/pull/12084)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add pivot_table and crosstab to docs. ([#12014](https://github.com/NVIDIA/cudf/pull/12014)) [@bdice](https://github.com/bdice) +- Fix doxygen text for cudf::dictionary::encode ([#11991](https://github.com/NVIDIA/cudf/pull/11991)) [@davidwendt](https://github.com/davidwendt) +- Replace default_stream_value with get_default_stream in docs. ([#11985](https://github.com/NVIDIA/cudf/pull/11985)) [@vyasr](https://github.com/vyasr) +- Add dtype docs pages and docstrings for `cudf` specific dtypes ([#11974](https://github.com/NVIDIA/cudf/pull/11974)) [@galipremsagar](https://github.com/galipremsagar) +- Update Unit Testing in libcudf guidelines to code tests outside the cudf::test namespace ([#11959](https://github.com/NVIDIA/cudf/pull/11959)) [@davidwendt](https://github.com/davidwendt) +- Rename libcudf++ to libcudf. ([#11953](https://github.com/NVIDIA/cudf/pull/11953)) [@bdice](https://github.com/bdice) +- Fix documentation referring to removed as_gpu_matrix method. ([#11937](https://github.com/NVIDIA/cudf/pull/11937)) [@bdice](https://github.com/bdice) +- Remove "experimental" warning for struct columns in ORC reader and writer ([#11880](https://github.com/NVIDIA/cudf/pull/11880)) [@vuule](https://github.com/vuule) +- Initial draft of policies and guidelines for libcudf usage. ([#11853](https://github.com/NVIDIA/cudf/pull/11853)) [@vyasr](https://github.com/vyasr) +- Add clear indication of non-GPU accelerated parameters in read_json docstring ([#11825](https://github.com/NVIDIA/cudf/pull/11825)) [@GregoryKimball](https://github.com/GregoryKimball) +- Add developer docs for writing tests ([#11199](https://github.com/NVIDIA/cudf/pull/11199)) [@vyasr](https://github.com/vyasr) ## 🚀 New Features -- Adds an EventHandler to Java MemoryBuffer to be invoked on close ([#12125](https://github.com/rapidsai/cudf/pull/12125)) [@abellina](https://github.com/abellina) -- Support `+` in `strings_udf` ([#12117](https://github.com/rapidsai/cudf/pull/12117)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `upper` and `lower` in `strings_udf` ([#12099](https://github.com/rapidsai/cudf/pull/12099)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add wheel builds ([#12096](https://github.com/rapidsai/cudf/pull/12096)) [@vyasr](https://github.com/vyasr) -- Allow setting malloc heap size in string udfs ([#12094](https://github.com/rapidsai/cudf/pull/12094)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support `strip`, `lstrip`, and `rstrip` in `strings_udf` ([#12091](https://github.com/rapidsai/cudf/pull/12091)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Mark nvcomp zstd compression stable ([#12059](https://github.com/rapidsai/cudf/pull/12059)) [@jbrennan333](https://github.com/jbrennan333) -- Add debug-only onAllocated/onDeallocated to RmmEventHandler ([#12054](https://github.com/rapidsai/cudf/pull/12054)) [@abellina](https://github.com/abellina) -- Enable building against the libarrow contained in pyarrow ([#12034](https://github.com/rapidsai/cudf/pull/12034)) [@vyasr](https://github.com/vyasr) -- Add strings `like` jni and native method ([#12032](https://github.com/rapidsai/cudf/pull/12032)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) -- Cleanup common parsing code in JSON, CSV reader ([#12022](https://github.com/rapidsai/cudf/pull/12022)) [@karthikeyann](https://github.com/karthikeyann) -- byte_range support for JSON Lines format ([#12017](https://github.com/rapidsai/cudf/pull/12017)) [@karthikeyann](https://github.com/karthikeyann) -- Minor cleanup of root CMakeLists.txt for better organization ([#11988](https://github.com/rapidsai/cudf/pull/11988)) [@robertmaynard](https://github.com/robertmaynard) -- Add inplace arithmetic operators to `MaskedType` ([#11987](https://github.com/rapidsai/cudf/pull/11987)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Implement JNI for chunked Parquet reader ([#11961](https://github.com/rapidsai/cudf/pull/11961)) [@ttnghia](https://github.com/ttnghia) -- Add method argument to DataFrame.quantile ([#11957](https://github.com/rapidsai/cudf/pull/11957)) [@rjzamora](https://github.com/rjzamora) -- Add gpu memory watermark apis to JNI ([#11950](https://github.com/rapidsai/cudf/pull/11950)) [@abellina](https://github.com/abellina) -- Adds retryCount to RmmEventHandler.onAllocFailure ([#11940](https://github.com/rapidsai/cudf/pull/11940)) [@abellina](https://github.com/abellina) -- Enable returning string data from UDFs used through `apply` ([#11933](https://github.com/rapidsai/cudf/pull/11933)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Switch over to rapids-cmake patches for thrust ([#11921](https://github.com/rapidsai/cudf/pull/11921)) [@robertmaynard](https://github.com/robertmaynard) -- Add strings udf C++ classes and functions for phase II ([#11912](https://github.com/rapidsai/cudf/pull/11912)) [@davidwendt](https://github.com/davidwendt) -- Trim quotes for non-string values in nested json parsing ([#11898](https://github.com/rapidsai/cudf/pull/11898)) [@karthikeyann](https://github.com/karthikeyann) -- Enable CEC for `strings_udf` ([#11884](https://github.com/rapidsai/cudf/pull/11884)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- ArrowIPCTableWriter writes en empty batch in the case of an empty table. ([#11883](https://github.com/rapidsai/cudf/pull/11883)) [@firestarman](https://github.com/firestarman) -- Implement chunked Parquet reader ([#11867](https://github.com/rapidsai/cudf/pull/11867)) [@ttnghia](https://github.com/ttnghia) -- Add `read_orc_metadata` to libcudf ([#11815](https://github.com/rapidsai/cudf/pull/11815)) [@vuule](https://github.com/vuule) -- Support nested types as groupby keys in libcudf ([#11792](https://github.com/rapidsai/cudf/pull/11792)) [@PointKernel](https://github.com/PointKernel) -- Adding feature Truncate to DataFrame and Series ([#11435](https://github.com/rapidsai/cudf/pull/11435)) [@VamsiTallam95](https://github.com/VamsiTallam95) +- Adds an EventHandler to Java MemoryBuffer to be invoked on close ([#12125](https://github.com/NVIDIA/cudf/pull/12125)) [@abellina](https://github.com/abellina) +- Support `+` in `strings_udf` ([#12117](https://github.com/NVIDIA/cudf/pull/12117)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `upper` and `lower` in `strings_udf` ([#12099](https://github.com/NVIDIA/cudf/pull/12099)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add wheel builds ([#12096](https://github.com/NVIDIA/cudf/pull/12096)) [@vyasr](https://github.com/vyasr) +- Allow setting malloc heap size in string udfs ([#12094](https://github.com/NVIDIA/cudf/pull/12094)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support `strip`, `lstrip`, and `rstrip` in `strings_udf` ([#12091](https://github.com/NVIDIA/cudf/pull/12091)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Mark nvcomp zstd compression stable ([#12059](https://github.com/NVIDIA/cudf/pull/12059)) [@jbrennan333](https://github.com/jbrennan333) +- Add debug-only onAllocated/onDeallocated to RmmEventHandler ([#12054](https://github.com/NVIDIA/cudf/pull/12054)) [@abellina](https://github.com/abellina) +- Enable building against the libarrow contained in pyarrow ([#12034](https://github.com/NVIDIA/cudf/pull/12034)) [@vyasr](https://github.com/vyasr) +- Add strings `like` jni and native method ([#12032](https://github.com/NVIDIA/cudf/pull/12032)) [@cindyyuanjiang](https://github.com/cindyyuanjiang) +- Cleanup common parsing code in JSON, CSV reader ([#12022](https://github.com/NVIDIA/cudf/pull/12022)) [@karthikeyann](https://github.com/karthikeyann) +- byte_range support for JSON Lines format ([#12017](https://github.com/NVIDIA/cudf/pull/12017)) [@karthikeyann](https://github.com/karthikeyann) +- Minor cleanup of root CMakeLists.txt for better organization ([#11988](https://github.com/NVIDIA/cudf/pull/11988)) [@robertmaynard](https://github.com/robertmaynard) +- Add inplace arithmetic operators to `MaskedType` ([#11987](https://github.com/NVIDIA/cudf/pull/11987)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Implement JNI for chunked Parquet reader ([#11961](https://github.com/NVIDIA/cudf/pull/11961)) [@ttnghia](https://github.com/ttnghia) +- Add method argument to DataFrame.quantile ([#11957](https://github.com/NVIDIA/cudf/pull/11957)) [@rjzamora](https://github.com/rjzamora) +- Add gpu memory watermark apis to JNI ([#11950](https://github.com/NVIDIA/cudf/pull/11950)) [@abellina](https://github.com/abellina) +- Adds retryCount to RmmEventHandler.onAllocFailure ([#11940](https://github.com/NVIDIA/cudf/pull/11940)) [@abellina](https://github.com/abellina) +- Enable returning string data from UDFs used through `apply` ([#11933](https://github.com/NVIDIA/cudf/pull/11933)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Switch over to rapids-cmake patches for thrust ([#11921](https://github.com/NVIDIA/cudf/pull/11921)) [@robertmaynard](https://github.com/robertmaynard) +- Add strings udf C++ classes and functions for phase II ([#11912](https://github.com/NVIDIA/cudf/pull/11912)) [@davidwendt](https://github.com/davidwendt) +- Trim quotes for non-string values in nested json parsing ([#11898](https://github.com/NVIDIA/cudf/pull/11898)) [@karthikeyann](https://github.com/karthikeyann) +- Enable CEC for `strings_udf` ([#11884](https://github.com/NVIDIA/cudf/pull/11884)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- ArrowIPCTableWriter writes en empty batch in the case of an empty table. ([#11883](https://github.com/NVIDIA/cudf/pull/11883)) [@firestarman](https://github.com/firestarman) +- Implement chunked Parquet reader ([#11867](https://github.com/NVIDIA/cudf/pull/11867)) [@ttnghia](https://github.com/ttnghia) +- Add `read_orc_metadata` to libcudf ([#11815](https://github.com/NVIDIA/cudf/pull/11815)) [@vuule](https://github.com/vuule) +- Support nested types as groupby keys in libcudf ([#11792](https://github.com/NVIDIA/cudf/pull/11792)) [@PointKernel](https://github.com/PointKernel) +- Adding feature Truncate to DataFrame and Series ([#11435](https://github.com/NVIDIA/cudf/pull/11435)) [@VamsiTallam95](https://github.com/VamsiTallam95) ## 🛠️ Improvements -- Reduce number of tests marked `spilling` ([#12197](https://github.com/rapidsai/cudf/pull/12197)) [@madsbk](https://github.com/madsbk) -- Pin `dask` and `distributed` for release ([#12165](https://github.com/rapidsai/cudf/pull/12165)) [@galipremsagar](https://github.com/galipremsagar) -- Don't rely on GNU find in headers_test.sh ([#12164](https://github.com/rapidsai/cudf/pull/12164)) [@wence-](https://github.com/wence-) -- Update cp.clip call ([#12148](https://github.com/rapidsai/cudf/pull/12148)) [@quasiben](https://github.com/quasiben) -- Enable automatic column projection in groupby().agg ([#12124](https://github.com/rapidsai/cudf/pull/12124)) [@rjzamora](https://github.com/rjzamora) -- Refactor `purge_nonempty_nulls` ([#12111](https://github.com/rapidsai/cudf/pull/12111)) [@ttnghia](https://github.com/ttnghia) -- Create an `int8` column in `read_csv` when all elements are missing ([#12110](https://github.com/rapidsai/cudf/pull/12110)) [@vuule](https://github.com/vuule) -- Spilling to host memory ([#12106](https://github.com/rapidsai/cudf/pull/12106)) [@madsbk](https://github.com/madsbk) -- First pass of `pd.read_orc` changes in tests ([#12103](https://github.com/rapidsai/cudf/pull/12103)) [@galipremsagar](https://github.com/galipremsagar) -- Expose engine argument in dask_cudf.read_json ([#12101](https://github.com/rapidsai/cudf/pull/12101)) [@rjzamora](https://github.com/rjzamora) -- Remove CUDA 10 compatibility code. ([#12088](https://github.com/rapidsai/cudf/pull/12088)) [@bdice](https://github.com/bdice) -- Move and update `dask` nigthly install in CI ([#12082](https://github.com/rapidsai/cudf/pull/12082)) [@galipremsagar](https://github.com/galipremsagar) -- Throw an error when libcudf is built without cuFile and `LIBCUDF_CUFILE_POLICY` is set to `"ALWAYS"` ([#12080](https://github.com/rapidsai/cudf/pull/12080)) [@vuule](https://github.com/vuule) -- Remove macros that inspect the contents of exceptions ([#12076](https://github.com/rapidsai/cudf/pull/12076)) [@vyasr](https://github.com/vyasr) -- Fix ingest_raw_data performance issue in Nested JSON reader due to RVO ([#12070](https://github.com/rapidsai/cudf/pull/12070)) [@karthikeyann](https://github.com/karthikeyann) -- Remove overflow error during decimal binops ([#12063](https://github.com/rapidsai/cudf/pull/12063)) [@galipremsagar](https://github.com/galipremsagar) -- Change cudf::detail::tdigest to cudf::tdigest::detail ([#12050](https://github.com/rapidsai/cudf/pull/12050)) [@davidwendt](https://github.com/davidwendt) -- Fix quantile gtests coded in namespace cudf::test ([#12049](https://github.com/rapidsai/cudf/pull/12049)) [@davidwendt](https://github.com/davidwendt) -- Add support for `DataFrame.from_dict`\`to_dict` and `Series.to_dict` ([#12048](https://github.com/rapidsai/cudf/pull/12048)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor Parquet reader ([#12046](https://github.com/rapidsai/cudf/pull/12046)) [@ttnghia](https://github.com/ttnghia) -- Forward merge 22.10 into 22.12 ([#12045](https://github.com/rapidsai/cudf/pull/12045)) [@vyasr](https://github.com/vyasr) -- Standardize newlines at ends of files. ([#12042](https://github.com/rapidsai/cudf/pull/12042)) [@bdice](https://github.com/bdice) -- Trim trailing whitespace from all files. ([#12041](https://github.com/rapidsai/cudf/pull/12041)) [@bdice](https://github.com/bdice) -- Use nosync policy in gather and scatter implementations. ([#12038](https://github.com/rapidsai/cudf/pull/12038)) [@bdice](https://github.com/bdice) -- Remove smart quotes from all docstrings. ([#12035](https://github.com/rapidsai/cudf/pull/12035)) [@bdice](https://github.com/bdice) -- Update cuda-python dependency to 11.7.1 ([#12030](https://github.com/rapidsai/cudf/pull/12030)) [@galipremsagar](https://github.com/galipremsagar) -- Add cython-lint to pre-commit checks. ([#12020](https://github.com/rapidsai/cudf/pull/12020)) [@bdice](https://github.com/bdice) -- Use pragma once ([#12019](https://github.com/rapidsai/cudf/pull/12019)) [@bdice](https://github.com/bdice) -- New GHA to add issues/prs to project board ([#12016](https://github.com/rapidsai/cudf/pull/12016)) [@jarmak-nv](https://github.com/jarmak-nv) -- Add DataFrame.pivot_table. ([#12015](https://github.com/rapidsai/cudf/pull/12015)) [@bdice](https://github.com/bdice) -- Rollback of `DeviceBufferLike` ([#12009](https://github.com/rapidsai/cudf/pull/12009)) [@madsbk](https://github.com/madsbk) -- Remove default parameters for nvtext::detail functions ([#12007](https://github.com/rapidsai/cudf/pull/12007)) [@davidwendt](https://github.com/davidwendt) -- Remove default parameters for cudf::dictionary::detail functions ([#12006](https://github.com/rapidsai/cudf/pull/12006)) [@davidwendt](https://github.com/davidwendt) -- Remove unused `managed_allocator` ([#12005](https://github.com/rapidsai/cudf/pull/12005)) [@vyasr](https://github.com/vyasr) -- Remove default parameters for cudf::strings::detail functions ([#12003](https://github.com/rapidsai/cudf/pull/12003)) [@davidwendt](https://github.com/davidwendt) -- Remove unnecessary code from dask-cudf _Frame ([#12001](https://github.com/rapidsai/cudf/pull/12001)) [@rjzamora](https://github.com/rjzamora) -- Ignore python docs build artifacts ([#12000](https://github.com/rapidsai/cudf/pull/12000)) [@galipremsagar](https://github.com/galipremsagar) -- Use rapids-cmake for google benchmark. ([#11997](https://github.com/rapidsai/cudf/pull/11997)) [@vyasr](https://github.com/vyasr) -- Leverage rapids_cython for more automated RPATH handling ([#11996](https://github.com/rapidsai/cudf/pull/11996)) [@vyasr](https://github.com/vyasr) -- Remove stale labeler ([#11995](https://github.com/rapidsai/cudf/pull/11995)) [@raydouglass](https://github.com/raydouglass) -- Move protobuf compilation to CMake ([#11986](https://github.com/rapidsai/cudf/pull/11986)) [@vyasr](https://github.com/vyasr) -- Replace most of preprocessor usage in nvcomp adapter with `constexpr` ([#11980](https://github.com/rapidsai/cudf/pull/11980)) [@vuule](https://github.com/vuule) -- Add missing noexcepts to column_in_metadata methods ([#11973](https://github.com/rapidsai/cudf/pull/11973)) [@vyasr](https://github.com/vyasr) -- Pass column names to `write_csv` instead of `table_metadata` pointer ([#11972](https://github.com/rapidsai/cudf/pull/11972)) [@vuule](https://github.com/vuule) -- Accelerate libcudf segmented sort with CUB segmented sort ([#11969](https://github.com/rapidsai/cudf/pull/11969)) [@davidwendt](https://github.com/davidwendt) -- Feature/remove default streams ([#11967](https://github.com/rapidsai/cudf/pull/11967)) [@vyasr](https://github.com/vyasr) -- Add pool memory resource to libcudf basic example ([#11966](https://github.com/rapidsai/cudf/pull/11966)) [@davidwendt](https://github.com/davidwendt) -- Fix some libcudf calls to cudf::detail::gather ([#11963](https://github.com/rapidsai/cudf/pull/11963)) [@davidwendt](https://github.com/davidwendt) -- Accept const refs instead of const unique_ptr refs in reduce and scan APIs. ([#11960](https://github.com/rapidsai/cudf/pull/11960)) [@vyasr](https://github.com/vyasr) -- Add deprecation warning for set_allocator. ([#11958](https://github.com/rapidsai/cudf/pull/11958)) [@vyasr](https://github.com/vyasr) -- Fix lists and structs gtests coded in namespace cudf::test ([#11956](https://github.com/rapidsai/cudf/pull/11956)) [@davidwendt](https://github.com/davidwendt) -- Add full page indexes to Parquet writer benchmarks ([#11955](https://github.com/rapidsai/cudf/pull/11955)) [@etseidl](https://github.com/etseidl) -- Use gather-based strings factory in cudf::strings::strip ([#11954](https://github.com/rapidsai/cudf/pull/11954)) [@davidwendt](https://github.com/davidwendt) -- Default to equal NaNs in make_merge_sets_aggregation. ([#11952](https://github.com/rapidsai/cudf/pull/11952)) [@bdice](https://github.com/bdice) -- Add `strip_delimiters` option to `read_text` ([#11946](https://github.com/rapidsai/cudf/pull/11946)) [@upsj](https://github.com/upsj) -- Refactor multibyte_split `output_builder` ([#11945](https://github.com/rapidsai/cudf/pull/11945)) [@upsj](https://github.com/upsj) -- Remove validation that requires introspection ([#11938](https://github.com/rapidsai/cudf/pull/11938)) [@vyasr](https://github.com/vyasr) -- Add `.str.find_multiple` API ([#11928](https://github.com/rapidsai/cudf/pull/11928)) [@galipremsagar](https://github.com/galipremsagar) -- Add regex_program class for use with all regex APIs ([#11927](https://github.com/rapidsai/cudf/pull/11927)) [@davidwendt](https://github.com/davidwendt) -- Enable backend dispatching for Dask-DataFrame creation ([#11920](https://github.com/rapidsai/cudf/pull/11920)) [@rjzamora](https://github.com/rjzamora) -- Performance improvement in JSON Tree traversal ([#11919](https://github.com/rapidsai/cudf/pull/11919)) [@karthikeyann](https://github.com/karthikeyann) -- Fix some gtests incorrectly coded in namespace cudf::test (part I) ([#11917](https://github.com/rapidsai/cudf/pull/11917)) [@davidwendt](https://github.com/davidwendt) -- Refactor pad/zfill functions for reuse with strings udf ([#11914](https://github.com/rapidsai/cudf/pull/11914)) [@davidwendt](https://github.com/davidwendt) -- Add `nanosecond` & `microsecond` to `DatetimeProperties` ([#11911](https://github.com/rapidsai/cudf/pull/11911)) [@galipremsagar](https://github.com/galipremsagar) -- Pin mimesis version in setup.py. ([#11906](https://github.com/rapidsai/cudf/pull/11906)) [@bdice](https://github.com/bdice) -- Error on `ListColumn` or any new unsupported column in `cudf.Index` ([#11902](https://github.com/rapidsai/cudf/pull/11902)) [@galipremsagar](https://github.com/galipremsagar) -- Add thrust output iterator fix (1805) to thrust.patch ([#11900](https://github.com/rapidsai/cudf/pull/11900)) [@davidwendt](https://github.com/davidwendt) -- Relax `codecov` threshold diff ([#11899](https://github.com/rapidsai/cudf/pull/11899)) [@galipremsagar](https://github.com/galipremsagar) -- Use public APIs in STREAM_COMPACTION_NVBENCH ([#11892](https://github.com/rapidsai/cudf/pull/11892)) [@GregoryKimball](https://github.com/GregoryKimball) -- Add coverage for string UDF tests. ([#11891](https://github.com/rapidsai/cudf/pull/11891)) [@vyasr](https://github.com/vyasr) -- Provide `data_chunk_source` wrapper for `datasource` ([#11886](https://github.com/rapidsai/cudf/pull/11886)) [@upsj](https://github.com/upsj) -- Handle `multibyte_split` byte_range out-of-bounds offsets on host ([#11885](https://github.com/rapidsai/cudf/pull/11885)) [@upsj](https://github.com/upsj) -- Add tests ensuring that cudf's default stream is always used ([#11875](https://github.com/rapidsai/cudf/pull/11875)) [@vyasr](https://github.com/vyasr) -- Change expect_strings_empty into expect_column_empty libcudf test utility ([#11873](https://github.com/rapidsai/cudf/pull/11873)) [@davidwendt](https://github.com/davidwendt) -- Add ngroup ([#11871](https://github.com/rapidsai/cudf/pull/11871)) [@shwina](https://github.com/shwina) -- Reduce memory usage in nested JSON parser - tree generation ([#11864](https://github.com/rapidsai/cudf/pull/11864)) [@karthikeyann](https://github.com/karthikeyann) -- Unpin `dask` and `distributed` for development ([#11859](https://github.com/rapidsai/cudf/pull/11859)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unused includes for table/row_operators ([#11857](https://github.com/rapidsai/cudf/pull/11857)) [@GregoryKimball](https://github.com/GregoryKimball) -- Use conda-forge's `pyorc` ([#11855](https://github.com/rapidsai/cudf/pull/11855)) [@jakirkham](https://github.com/jakirkham) -- Add libcudf strings examples ([#11849](https://github.com/rapidsai/cudf/pull/11849)) [@davidwendt](https://github.com/davidwendt) -- Remove `cudf_io` namespace alias ([#11827](https://github.com/rapidsai/cudf/pull/11827)) [@vuule](https://github.com/vuule) -- Test/remove thrust vector usage ([#11813](https://github.com/rapidsai/cudf/pull/11813)) [@vyasr](https://github.com/vyasr) -- Add BGZIP reader to python `read_text` ([#11802](https://github.com/rapidsai/cudf/pull/11802)) [@upsj](https://github.com/upsj) -- Merge branch-22.10 into branch-22.12 ([#11801](https://github.com/rapidsai/cudf/pull/11801)) [@davidwendt](https://github.com/davidwendt) -- Fix compile warning from CUDF_FUNC_RANGE in a member function ([#11798](https://github.com/rapidsai/cudf/pull/11798)) [@davidwendt](https://github.com/davidwendt) -- Update cudf JNI version to 22.12.0-SNAPSHOT ([#11764](https://github.com/rapidsai/cudf/pull/11764)) [@pxLi](https://github.com/pxLi) -- Update flake8 to 5.0.4 and use flake8-force to check Cython. ([#11736](https://github.com/rapidsai/cudf/pull/11736)) [@bdice](https://github.com/bdice) -- Add BGZIP multibyte_split benchmark ([#11723](https://github.com/rapidsai/cudf/pull/11723)) [@upsj](https://github.com/upsj) -- Bifurcate Dependency Lists ([#11674](https://github.com/rapidsai/cudf/pull/11674)) [@bdice](https://github.com/bdice) -- Default to equal NaNs in make_collect_set_aggregation. ([#11621](https://github.com/rapidsai/cudf/pull/11621)) [@bdice](https://github.com/bdice) -- Conform "bench_isin" to match generator column names ([#11549](https://github.com/rapidsai/cudf/pull/11549)) [@GregoryKimball](https://github.com/GregoryKimball) -- Removing int8 column option from parquet byte_array writing ([#11539](https://github.com/rapidsai/cudf/pull/11539)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add checks for HLG layers in dask-cudf groupby tests ([#10853](https://github.com/rapidsai/cudf/pull/10853)) [@charlesbluca](https://github.com/charlesbluca) -- part1: Simplify BaseIndex to an abstract class ([#10389](https://github.com/rapidsai/cudf/pull/10389)) [@skirui-source](https://github.com/skirui-source) -- Make all `nvcc` warnings into errors ([#8916](https://github.com/rapidsai/cudf/pull/8916)) [@trxcllnt](https://github.com/trxcllnt) +- Reduce number of tests marked `spilling` ([#12197](https://github.com/NVIDIA/cudf/pull/12197)) [@madsbk](https://github.com/madsbk) +- Pin `dask` and `distributed` for release ([#12165](https://github.com/NVIDIA/cudf/pull/12165)) [@galipremsagar](https://github.com/galipremsagar) +- Don't rely on GNU find in headers_test.sh ([#12164](https://github.com/NVIDIA/cudf/pull/12164)) [@wence-](https://github.com/wence-) +- Update cp.clip call ([#12148](https://github.com/NVIDIA/cudf/pull/12148)) [@quasiben](https://github.com/quasiben) +- Enable automatic column projection in groupby().agg ([#12124](https://github.com/NVIDIA/cudf/pull/12124)) [@rjzamora](https://github.com/rjzamora) +- Refactor `purge_nonempty_nulls` ([#12111](https://github.com/NVIDIA/cudf/pull/12111)) [@ttnghia](https://github.com/ttnghia) +- Create an `int8` column in `read_csv` when all elements are missing ([#12110](https://github.com/NVIDIA/cudf/pull/12110)) [@vuule](https://github.com/vuule) +- Spilling to host memory ([#12106](https://github.com/NVIDIA/cudf/pull/12106)) [@madsbk](https://github.com/madsbk) +- First pass of `pd.read_orc` changes in tests ([#12103](https://github.com/NVIDIA/cudf/pull/12103)) [@galipremsagar](https://github.com/galipremsagar) +- Expose engine argument in dask_cudf.read_json ([#12101](https://github.com/NVIDIA/cudf/pull/12101)) [@rjzamora](https://github.com/rjzamora) +- Remove CUDA 10 compatibility code. ([#12088](https://github.com/NVIDIA/cudf/pull/12088)) [@bdice](https://github.com/bdice) +- Move and update `dask` nigthly install in CI ([#12082](https://github.com/NVIDIA/cudf/pull/12082)) [@galipremsagar](https://github.com/galipremsagar) +- Throw an error when libcudf is built without cuFile and `LIBCUDF_CUFILE_POLICY` is set to `"ALWAYS"` ([#12080](https://github.com/NVIDIA/cudf/pull/12080)) [@vuule](https://github.com/vuule) +- Remove macros that inspect the contents of exceptions ([#12076](https://github.com/NVIDIA/cudf/pull/12076)) [@vyasr](https://github.com/vyasr) +- Fix ingest_raw_data performance issue in Nested JSON reader due to RVO ([#12070](https://github.com/NVIDIA/cudf/pull/12070)) [@karthikeyann](https://github.com/karthikeyann) +- Remove overflow error during decimal binops ([#12063](https://github.com/NVIDIA/cudf/pull/12063)) [@galipremsagar](https://github.com/galipremsagar) +- Change cudf::detail::tdigest to cudf::tdigest::detail ([#12050](https://github.com/NVIDIA/cudf/pull/12050)) [@davidwendt](https://github.com/davidwendt) +- Fix quantile gtests coded in namespace cudf::test ([#12049](https://github.com/NVIDIA/cudf/pull/12049)) [@davidwendt](https://github.com/davidwendt) +- Add support for `DataFrame.from_dict`\`to_dict` and `Series.to_dict` ([#12048](https://github.com/NVIDIA/cudf/pull/12048)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor Parquet reader ([#12046](https://github.com/NVIDIA/cudf/pull/12046)) [@ttnghia](https://github.com/ttnghia) +- Forward merge 22.10 into 22.12 ([#12045](https://github.com/NVIDIA/cudf/pull/12045)) [@vyasr](https://github.com/vyasr) +- Standardize newlines at ends of files. ([#12042](https://github.com/NVIDIA/cudf/pull/12042)) [@bdice](https://github.com/bdice) +- Trim trailing whitespace from all files. ([#12041](https://github.com/NVIDIA/cudf/pull/12041)) [@bdice](https://github.com/bdice) +- Use nosync policy in gather and scatter implementations. ([#12038](https://github.com/NVIDIA/cudf/pull/12038)) [@bdice](https://github.com/bdice) +- Remove smart quotes from all docstrings. ([#12035](https://github.com/NVIDIA/cudf/pull/12035)) [@bdice](https://github.com/bdice) +- Update cuda-python dependency to 11.7.1 ([#12030](https://github.com/NVIDIA/cudf/pull/12030)) [@galipremsagar](https://github.com/galipremsagar) +- Add cython-lint to pre-commit checks. ([#12020](https://github.com/NVIDIA/cudf/pull/12020)) [@bdice](https://github.com/bdice) +- Use pragma once ([#12019](https://github.com/NVIDIA/cudf/pull/12019)) [@bdice](https://github.com/bdice) +- New GHA to add issues/prs to project board ([#12016](https://github.com/NVIDIA/cudf/pull/12016)) [@jarmak-nv](https://github.com/jarmak-nv) +- Add DataFrame.pivot_table. ([#12015](https://github.com/NVIDIA/cudf/pull/12015)) [@bdice](https://github.com/bdice) +- Rollback of `DeviceBufferLike` ([#12009](https://github.com/NVIDIA/cudf/pull/12009)) [@madsbk](https://github.com/madsbk) +- Remove default parameters for nvtext::detail functions ([#12007](https://github.com/NVIDIA/cudf/pull/12007)) [@davidwendt](https://github.com/davidwendt) +- Remove default parameters for cudf::dictionary::detail functions ([#12006](https://github.com/NVIDIA/cudf/pull/12006)) [@davidwendt](https://github.com/davidwendt) +- Remove unused `managed_allocator` ([#12005](https://github.com/NVIDIA/cudf/pull/12005)) [@vyasr](https://github.com/vyasr) +- Remove default parameters for cudf::strings::detail functions ([#12003](https://github.com/NVIDIA/cudf/pull/12003)) [@davidwendt](https://github.com/davidwendt) +- Remove unnecessary code from dask-cudf _Frame ([#12001](https://github.com/NVIDIA/cudf/pull/12001)) [@rjzamora](https://github.com/rjzamora) +- Ignore python docs build artifacts ([#12000](https://github.com/NVIDIA/cudf/pull/12000)) [@galipremsagar](https://github.com/galipremsagar) +- Use rapids-cmake for google benchmark. ([#11997](https://github.com/NVIDIA/cudf/pull/11997)) [@vyasr](https://github.com/vyasr) +- Leverage rapids_cython for more automated RPATH handling ([#11996](https://github.com/NVIDIA/cudf/pull/11996)) [@vyasr](https://github.com/vyasr) +- Remove stale labeler ([#11995](https://github.com/NVIDIA/cudf/pull/11995)) [@raydouglass](https://github.com/raydouglass) +- Move protobuf compilation to CMake ([#11986](https://github.com/NVIDIA/cudf/pull/11986)) [@vyasr](https://github.com/vyasr) +- Replace most of preprocessor usage in nvcomp adapter with `constexpr` ([#11980](https://github.com/NVIDIA/cudf/pull/11980)) [@vuule](https://github.com/vuule) +- Add missing noexcepts to column_in_metadata methods ([#11973](https://github.com/NVIDIA/cudf/pull/11973)) [@vyasr](https://github.com/vyasr) +- Pass column names to `write_csv` instead of `table_metadata` pointer ([#11972](https://github.com/NVIDIA/cudf/pull/11972)) [@vuule](https://github.com/vuule) +- Accelerate libcudf segmented sort with CUB segmented sort ([#11969](https://github.com/NVIDIA/cudf/pull/11969)) [@davidwendt](https://github.com/davidwendt) +- Feature/remove default streams ([#11967](https://github.com/NVIDIA/cudf/pull/11967)) [@vyasr](https://github.com/vyasr) +- Add pool memory resource to libcudf basic example ([#11966](https://github.com/NVIDIA/cudf/pull/11966)) [@davidwendt](https://github.com/davidwendt) +- Fix some libcudf calls to cudf::detail::gather ([#11963](https://github.com/NVIDIA/cudf/pull/11963)) [@davidwendt](https://github.com/davidwendt) +- Accept const refs instead of const unique_ptr refs in reduce and scan APIs. ([#11960](https://github.com/NVIDIA/cudf/pull/11960)) [@vyasr](https://github.com/vyasr) +- Add deprecation warning for set_allocator. ([#11958](https://github.com/NVIDIA/cudf/pull/11958)) [@vyasr](https://github.com/vyasr) +- Fix lists and structs gtests coded in namespace cudf::test ([#11956](https://github.com/NVIDIA/cudf/pull/11956)) [@davidwendt](https://github.com/davidwendt) +- Add full page indexes to Parquet writer benchmarks ([#11955](https://github.com/NVIDIA/cudf/pull/11955)) [@etseidl](https://github.com/etseidl) +- Use gather-based strings factory in cudf::strings::strip ([#11954](https://github.com/NVIDIA/cudf/pull/11954)) [@davidwendt](https://github.com/davidwendt) +- Default to equal NaNs in make_merge_sets_aggregation. ([#11952](https://github.com/NVIDIA/cudf/pull/11952)) [@bdice](https://github.com/bdice) +- Add `strip_delimiters` option to `read_text` ([#11946](https://github.com/NVIDIA/cudf/pull/11946)) [@upsj](https://github.com/upsj) +- Refactor multibyte_split `output_builder` ([#11945](https://github.com/NVIDIA/cudf/pull/11945)) [@upsj](https://github.com/upsj) +- Remove validation that requires introspection ([#11938](https://github.com/NVIDIA/cudf/pull/11938)) [@vyasr](https://github.com/vyasr) +- Add `.str.find_multiple` API ([#11928](https://github.com/NVIDIA/cudf/pull/11928)) [@galipremsagar](https://github.com/galipremsagar) +- Add regex_program class for use with all regex APIs ([#11927](https://github.com/NVIDIA/cudf/pull/11927)) [@davidwendt](https://github.com/davidwendt) +- Enable backend dispatching for Dask-DataFrame creation ([#11920](https://github.com/NVIDIA/cudf/pull/11920)) [@rjzamora](https://github.com/rjzamora) +- Performance improvement in JSON Tree traversal ([#11919](https://github.com/NVIDIA/cudf/pull/11919)) [@karthikeyann](https://github.com/karthikeyann) +- Fix some gtests incorrectly coded in namespace cudf::test (part I) ([#11917](https://github.com/NVIDIA/cudf/pull/11917)) [@davidwendt](https://github.com/davidwendt) +- Refactor pad/zfill functions for reuse with strings udf ([#11914](https://github.com/NVIDIA/cudf/pull/11914)) [@davidwendt](https://github.com/davidwendt) +- Add `nanosecond` & `microsecond` to `DatetimeProperties` ([#11911](https://github.com/NVIDIA/cudf/pull/11911)) [@galipremsagar](https://github.com/galipremsagar) +- Pin mimesis version in setup.py. ([#11906](https://github.com/NVIDIA/cudf/pull/11906)) [@bdice](https://github.com/bdice) +- Error on `ListColumn` or any new unsupported column in `cudf.Index` ([#11902](https://github.com/NVIDIA/cudf/pull/11902)) [@galipremsagar](https://github.com/galipremsagar) +- Add thrust output iterator fix (1805) to thrust.patch ([#11900](https://github.com/NVIDIA/cudf/pull/11900)) [@davidwendt](https://github.com/davidwendt) +- Relax `codecov` threshold diff ([#11899](https://github.com/NVIDIA/cudf/pull/11899)) [@galipremsagar](https://github.com/galipremsagar) +- Use public APIs in STREAM_COMPACTION_NVBENCH ([#11892](https://github.com/NVIDIA/cudf/pull/11892)) [@GregoryKimball](https://github.com/GregoryKimball) +- Add coverage for string UDF tests. ([#11891](https://github.com/NVIDIA/cudf/pull/11891)) [@vyasr](https://github.com/vyasr) +- Provide `data_chunk_source` wrapper for `datasource` ([#11886](https://github.com/NVIDIA/cudf/pull/11886)) [@upsj](https://github.com/upsj) +- Handle `multibyte_split` byte_range out-of-bounds offsets on host ([#11885](https://github.com/NVIDIA/cudf/pull/11885)) [@upsj](https://github.com/upsj) +- Add tests ensuring that cudf's default stream is always used ([#11875](https://github.com/NVIDIA/cudf/pull/11875)) [@vyasr](https://github.com/vyasr) +- Change expect_strings_empty into expect_column_empty libcudf test utility ([#11873](https://github.com/NVIDIA/cudf/pull/11873)) [@davidwendt](https://github.com/davidwendt) +- Add ngroup ([#11871](https://github.com/NVIDIA/cudf/pull/11871)) [@shwina](https://github.com/shwina) +- Reduce memory usage in nested JSON parser - tree generation ([#11864](https://github.com/NVIDIA/cudf/pull/11864)) [@karthikeyann](https://github.com/karthikeyann) +- Unpin `dask` and `distributed` for development ([#11859](https://github.com/NVIDIA/cudf/pull/11859)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unused includes for table/row_operators ([#11857](https://github.com/NVIDIA/cudf/pull/11857)) [@GregoryKimball](https://github.com/GregoryKimball) +- Use conda-forge's `pyorc` ([#11855](https://github.com/NVIDIA/cudf/pull/11855)) [@jakirkham](https://github.com/jakirkham) +- Add libcudf strings examples ([#11849](https://github.com/NVIDIA/cudf/pull/11849)) [@davidwendt](https://github.com/davidwendt) +- Remove `cudf_io` namespace alias ([#11827](https://github.com/NVIDIA/cudf/pull/11827)) [@vuule](https://github.com/vuule) +- Test/remove thrust vector usage ([#11813](https://github.com/NVIDIA/cudf/pull/11813)) [@vyasr](https://github.com/vyasr) +- Add BGZIP reader to python `read_text` ([#11802](https://github.com/NVIDIA/cudf/pull/11802)) [@upsj](https://github.com/upsj) +- Merge branch-22.10 into branch-22.12 ([#11801](https://github.com/NVIDIA/cudf/pull/11801)) [@davidwendt](https://github.com/davidwendt) +- Fix compile warning from CUDF_FUNC_RANGE in a member function ([#11798](https://github.com/NVIDIA/cudf/pull/11798)) [@davidwendt](https://github.com/davidwendt) +- Update cudf JNI version to 22.12.0-SNAPSHOT ([#11764](https://github.com/NVIDIA/cudf/pull/11764)) [@pxLi](https://github.com/pxLi) +- Update flake8 to 5.0.4 and use flake8-force to check Cython. ([#11736](https://github.com/NVIDIA/cudf/pull/11736)) [@bdice](https://github.com/bdice) +- Add BGZIP multibyte_split benchmark ([#11723](https://github.com/NVIDIA/cudf/pull/11723)) [@upsj](https://github.com/upsj) +- Bifurcate Dependency Lists ([#11674](https://github.com/NVIDIA/cudf/pull/11674)) [@bdice](https://github.com/bdice) +- Default to equal NaNs in make_collect_set_aggregation. ([#11621](https://github.com/NVIDIA/cudf/pull/11621)) [@bdice](https://github.com/bdice) +- Conform "bench_isin" to match generator column names ([#11549](https://github.com/NVIDIA/cudf/pull/11549)) [@GregoryKimball](https://github.com/GregoryKimball) +- Removing int8 column option from parquet byte_array writing ([#11539](https://github.com/NVIDIA/cudf/pull/11539)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add checks for HLG layers in dask-cudf groupby tests ([#10853](https://github.com/NVIDIA/cudf/pull/10853)) [@charlesbluca](https://github.com/charlesbluca) +- part1: Simplify BaseIndex to an abstract class ([#10389](https://github.com/NVIDIA/cudf/pull/10389)) [@skirui-source](https://github.com/skirui-source) +- Make all `nvcc` warnings into errors ([#8916](https://github.com/NVIDIA/cudf/pull/8916)) [@trxcllnt](https://github.com/trxcllnt) # cuDF 22.10.00 (12 Oct 2022) ## 🚨 Breaking Changes -- Disable Zstandard decompression on nvCOMP 2.4 and Pascal GPus ([#11856](https://github.com/rapidsai/cudf/pull/11856)) [@vuule](https://github.com/vuule) -- Disable nvCOMP DEFLATE integration ([#11811](https://github.com/rapidsai/cudf/pull/11811)) [@vuule](https://github.com/vuule) -- Fix return type of `Index.isna` & `Index.notna` ([#11769](https://github.com/rapidsai/cudf/pull/11769)) [@galipremsagar](https://github.com/galipremsagar) -- Remove `kwargs` in `read_csv` & `to_csv` ([#11762](https://github.com/rapidsai/cudf/pull/11762)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `cudf::partition*` APIs that do not return offsets for empty output table ([#11709](https://github.com/rapidsai/cudf/pull/11709)) [@ttnghia](https://github.com/ttnghia) -- Fix regex negated classes to not automatically include new-lines ([#11644](https://github.com/rapidsai/cudf/pull/11644)) [@davidwendt](https://github.com/davidwendt) -- Update zfill to match Python output ([#11634](https://github.com/rapidsai/cudf/pull/11634)) [@davidwendt](https://github.com/davidwendt) -- Upgrade `pandas` to `1.5` ([#11617](https://github.com/rapidsai/cudf/pull/11617)) [@galipremsagar](https://github.com/galipremsagar) -- Change default value of `ordered` to `False` in `CategoricalDtype` ([#11604](https://github.com/rapidsai/cudf/pull/11604)) [@galipremsagar](https://github.com/galipremsagar) -- Move cudf::strings::findall_record to cudf::strings::findall ([#11575](https://github.com/rapidsai/cudf/pull/11575)) [@davidwendt](https://github.com/davidwendt) -- Adding optional parquet reader schema ([#11524](https://github.com/rapidsai/cudf/pull/11524)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Deprecate `skiprows` and `num_rows` in `read_orc` ([#11522](https://github.com/rapidsai/cudf/pull/11522)) [@galipremsagar](https://github.com/galipremsagar) -- Remove support for skip_rows / num_rows options in the parquet reader. ([#11503](https://github.com/rapidsai/cudf/pull/11503)) [@nvdbaranec](https://github.com/nvdbaranec) -- Drop support for `skiprows` and `num_rows` in `cudf.read_parquet` ([#11480](https://github.com/rapidsai/cudf/pull/11480)) [@galipremsagar](https://github.com/galipremsagar) -- Disable Arrow S3 support by default. ([#11470](https://github.com/rapidsai/cudf/pull/11470)) [@bdice](https://github.com/bdice) -- Convert thrust::optional usages to std::optional ([#11455](https://github.com/rapidsai/cudf/pull/11455)) [@robertmaynard](https://github.com/robertmaynard) -- Remove unused is_struct trait. ([#11450](https://github.com/rapidsai/cudf/pull/11450)) [@bdice](https://github.com/bdice) -- Refactor the `Buffer` class ([#11447](https://github.com/rapidsai/cudf/pull/11447)) [@madsbk](https://github.com/madsbk) -- Return empty dataframe when reading an ORC file using empty `columns` option ([#11446](https://github.com/rapidsai/cudf/pull/11446)) [@vuule](https://github.com/vuule) -- Refactor pad_side and strip_type enums into side_type enum ([#11438](https://github.com/rapidsai/cudf/pull/11438)) [@davidwendt](https://github.com/davidwendt) -- Remove HASH_SERIAL_MURMUR3 / serial32BitMurmurHash3 ([#11383](https://github.com/rapidsai/cudf/pull/11383)) [@bdice](https://github.com/bdice) -- Use the new JSON parser when the experimental reader is selected ([#11364](https://github.com/rapidsai/cudf/pull/11364)) [@vuule](https://github.com/vuule) -- Remove deprecated Series.applymap. ([#11031](https://github.com/rapidsai/cudf/pull/11031)) [@bdice](https://github.com/bdice) -- Remove deprecated expand parameter from str.findall. ([#11030](https://github.com/rapidsai/cudf/pull/11030)) [@bdice](https://github.com/bdice) +- Disable Zstandard decompression on nvCOMP 2.4 and Pascal GPus ([#11856](https://github.com/NVIDIA/cudf/pull/11856)) [@vuule](https://github.com/vuule) +- Disable nvCOMP DEFLATE integration ([#11811](https://github.com/NVIDIA/cudf/pull/11811)) [@vuule](https://github.com/vuule) +- Fix return type of `Index.isna` & `Index.notna` ([#11769](https://github.com/NVIDIA/cudf/pull/11769)) [@galipremsagar](https://github.com/galipremsagar) +- Remove `kwargs` in `read_csv` & `to_csv` ([#11762](https://github.com/NVIDIA/cudf/pull/11762)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `cudf::partition*` APIs that do not return offsets for empty output table ([#11709](https://github.com/NVIDIA/cudf/pull/11709)) [@ttnghia](https://github.com/ttnghia) +- Fix regex negated classes to not automatically include new-lines ([#11644](https://github.com/NVIDIA/cudf/pull/11644)) [@davidwendt](https://github.com/davidwendt) +- Update zfill to match Python output ([#11634](https://github.com/NVIDIA/cudf/pull/11634)) [@davidwendt](https://github.com/davidwendt) +- Upgrade `pandas` to `1.5` ([#11617](https://github.com/NVIDIA/cudf/pull/11617)) [@galipremsagar](https://github.com/galipremsagar) +- Change default value of `ordered` to `False` in `CategoricalDtype` ([#11604](https://github.com/NVIDIA/cudf/pull/11604)) [@galipremsagar](https://github.com/galipremsagar) +- Move cudf::strings::findall_record to cudf::strings::findall ([#11575](https://github.com/NVIDIA/cudf/pull/11575)) [@davidwendt](https://github.com/davidwendt) +- Adding optional parquet reader schema ([#11524](https://github.com/NVIDIA/cudf/pull/11524)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Deprecate `skiprows` and `num_rows` in `read_orc` ([#11522](https://github.com/NVIDIA/cudf/pull/11522)) [@galipremsagar](https://github.com/galipremsagar) +- Remove support for skip_rows / num_rows options in the parquet reader. ([#11503](https://github.com/NVIDIA/cudf/pull/11503)) [@nvdbaranec](https://github.com/nvdbaranec) +- Drop support for `skiprows` and `num_rows` in `cudf.read_parquet` ([#11480](https://github.com/NVIDIA/cudf/pull/11480)) [@galipremsagar](https://github.com/galipremsagar) +- Disable Arrow S3 support by default. ([#11470](https://github.com/NVIDIA/cudf/pull/11470)) [@bdice](https://github.com/bdice) +- Convert thrust::optional usages to std::optional ([#11455](https://github.com/NVIDIA/cudf/pull/11455)) [@robertmaynard](https://github.com/robertmaynard) +- Remove unused is_struct trait. ([#11450](https://github.com/NVIDIA/cudf/pull/11450)) [@bdice](https://github.com/bdice) +- Refactor the `Buffer` class ([#11447](https://github.com/NVIDIA/cudf/pull/11447)) [@madsbk](https://github.com/madsbk) +- Return empty dataframe when reading an ORC file using empty `columns` option ([#11446](https://github.com/NVIDIA/cudf/pull/11446)) [@vuule](https://github.com/vuule) +- Refactor pad_side and strip_type enums into side_type enum ([#11438](https://github.com/NVIDIA/cudf/pull/11438)) [@davidwendt](https://github.com/davidwendt) +- Remove HASH_SERIAL_MURMUR3 / serial32BitMurmurHash3 ([#11383](https://github.com/NVIDIA/cudf/pull/11383)) [@bdice](https://github.com/bdice) +- Use the new JSON parser when the experimental reader is selected ([#11364](https://github.com/NVIDIA/cudf/pull/11364)) [@vuule](https://github.com/vuule) +- Remove deprecated Series.applymap. ([#11031](https://github.com/NVIDIA/cudf/pull/11031)) [@bdice](https://github.com/bdice) +- Remove deprecated expand parameter from str.findall. ([#11030](https://github.com/NVIDIA/cudf/pull/11030)) [@bdice](https://github.com/bdice) ## 🐛 Bug Fixes -- Fixes bug in temporary decompression space estimation before calling nvcomp ([#11879](https://github.com/rapidsai/cudf/pull/11879)) [@abellina](https://github.com/abellina) -- Handle `ptx` file paths during `strings_udf` import ([#11862](https://github.com/rapidsai/cudf/pull/11862)) [@galipremsagar](https://github.com/galipremsagar) -- Disable Zstandard decompression on nvCOMP 2.4 and Pascal GPus ([#11856](https://github.com/rapidsai/cudf/pull/11856)) [@vuule](https://github.com/vuule) -- Reset `strings_udf` CEC and solve several related issues ([#11846](https://github.com/rapidsai/cudf/pull/11846)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix bug in new shuffle-based groupby implementation ([#11836](https://github.com/rapidsai/cudf/pull/11836)) [@rjzamora](https://github.com/rjzamora) -- Fix `is_valid` checks in `Scalar._binaryop` ([#11818](https://github.com/rapidsai/cudf/pull/11818)) [@wence-](https://github.com/wence-) -- Fix operator `NotImplemented` issue with `numpy` ([#11816](https://github.com/rapidsai/cudf/pull/11816)) [@galipremsagar](https://github.com/galipremsagar) -- Disable nvCOMP DEFLATE integration ([#11811](https://github.com/rapidsai/cudf/pull/11811)) [@vuule](https://github.com/vuule) -- Build `strings_udf` package with other python packages in nightlies ([#11808](https://github.com/rapidsai/cudf/pull/11808)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Revert problematic shuffle=explicit-comms changes ([#11803](https://github.com/rapidsai/cudf/pull/11803)) [@rjzamora](https://github.com/rjzamora) -- Fix regex out-of-bounds write in strided rows logic ([#11797](https://github.com/rapidsai/cudf/pull/11797)) [@davidwendt](https://github.com/davidwendt) -- Build `cudf` locally before building `strings_udf` conda packages in CI ([#11785](https://github.com/rapidsai/cudf/pull/11785)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix an issue in cudf::row_bit_count involving structs and lists at multiple levels. ([#11779](https://github.com/rapidsai/cudf/pull/11779)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix return type of `Index.isna` & `Index.notna` ([#11769](https://github.com/rapidsai/cudf/pull/11769)) [@galipremsagar](https://github.com/galipremsagar) -- Fix issue with set-item in case of `list` and `struct` types ([#11760](https://github.com/rapidsai/cudf/pull/11760)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure all libcudf APIs run on cudf's default stream ([#11759](https://github.com/rapidsai/cudf/pull/11759)) [@vyasr](https://github.com/vyasr) -- Resolve dask_cudf failures caused by upstream groupby changes ([#11755](https://github.com/rapidsai/cudf/pull/11755)) [@rjzamora](https://github.com/rjzamora) -- Fix ORC string sum statistics ([#11740](https://github.com/rapidsai/cudf/pull/11740)) [@vuule](https://github.com/vuule) -- Add `strings_udf` package for python 3.9 ([#11730](https://github.com/rapidsai/cudf/pull/11730)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Ensure that all tests launch kernels on cudf's default stream ([#11726](https://github.com/rapidsai/cudf/pull/11726)) [@vyasr](https://github.com/vyasr) -- Don't assume stream is a compile-time constant expression ([#11725](https://github.com/rapidsai/cudf/pull/11725)) [@vyasr](https://github.com/vyasr) -- Fix get_thrust.cmake format at patch command ([#11715](https://github.com/rapidsai/cudf/pull/11715)) [@davidwendt](https://github.com/davidwendt) -- Fix `cudf::partition*` APIs that do not return offsets for empty output table ([#11709](https://github.com/rapidsai/cudf/pull/11709)) [@ttnghia](https://github.com/ttnghia) -- Fix cudf::lists::sort_lists for NaN and Infinity values ([#11703](https://github.com/rapidsai/cudf/pull/11703)) [@davidwendt](https://github.com/davidwendt) -- Modify ORC reader timestamp parsing to match the apache reader behavior ([#11699](https://github.com/rapidsai/cudf/pull/11699)) [@vuule](https://github.com/vuule) -- Fix `DataFrame.from_arrow` to preserve type metadata ([#11698](https://github.com/rapidsai/cudf/pull/11698)) [@galipremsagar](https://github.com/galipremsagar) -- Fix compile error due to missing header ([#11697](https://github.com/rapidsai/cudf/pull/11697)) [@ttnghia](https://github.com/ttnghia) -- Default to Snappy compression in `to_orc` when using cuDF or Dask ([#11690](https://github.com/rapidsai/cudf/pull/11690)) [@vuule](https://github.com/vuule) -- Fix an issue related to `Multindex` when `group_keys=True` ([#11689](https://github.com/rapidsai/cudf/pull/11689)) [@galipremsagar](https://github.com/galipremsagar) -- Transfer correct dtype to exploded column ([#11687](https://github.com/rapidsai/cudf/pull/11687)) [@wence-](https://github.com/wence-) -- Ignore protobuf generated files in `mypy` checks ([#11685](https://github.com/rapidsai/cudf/pull/11685)) [@galipremsagar](https://github.com/galipremsagar) -- Maintain the index name after `.loc` ([#11677](https://github.com/rapidsai/cudf/pull/11677)) [@shwina](https://github.com/shwina) -- Fix issue with extracting nested column data & dtype preservation ([#11671](https://github.com/rapidsai/cudf/pull/11671)) [@galipremsagar](https://github.com/galipremsagar) -- Ensure that all cudf tests and benchmarks are conda env aware ([#11666](https://github.com/rapidsai/cudf/pull/11666)) [@robertmaynard](https://github.com/robertmaynard) -- Update to Thrust 1.17.2 to fix cub ODR issues ([#11665](https://github.com/rapidsai/cudf/pull/11665)) [@robertmaynard](https://github.com/robertmaynard) -- Fix multi-file remote datasource bug ([#11655](https://github.com/rapidsai/cudf/pull/11655)) [@rjzamora](https://github.com/rjzamora) -- Fix invalid regex quantifier check to not include alternation ([#11654](https://github.com/rapidsai/cudf/pull/11654)) [@davidwendt](https://github.com/davidwendt) -- Fix bug in `device_write()`: it uses an incorrect size ([#11651](https://github.com/rapidsai/cudf/pull/11651)) [@madsbk](https://github.com/madsbk) -- fixes overflows in benchmarks ([#11649](https://github.com/rapidsai/cudf/pull/11649)) [@elstehle](https://github.com/elstehle) -- Fix regex negated classes to not automatically include new-lines ([#11644](https://github.com/rapidsai/cudf/pull/11644)) [@davidwendt](https://github.com/davidwendt) -- Fix compile error in benchmark nested_json.cpp ([#11637](https://github.com/rapidsai/cudf/pull/11637)) [@davidwendt](https://github.com/davidwendt) -- Update zfill to match Python output ([#11634](https://github.com/rapidsai/cudf/pull/11634)) [@davidwendt](https://github.com/davidwendt) -- Removed converted type for INT32 and INT64 since they do not convert ([#11627](https://github.com/rapidsai/cudf/pull/11627)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix host scalars construction of nested types ([#11612](https://github.com/rapidsai/cudf/pull/11612)) [@galipremsagar](https://github.com/galipremsagar) -- Fix compile warning in nested_json_gpu.cu ([#11607](https://github.com/rapidsai/cudf/pull/11607)) [@davidwendt](https://github.com/davidwendt) -- Change default value of `ordered` to `False` in `CategoricalDtype` ([#11604](https://github.com/rapidsai/cudf/pull/11604)) [@galipremsagar](https://github.com/galipremsagar) -- Preserve order if necessary when deduping categoricals internally ([#11597](https://github.com/rapidsai/cudf/pull/11597)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add is_timestamp test for leap second (60) ([#11594](https://github.com/rapidsai/cudf/pull/11594)) [@davidwendt](https://github.com/davidwendt) -- Fix an issue with `to_arrow` when column name type is not a string ([#11590](https://github.com/rapidsai/cudf/pull/11590)) [@galipremsagar](https://github.com/galipremsagar) -- Fix exception in segmented-reduce benchmark ([#11588](https://github.com/rapidsai/cudf/pull/11588)) [@davidwendt](https://github.com/davidwendt) -- Fix encode/decode of negative timestamps in ORC reader/writer ([#11586](https://github.com/rapidsai/cudf/pull/11586)) [@vuule](https://github.com/vuule) -- Correct distribution data type in `quantiles` benchmark ([#11584](https://github.com/rapidsai/cudf/pull/11584)) [@vuule](https://github.com/vuule) -- Fix multibyte_split benchmark for host buffers ([#11583](https://github.com/rapidsai/cudf/pull/11583)) [@upsj](https://github.com/upsj) -- xfail custreamz display test for now ([#11567](https://github.com/rapidsai/cudf/pull/11567)) [@shwina](https://github.com/shwina) -- Fix JNI for TableWithMeta to use schema_info instead of column_names ([#11566](https://github.com/rapidsai/cudf/pull/11566)) [@jlowe](https://github.com/jlowe) -- Reduce code duplication for `dask` & `distributed` nightly/stable installs ([#11565](https://github.com/rapidsai/cudf/pull/11565)) [@galipremsagar](https://github.com/galipremsagar) -- Fix groupby failures in dask_cudf CI ([#11561](https://github.com/rapidsai/cudf/pull/11561)) [@rjzamora](https://github.com/rjzamora) -- Fix for pivot: error when 'values' is a multicharacter string ([#11538](https://github.com/rapidsai/cudf/pull/11538)) [@shaswat-indian](https://github.com/shaswat-indian) -- find_package(cudf) + arrow9 usable with cudf build directory ([#11535](https://github.com/rapidsai/cudf/pull/11535)) [@robertmaynard](https://github.com/robertmaynard) -- Fixing crash when writing binary nested data in parquet ([#11526](https://github.com/rapidsai/cudf/pull/11526)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix for: error when assigning a value to an empty series ([#11523](https://github.com/rapidsai/cudf/pull/11523)) [@shaswat-indian](https://github.com/shaswat-indian) -- Fix invalid results from conditional-left-anti-join in debug build ([#11517](https://github.com/rapidsai/cudf/pull/11517)) [@davidwendt](https://github.com/davidwendt) -- Fix cmake error after upgrading to Arrow 9 ([#11513](https://github.com/rapidsai/cudf/pull/11513)) [@ttnghia](https://github.com/ttnghia) -- Fix reverse binary operators acting on a host value and cudf.Scalar ([#11512](https://github.com/rapidsai/cudf/pull/11512)) [@bdice](https://github.com/bdice) -- Update parquet fuzz tests to drop support for `skiprows` & `num_rows` ([#11505](https://github.com/rapidsai/cudf/pull/11505)) [@galipremsagar](https://github.com/galipremsagar) -- Use rapids-cmake 22.10 best practice for RAPIDS.cmake location ([#11493](https://github.com/rapidsai/cudf/pull/11493)) [@robertmaynard](https://github.com/robertmaynard) -- Handle some zero-sized corner cases in dlpack interop ([#11449](https://github.com/rapidsai/cudf/pull/11449)) [@wence-](https://github.com/wence-) -- Return empty dataframe when reading an ORC file using empty `columns` option ([#11446](https://github.com/rapidsai/cudf/pull/11446)) [@vuule](https://github.com/vuule) -- libcudf c++ example updated to CPM version 0.35.3 ([#11417](https://github.com/rapidsai/cudf/pull/11417)) [@robertmaynard](https://github.com/robertmaynard) -- Fix regex quantifier check to include capture groups ([#11373](https://github.com/rapidsai/cudf/pull/11373)) [@davidwendt](https://github.com/davidwendt) -- Fix read_text when byte_range is aligned with field ([#11371](https://github.com/rapidsai/cudf/pull/11371)) [@upsj](https://github.com/upsj) -- Fix to_timestamps truncated subsecond calculation ([#11367](https://github.com/rapidsai/cudf/pull/11367)) [@davidwendt](https://github.com/davidwendt) -- column: calculate null_count before release()ing the cudf::column ([#11365](https://github.com/rapidsai/cudf/pull/11365)) [@wence-](https://github.com/wence-) +- Fixes bug in temporary decompression space estimation before calling nvcomp ([#11879](https://github.com/NVIDIA/cudf/pull/11879)) [@abellina](https://github.com/abellina) +- Handle `ptx` file paths during `strings_udf` import ([#11862](https://github.com/NVIDIA/cudf/pull/11862)) [@galipremsagar](https://github.com/galipremsagar) +- Disable Zstandard decompression on nvCOMP 2.4 and Pascal GPus ([#11856](https://github.com/NVIDIA/cudf/pull/11856)) [@vuule](https://github.com/vuule) +- Reset `strings_udf` CEC and solve several related issues ([#11846](https://github.com/NVIDIA/cudf/pull/11846)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix bug in new shuffle-based groupby implementation ([#11836](https://github.com/NVIDIA/cudf/pull/11836)) [@rjzamora](https://github.com/rjzamora) +- Fix `is_valid` checks in `Scalar._binaryop` ([#11818](https://github.com/NVIDIA/cudf/pull/11818)) [@wence-](https://github.com/wence-) +- Fix operator `NotImplemented` issue with `numpy` ([#11816](https://github.com/NVIDIA/cudf/pull/11816)) [@galipremsagar](https://github.com/galipremsagar) +- Disable nvCOMP DEFLATE integration ([#11811](https://github.com/NVIDIA/cudf/pull/11811)) [@vuule](https://github.com/vuule) +- Build `strings_udf` package with other python packages in nightlies ([#11808](https://github.com/NVIDIA/cudf/pull/11808)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Revert problematic shuffle=explicit-comms changes ([#11803](https://github.com/NVIDIA/cudf/pull/11803)) [@rjzamora](https://github.com/rjzamora) +- Fix regex out-of-bounds write in strided rows logic ([#11797](https://github.com/NVIDIA/cudf/pull/11797)) [@davidwendt](https://github.com/davidwendt) +- Build `cudf` locally before building `strings_udf` conda packages in CI ([#11785](https://github.com/NVIDIA/cudf/pull/11785)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix an issue in cudf::row_bit_count involving structs and lists at multiple levels. ([#11779](https://github.com/NVIDIA/cudf/pull/11779)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix return type of `Index.isna` & `Index.notna` ([#11769](https://github.com/NVIDIA/cudf/pull/11769)) [@galipremsagar](https://github.com/galipremsagar) +- Fix issue with set-item in case of `list` and `struct` types ([#11760](https://github.com/NVIDIA/cudf/pull/11760)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure all libcudf APIs run on cudf's default stream ([#11759](https://github.com/NVIDIA/cudf/pull/11759)) [@vyasr](https://github.com/vyasr) +- Resolve dask_cudf failures caused by upstream groupby changes ([#11755](https://github.com/NVIDIA/cudf/pull/11755)) [@rjzamora](https://github.com/rjzamora) +- Fix ORC string sum statistics ([#11740](https://github.com/NVIDIA/cudf/pull/11740)) [@vuule](https://github.com/vuule) +- Add `strings_udf` package for python 3.9 ([#11730](https://github.com/NVIDIA/cudf/pull/11730)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Ensure that all tests launch kernels on cudf's default stream ([#11726](https://github.com/NVIDIA/cudf/pull/11726)) [@vyasr](https://github.com/vyasr) +- Don't assume stream is a compile-time constant expression ([#11725](https://github.com/NVIDIA/cudf/pull/11725)) [@vyasr](https://github.com/vyasr) +- Fix get_thrust.cmake format at patch command ([#11715](https://github.com/NVIDIA/cudf/pull/11715)) [@davidwendt](https://github.com/davidwendt) +- Fix `cudf::partition*` APIs that do not return offsets for empty output table ([#11709](https://github.com/NVIDIA/cudf/pull/11709)) [@ttnghia](https://github.com/ttnghia) +- Fix cudf::lists::sort_lists for NaN and Infinity values ([#11703](https://github.com/NVIDIA/cudf/pull/11703)) [@davidwendt](https://github.com/davidwendt) +- Modify ORC reader timestamp parsing to match the apache reader behavior ([#11699](https://github.com/NVIDIA/cudf/pull/11699)) [@vuule](https://github.com/vuule) +- Fix `DataFrame.from_arrow` to preserve type metadata ([#11698](https://github.com/NVIDIA/cudf/pull/11698)) [@galipremsagar](https://github.com/galipremsagar) +- Fix compile error due to missing header ([#11697](https://github.com/NVIDIA/cudf/pull/11697)) [@ttnghia](https://github.com/ttnghia) +- Default to Snappy compression in `to_orc` when using cuDF or Dask ([#11690](https://github.com/NVIDIA/cudf/pull/11690)) [@vuule](https://github.com/vuule) +- Fix an issue related to `Multindex` when `group_keys=True` ([#11689](https://github.com/NVIDIA/cudf/pull/11689)) [@galipremsagar](https://github.com/galipremsagar) +- Transfer correct dtype to exploded column ([#11687](https://github.com/NVIDIA/cudf/pull/11687)) [@wence-](https://github.com/wence-) +- Ignore protobuf generated files in `mypy` checks ([#11685](https://github.com/NVIDIA/cudf/pull/11685)) [@galipremsagar](https://github.com/galipremsagar) +- Maintain the index name after `.loc` ([#11677](https://github.com/NVIDIA/cudf/pull/11677)) [@shwina](https://github.com/shwina) +- Fix issue with extracting nested column data & dtype preservation ([#11671](https://github.com/NVIDIA/cudf/pull/11671)) [@galipremsagar](https://github.com/galipremsagar) +- Ensure that all cudf tests and benchmarks are conda env aware ([#11666](https://github.com/NVIDIA/cudf/pull/11666)) [@robertmaynard](https://github.com/robertmaynard) +- Update to Thrust 1.17.2 to fix cub ODR issues ([#11665](https://github.com/NVIDIA/cudf/pull/11665)) [@robertmaynard](https://github.com/robertmaynard) +- Fix multi-file remote datasource bug ([#11655](https://github.com/NVIDIA/cudf/pull/11655)) [@rjzamora](https://github.com/rjzamora) +- Fix invalid regex quantifier check to not include alternation ([#11654](https://github.com/NVIDIA/cudf/pull/11654)) [@davidwendt](https://github.com/davidwendt) +- Fix bug in `device_write()`: it uses an incorrect size ([#11651](https://github.com/NVIDIA/cudf/pull/11651)) [@madsbk](https://github.com/madsbk) +- fixes overflows in benchmarks ([#11649](https://github.com/NVIDIA/cudf/pull/11649)) [@elstehle](https://github.com/elstehle) +- Fix regex negated classes to not automatically include new-lines ([#11644](https://github.com/NVIDIA/cudf/pull/11644)) [@davidwendt](https://github.com/davidwendt) +- Fix compile error in benchmark nested_json.cpp ([#11637](https://github.com/NVIDIA/cudf/pull/11637)) [@davidwendt](https://github.com/davidwendt) +- Update zfill to match Python output ([#11634](https://github.com/NVIDIA/cudf/pull/11634)) [@davidwendt](https://github.com/davidwendt) +- Removed converted type for INT32 and INT64 since they do not convert ([#11627](https://github.com/NVIDIA/cudf/pull/11627)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix host scalars construction of nested types ([#11612](https://github.com/NVIDIA/cudf/pull/11612)) [@galipremsagar](https://github.com/galipremsagar) +- Fix compile warning in nested_json_gpu.cu ([#11607](https://github.com/NVIDIA/cudf/pull/11607)) [@davidwendt](https://github.com/davidwendt) +- Change default value of `ordered` to `False` in `CategoricalDtype` ([#11604](https://github.com/NVIDIA/cudf/pull/11604)) [@galipremsagar](https://github.com/galipremsagar) +- Preserve order if necessary when deduping categoricals internally ([#11597](https://github.com/NVIDIA/cudf/pull/11597)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add is_timestamp test for leap second (60) ([#11594](https://github.com/NVIDIA/cudf/pull/11594)) [@davidwendt](https://github.com/davidwendt) +- Fix an issue with `to_arrow` when column name type is not a string ([#11590](https://github.com/NVIDIA/cudf/pull/11590)) [@galipremsagar](https://github.com/galipremsagar) +- Fix exception in segmented-reduce benchmark ([#11588](https://github.com/NVIDIA/cudf/pull/11588)) [@davidwendt](https://github.com/davidwendt) +- Fix encode/decode of negative timestamps in ORC reader/writer ([#11586](https://github.com/NVIDIA/cudf/pull/11586)) [@vuule](https://github.com/vuule) +- Correct distribution data type in `quantiles` benchmark ([#11584](https://github.com/NVIDIA/cudf/pull/11584)) [@vuule](https://github.com/vuule) +- Fix multibyte_split benchmark for host buffers ([#11583](https://github.com/NVIDIA/cudf/pull/11583)) [@upsj](https://github.com/upsj) +- xfail custreamz display test for now ([#11567](https://github.com/NVIDIA/cudf/pull/11567)) [@shwina](https://github.com/shwina) +- Fix JNI for TableWithMeta to use schema_info instead of column_names ([#11566](https://github.com/NVIDIA/cudf/pull/11566)) [@jlowe](https://github.com/jlowe) +- Reduce code duplication for `dask` & `distributed` nightly/stable installs ([#11565](https://github.com/NVIDIA/cudf/pull/11565)) [@galipremsagar](https://github.com/galipremsagar) +- Fix groupby failures in dask_cudf CI ([#11561](https://github.com/NVIDIA/cudf/pull/11561)) [@rjzamora](https://github.com/rjzamora) +- Fix for pivot: error when 'values' is a multicharacter string ([#11538](https://github.com/NVIDIA/cudf/pull/11538)) [@shaswat-indian](https://github.com/shaswat-indian) +- find_package(cudf) + arrow9 usable with cudf build directory ([#11535](https://github.com/NVIDIA/cudf/pull/11535)) [@robertmaynard](https://github.com/robertmaynard) +- Fixing crash when writing binary nested data in parquet ([#11526](https://github.com/NVIDIA/cudf/pull/11526)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix for: error when assigning a value to an empty series ([#11523](https://github.com/NVIDIA/cudf/pull/11523)) [@shaswat-indian](https://github.com/shaswat-indian) +- Fix invalid results from conditional-left-anti-join in debug build ([#11517](https://github.com/NVIDIA/cudf/pull/11517)) [@davidwendt](https://github.com/davidwendt) +- Fix cmake error after upgrading to Arrow 9 ([#11513](https://github.com/NVIDIA/cudf/pull/11513)) [@ttnghia](https://github.com/ttnghia) +- Fix reverse binary operators acting on a host value and cudf.Scalar ([#11512](https://github.com/NVIDIA/cudf/pull/11512)) [@bdice](https://github.com/bdice) +- Update parquet fuzz tests to drop support for `skiprows` & `num_rows` ([#11505](https://github.com/NVIDIA/cudf/pull/11505)) [@galipremsagar](https://github.com/galipremsagar) +- Use rapids-cmake 22.10 best practice for RAPIDS.cmake location ([#11493](https://github.com/NVIDIA/cudf/pull/11493)) [@robertmaynard](https://github.com/robertmaynard) +- Handle some zero-sized corner cases in dlpack interop ([#11449](https://github.com/NVIDIA/cudf/pull/11449)) [@wence-](https://github.com/wence-) +- Return empty dataframe when reading an ORC file using empty `columns` option ([#11446](https://github.com/NVIDIA/cudf/pull/11446)) [@vuule](https://github.com/vuule) +- libcudf c++ example updated to CPM version 0.35.3 ([#11417](https://github.com/NVIDIA/cudf/pull/11417)) [@robertmaynard](https://github.com/robertmaynard) +- Fix regex quantifier check to include capture groups ([#11373](https://github.com/NVIDIA/cudf/pull/11373)) [@davidwendt](https://github.com/davidwendt) +- Fix read_text when byte_range is aligned with field ([#11371](https://github.com/NVIDIA/cudf/pull/11371)) [@upsj](https://github.com/upsj) +- Fix to_timestamps truncated subsecond calculation ([#11367](https://github.com/NVIDIA/cudf/pull/11367)) [@davidwendt](https://github.com/davidwendt) +- column: calculate null_count before release()ing the cudf::column ([#11365](https://github.com/NVIDIA/cudf/pull/11365)) [@wence-](https://github.com/wence-) ## 📖 Documentation -- Update `guide-to-udfs` notebook ([#11861](https://github.com/rapidsai/cudf/pull/11861)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Update docstring for cudf.read_text ([#11799](https://github.com/rapidsai/cudf/pull/11799)) [@GregoryKimball](https://github.com/GregoryKimball) -- Add doc section for `list` & `struct` handling ([#11770](https://github.com/rapidsai/cudf/pull/11770)) [@galipremsagar](https://github.com/galipremsagar) -- Document that minimum required CMake version is now 3.23.1 ([#11751](https://github.com/rapidsai/cudf/pull/11751)) [@robertmaynard](https://github.com/robertmaynard) -- Update libcudf documentation build command in DOCUMENTATION.md ([#11735](https://github.com/rapidsai/cudf/pull/11735)) [@davidwendt](https://github.com/davidwendt) -- Add docs for use of string data to `DataFrame.apply` and `Series.apply` and update guide to UDFs notebook ([#11733](https://github.com/rapidsai/cudf/pull/11733)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Enable more Pydocstyle rules ([#11582](https://github.com/rapidsai/cudf/pull/11582)) [@bdice](https://github.com/bdice) -- Remove unused cpp/img folder ([#11554](https://github.com/rapidsai/cudf/pull/11554)) [@davidwendt](https://github.com/davidwendt) -- Publish C++ developer docs ([#11475](https://github.com/rapidsai/cudf/pull/11475)) [@vyasr](https://github.com/vyasr) -- Fix a misalignment in `cudf.get_dummies` docstring ([#11443](https://github.com/rapidsai/cudf/pull/11443)) [@galipremsagar](https://github.com/galipremsagar) -- Update contributing doc to include links to the developer guides ([#11390](https://github.com/rapidsai/cudf/pull/11390)) [@davidwendt](https://github.com/davidwendt) -- Fix table_view_base doxygen format ([#11340](https://github.com/rapidsai/cudf/pull/11340)) [@davidwendt](https://github.com/davidwendt) -- Create main developer guide for Python ([#11235](https://github.com/rapidsai/cudf/pull/11235)) [@vyasr](https://github.com/vyasr) -- Add developer documentation for benchmarking ([#11122](https://github.com/rapidsai/cudf/pull/11122)) [@vyasr](https://github.com/vyasr) -- cuDF error handling document ([#7917](https://github.com/rapidsai/cudf/pull/7917)) [@isVoid](https://github.com/isVoid) +- Update `guide-to-udfs` notebook ([#11861](https://github.com/NVIDIA/cudf/pull/11861)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Update docstring for cudf.read_text ([#11799](https://github.com/NVIDIA/cudf/pull/11799)) [@GregoryKimball](https://github.com/GregoryKimball) +- Add doc section for `list` & `struct` handling ([#11770](https://github.com/NVIDIA/cudf/pull/11770)) [@galipremsagar](https://github.com/galipremsagar) +- Document that minimum required CMake version is now 3.23.1 ([#11751](https://github.com/NVIDIA/cudf/pull/11751)) [@robertmaynard](https://github.com/robertmaynard) +- Update libcudf documentation build command in DOCUMENTATION.md ([#11735](https://github.com/NVIDIA/cudf/pull/11735)) [@davidwendt](https://github.com/davidwendt) +- Add docs for use of string data to `DataFrame.apply` and `Series.apply` and update guide to UDFs notebook ([#11733](https://github.com/NVIDIA/cudf/pull/11733)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Enable more Pydocstyle rules ([#11582](https://github.com/NVIDIA/cudf/pull/11582)) [@bdice](https://github.com/bdice) +- Remove unused cpp/img folder ([#11554](https://github.com/NVIDIA/cudf/pull/11554)) [@davidwendt](https://github.com/davidwendt) +- Publish C++ developer docs ([#11475](https://github.com/NVIDIA/cudf/pull/11475)) [@vyasr](https://github.com/vyasr) +- Fix a misalignment in `cudf.get_dummies` docstring ([#11443](https://github.com/NVIDIA/cudf/pull/11443)) [@galipremsagar](https://github.com/galipremsagar) +- Update contributing doc to include links to the developer guides ([#11390](https://github.com/NVIDIA/cudf/pull/11390)) [@davidwendt](https://github.com/davidwendt) +- Fix table_view_base doxygen format ([#11340](https://github.com/NVIDIA/cudf/pull/11340)) [@davidwendt](https://github.com/davidwendt) +- Create main developer guide for Python ([#11235](https://github.com/NVIDIA/cudf/pull/11235)) [@vyasr](https://github.com/vyasr) +- Add developer documentation for benchmarking ([#11122](https://github.com/NVIDIA/cudf/pull/11122)) [@vyasr](https://github.com/vyasr) +- cuDF error handling document ([#7917](https://github.com/NVIDIA/cudf/pull/7917)) [@isVoid](https://github.com/isVoid) ## 🚀 New Features -- Add hasNull statistic reading ability to ORC ([#11747](https://github.com/rapidsai/cudf/pull/11747)) [@devavret](https://github.com/devavret) -- Add `istitle` to string UDFs ([#11738](https://github.com/rapidsai/cudf/pull/11738)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- JSON Column creation in GPU ([#11714](https://github.com/rapidsai/cudf/pull/11714)) [@karthikeyann](https://github.com/karthikeyann) -- Adds option to take explicit nested schema for nested JSON reader ([#11682](https://github.com/rapidsai/cudf/pull/11682)) [@elstehle](https://github.com/elstehle) -- Add BGZIP `data_chunk_reader` ([#11652](https://github.com/rapidsai/cudf/pull/11652)) [@upsj](https://github.com/upsj) -- Support DECIMAL order-by for RANGE window functions ([#11645](https://github.com/rapidsai/cudf/pull/11645)) [@mythrocks](https://github.com/mythrocks) -- changing version of cmake to 3.23.3 ([#11619](https://github.com/rapidsai/cudf/pull/11619)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Generate unique keys table in java JNI `contiguousSplitGroups` ([#11614](https://github.com/rapidsai/cudf/pull/11614)) [@res-life](https://github.com/res-life) -- Generic type casting to support the new nested JSON reader ([#11613](https://github.com/rapidsai/cudf/pull/11613)) [@elstehle](https://github.com/elstehle) -- JSON tree traversal ([#11610](https://github.com/rapidsai/cudf/pull/11610)) [@karthikeyann](https://github.com/karthikeyann) -- Add casting operators to masked UDFs ([#11578](https://github.com/rapidsai/cudf/pull/11578)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Adds type inference and type conversion for leaf-columns to the nested JSON parser ([#11574](https://github.com/rapidsai/cudf/pull/11574)) [@elstehle](https://github.com/elstehle) -- Add strings 'like' function ([#11558](https://github.com/rapidsai/cudf/pull/11558)) [@davidwendt](https://github.com/davidwendt) -- Handle hyphen as literal for regex cclass when incomplete range ([#11557](https://github.com/rapidsai/cudf/pull/11557)) [@davidwendt](https://github.com/davidwendt) -- Enable ZSTD compression in ORC and Parquet writers ([#11551](https://github.com/rapidsai/cudf/pull/11551)) [@vuule](https://github.com/vuule) -- Adds support for json lines format to the nested JSON reader ([#11534](https://github.com/rapidsai/cudf/pull/11534)) [@elstehle](https://github.com/elstehle) -- Adding optional parquet reader schema ([#11524](https://github.com/rapidsai/cudf/pull/11524)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Adds GPU implementation of JSON-token-stream to JSON-tree ([#11518](https://github.com/rapidsai/cudf/pull/11518)) [@karthikeyann](https://github.com/karthikeyann) -- Add `gdb` pretty-printers for simple types ([#11499](https://github.com/rapidsai/cudf/pull/11499)) [@upsj](https://github.com/upsj) -- Add `create_random_column` function to the data generator ([#11490](https://github.com/rapidsai/cudf/pull/11490)) [@vuule](https://github.com/vuule) -- Add fluent API builder to `data_profile` ([#11479](https://github.com/rapidsai/cudf/pull/11479)) [@vuule](https://github.com/vuule) -- Adds Nested Json benchmark ([#11466](https://github.com/rapidsai/cudf/pull/11466)) [@karthikeyann](https://github.com/karthikeyann) -- Convert thrust::optional usages to std::optional ([#11455](https://github.com/rapidsai/cudf/pull/11455)) [@robertmaynard](https://github.com/robertmaynard) -- Python API for the future experimental JSON reader ([#11426](https://github.com/rapidsai/cudf/pull/11426)) [@vuule](https://github.com/vuule) -- Return schema info from JSON reader ([#11419](https://github.com/rapidsai/cudf/pull/11419)) [@vuule](https://github.com/vuule) -- Add regex ASCII flag support for matching builtin character classes ([#11404](https://github.com/rapidsai/cudf/pull/11404)) [@davidwendt](https://github.com/davidwendt) -- Truncate parquet column indexes ([#11403](https://github.com/rapidsai/cudf/pull/11403)) [@etseidl](https://github.com/etseidl) -- Adds the end-to-end JSON parser implementation ([#11388](https://github.com/rapidsai/cudf/pull/11388)) [@elstehle](https://github.com/elstehle) -- Use the new JSON parser when the experimental reader is selected ([#11364](https://github.com/rapidsai/cudf/pull/11364)) [@vuule](https://github.com/vuule) -- Add placeholder for the experimental JSON reader ([#11334](https://github.com/rapidsai/cudf/pull/11334)) [@vuule](https://github.com/vuule) -- Add read-only functions on string dtypes to `DataFrame.apply` and `Series.apply` ([#11319](https://github.com/rapidsai/cudf/pull/11319)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Added 'crosstab' and 'pivot_table' features ([#11314](https://github.com/rapidsai/cudf/pull/11314)) [@shaswat-indian](https://github.com/shaswat-indian) -- Quickly error out when trying to build with unsupported nvcc versions ([#11297](https://github.com/rapidsai/cudf/pull/11297)) [@robertmaynard](https://github.com/robertmaynard) -- Adds JSON tokenizer ([#11264](https://github.com/rapidsai/cudf/pull/11264)) [@elstehle](https://github.com/elstehle) -- List lexicographic comparator ([#11129](https://github.com/rapidsai/cudf/pull/11129)) [@devavret](https://github.com/devavret) -- Add generic type inference for cuIO ([#11121](https://github.com/rapidsai/cudf/pull/11121)) [@PointKernel](https://github.com/PointKernel) -- Fully support nested types in `cudf::contains` ([#10656](https://github.com/rapidsai/cudf/pull/10656)) [@ttnghia](https://github.com/ttnghia) -- Support nested types in `lists::contains` ([#10548](https://github.com/rapidsai/cudf/pull/10548)) [@ttnghia](https://github.com/ttnghia) +- Add hasNull statistic reading ability to ORC ([#11747](https://github.com/NVIDIA/cudf/pull/11747)) [@devavret](https://github.com/devavret) +- Add `istitle` to string UDFs ([#11738](https://github.com/NVIDIA/cudf/pull/11738)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- JSON Column creation in GPU ([#11714](https://github.com/NVIDIA/cudf/pull/11714)) [@karthikeyann](https://github.com/karthikeyann) +- Adds option to take explicit nested schema for nested JSON reader ([#11682](https://github.com/NVIDIA/cudf/pull/11682)) [@elstehle](https://github.com/elstehle) +- Add BGZIP `data_chunk_reader` ([#11652](https://github.com/NVIDIA/cudf/pull/11652)) [@upsj](https://github.com/upsj) +- Support DECIMAL order-by for RANGE window functions ([#11645](https://github.com/NVIDIA/cudf/pull/11645)) [@mythrocks](https://github.com/mythrocks) +- changing version of cmake to 3.23.3 ([#11619](https://github.com/NVIDIA/cudf/pull/11619)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Generate unique keys table in java JNI `contiguousSplitGroups` ([#11614](https://github.com/NVIDIA/cudf/pull/11614)) [@res-life](https://github.com/res-life) +- Generic type casting to support the new nested JSON reader ([#11613](https://github.com/NVIDIA/cudf/pull/11613)) [@elstehle](https://github.com/elstehle) +- JSON tree traversal ([#11610](https://github.com/NVIDIA/cudf/pull/11610)) [@karthikeyann](https://github.com/karthikeyann) +- Add casting operators to masked UDFs ([#11578](https://github.com/NVIDIA/cudf/pull/11578)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Adds type inference and type conversion for leaf-columns to the nested JSON parser ([#11574](https://github.com/NVIDIA/cudf/pull/11574)) [@elstehle](https://github.com/elstehle) +- Add strings 'like' function ([#11558](https://github.com/NVIDIA/cudf/pull/11558)) [@davidwendt](https://github.com/davidwendt) +- Handle hyphen as literal for regex cclass when incomplete range ([#11557](https://github.com/NVIDIA/cudf/pull/11557)) [@davidwendt](https://github.com/davidwendt) +- Enable ZSTD compression in ORC and Parquet writers ([#11551](https://github.com/NVIDIA/cudf/pull/11551)) [@vuule](https://github.com/vuule) +- Adds support for json lines format to the nested JSON reader ([#11534](https://github.com/NVIDIA/cudf/pull/11534)) [@elstehle](https://github.com/elstehle) +- Adding optional parquet reader schema ([#11524](https://github.com/NVIDIA/cudf/pull/11524)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Adds GPU implementation of JSON-token-stream to JSON-tree ([#11518](https://github.com/NVIDIA/cudf/pull/11518)) [@karthikeyann](https://github.com/karthikeyann) +- Add `gdb` pretty-printers for simple types ([#11499](https://github.com/NVIDIA/cudf/pull/11499)) [@upsj](https://github.com/upsj) +- Add `create_random_column` function to the data generator ([#11490](https://github.com/NVIDIA/cudf/pull/11490)) [@vuule](https://github.com/vuule) +- Add fluent API builder to `data_profile` ([#11479](https://github.com/NVIDIA/cudf/pull/11479)) [@vuule](https://github.com/vuule) +- Adds Nested Json benchmark ([#11466](https://github.com/NVIDIA/cudf/pull/11466)) [@karthikeyann](https://github.com/karthikeyann) +- Convert thrust::optional usages to std::optional ([#11455](https://github.com/NVIDIA/cudf/pull/11455)) [@robertmaynard](https://github.com/robertmaynard) +- Python API for the future experimental JSON reader ([#11426](https://github.com/NVIDIA/cudf/pull/11426)) [@vuule](https://github.com/vuule) +- Return schema info from JSON reader ([#11419](https://github.com/NVIDIA/cudf/pull/11419)) [@vuule](https://github.com/vuule) +- Add regex ASCII flag support for matching builtin character classes ([#11404](https://github.com/NVIDIA/cudf/pull/11404)) [@davidwendt](https://github.com/davidwendt) +- Truncate parquet column indexes ([#11403](https://github.com/NVIDIA/cudf/pull/11403)) [@etseidl](https://github.com/etseidl) +- Adds the end-to-end JSON parser implementation ([#11388](https://github.com/NVIDIA/cudf/pull/11388)) [@elstehle](https://github.com/elstehle) +- Use the new JSON parser when the experimental reader is selected ([#11364](https://github.com/NVIDIA/cudf/pull/11364)) [@vuule](https://github.com/vuule) +- Add placeholder for the experimental JSON reader ([#11334](https://github.com/NVIDIA/cudf/pull/11334)) [@vuule](https://github.com/vuule) +- Add read-only functions on string dtypes to `DataFrame.apply` and `Series.apply` ([#11319](https://github.com/NVIDIA/cudf/pull/11319)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Added 'crosstab' and 'pivot_table' features ([#11314](https://github.com/NVIDIA/cudf/pull/11314)) [@shaswat-indian](https://github.com/shaswat-indian) +- Quickly error out when trying to build with unsupported nvcc versions ([#11297](https://github.com/NVIDIA/cudf/pull/11297)) [@robertmaynard](https://github.com/robertmaynard) +- Adds JSON tokenizer ([#11264](https://github.com/NVIDIA/cudf/pull/11264)) [@elstehle](https://github.com/elstehle) +- List lexicographic comparator ([#11129](https://github.com/NVIDIA/cudf/pull/11129)) [@devavret](https://github.com/devavret) +- Add generic type inference for cuIO ([#11121](https://github.com/NVIDIA/cudf/pull/11121)) [@PointKernel](https://github.com/PointKernel) +- Fully support nested types in `cudf::contains` ([#10656](https://github.com/NVIDIA/cudf/pull/10656)) [@ttnghia](https://github.com/ttnghia) +- Support nested types in `lists::contains` ([#10548](https://github.com/NVIDIA/cudf/pull/10548)) [@ttnghia](https://github.com/ttnghia) ## 🛠️ Improvements -- Pin `dask` and `distributed` for release ([#11822](https://github.com/rapidsai/cudf/pull/11822)) [@galipremsagar](https://github.com/galipremsagar) -- Add examples for Nested JSON reader ([#11814](https://github.com/rapidsai/cudf/pull/11814)) [@GregoryKimball](https://github.com/GregoryKimball) -- Support shuffle-based groupby aggregations in dask_cudf ([#11800](https://github.com/rapidsai/cudf/pull/11800)) [@rjzamora](https://github.com/rjzamora) -- Update strings udf version updater script ([#11772](https://github.com/rapidsai/cudf/pull/11772)) [@galipremsagar](https://github.com/galipremsagar) -- Remove `kwargs` in `read_csv` & `to_csv` ([#11762](https://github.com/rapidsai/cudf/pull/11762)) [@galipremsagar](https://github.com/galipremsagar) -- Pass `dtype` param to avoid `pd.Series` warnings ([#11761](https://github.com/rapidsai/cudf/pull/11761)) [@galipremsagar](https://github.com/galipremsagar) -- Enable `schema_element` & `keep_quotes` support in json reader ([#11746](https://github.com/rapidsai/cudf/pull/11746)) [@galipremsagar](https://github.com/galipremsagar) -- Add ability to construct `ListColumn` when size is `None` ([#11745](https://github.com/rapidsai/cudf/pull/11745)) [@galipremsagar](https://github.com/galipremsagar) -- Reduces memory requirements in JSON parser and adds bytes/s and peak memory usage to benchmarks ([#11732](https://github.com/rapidsai/cudf/pull/11732)) [@elstehle](https://github.com/elstehle) -- Add missing copyright headers. ([#11712](https://github.com/rapidsai/cudf/pull/11712)) [@bdice](https://github.com/bdice) -- Fix copyright check issues in pre-commit ([#11711](https://github.com/rapidsai/cudf/pull/11711)) [@bdice](https://github.com/bdice) -- Include decimal in supported types for range window order-by columns ([#11710](https://github.com/rapidsai/cudf/pull/11710)) [@mythrocks](https://github.com/mythrocks) -- Disable very large column gtest for contiguous-split ([#11706](https://github.com/rapidsai/cudf/pull/11706)) [@davidwendt](https://github.com/davidwendt) -- Drop split_out=None test from groupby.agg ([#11704](https://github.com/rapidsai/cudf/pull/11704)) [@wence-](https://github.com/wence-) -- Use CubinLinker for CUDA Minor Version Compatibility ([#11701](https://github.com/rapidsai/cudf/pull/11701)) [@gmarkall](https://github.com/gmarkall) -- Add regex capture-group parameter to auto convert to non-capture groups ([#11695](https://github.com/rapidsai/cudf/pull/11695)) [@davidwendt](https://github.com/davidwendt) -- Add a `__dataframe__` method to the protocol dataframe object ([#11692](https://github.com/rapidsai/cudf/pull/11692)) [@rgommers](https://github.com/rgommers) -- Special-case multibyte_split for single-byte delimiter ([#11681](https://github.com/rapidsai/cudf/pull/11681)) [@upsj](https://github.com/upsj) -- Remove isort exclusions ([#11680](https://github.com/rapidsai/cudf/pull/11680)) [@bdice](https://github.com/bdice) -- Refactor CSV reader benchmarks with nvbench ([#11678](https://github.com/rapidsai/cudf/pull/11678)) [@PointKernel](https://github.com/PointKernel) -- Check conda recipe headers with pre-commit ([#11669](https://github.com/rapidsai/cudf/pull/11669)) [@bdice](https://github.com/bdice) -- Remove redundant style check for clang-format. ([#11668](https://github.com/rapidsai/cudf/pull/11668)) [@bdice](https://github.com/bdice) -- Add support for `group_keys` in `groupby` ([#11659](https://github.com/rapidsai/cudf/pull/11659)) [@galipremsagar](https://github.com/galipremsagar) -- Fix pandoc pinning. ([#11658](https://github.com/rapidsai/cudf/pull/11658)) [@bdice](https://github.com/bdice) -- Revert removal of skip_rows / num_rows options from the Parquet reader. ([#11657](https://github.com/rapidsai/cudf/pull/11657)) [@nvdbaranec](https://github.com/nvdbaranec) -- Update git metadata ([#11647](https://github.com/rapidsai/cudf/pull/11647)) [@bdice](https://github.com/bdice) -- Call set_null_count on a returning column if null-count is known ([#11646](https://github.com/rapidsai/cudf/pull/11646)) [@davidwendt](https://github.com/davidwendt) -- Fix some libcudf detail calls not passing the stream variable ([#11642](https://github.com/rapidsai/cudf/pull/11642)) [@davidwendt](https://github.com/davidwendt) -- Update to mypy 0.971 ([#11640](https://github.com/rapidsai/cudf/pull/11640)) [@wence-](https://github.com/wence-) -- Refactor strings strip functor to details header ([#11635](https://github.com/rapidsai/cudf/pull/11635)) [@davidwendt](https://github.com/davidwendt) -- Fix incorrect `nullCount` in `get_json_object` ([#11633](https://github.com/rapidsai/cudf/pull/11633)) [@trxcllnt](https://github.com/trxcllnt) -- Simplify `hostdevice_vector` ([#11631](https://github.com/rapidsai/cudf/pull/11631)) [@upsj](https://github.com/upsj) -- Refactor parquet writer benchmarks with nvbench ([#11623](https://github.com/rapidsai/cudf/pull/11623)) [@PointKernel](https://github.com/PointKernel) -- Rework contains_scalar to check nulls at runtime ([#11622](https://github.com/rapidsai/cudf/pull/11622)) [@davidwendt](https://github.com/davidwendt) -- Fix incorrect memory resource used in rolling temp columns ([#11618](https://github.com/rapidsai/cudf/pull/11618)) [@mythrocks](https://github.com/mythrocks) -- Upgrade `pandas` to `1.5` ([#11617](https://github.com/rapidsai/cudf/pull/11617)) [@galipremsagar](https://github.com/galipremsagar) -- Move type-dispatcher calls from traits.hpp to traits.cpp ([#11616](https://github.com/rapidsai/cudf/pull/11616)) [@davidwendt](https://github.com/davidwendt) -- Refactor parquet reader benchmarks with nvbench ([#11611](https://github.com/rapidsai/cudf/pull/11611)) [@PointKernel](https://github.com/PointKernel) -- Forward-merge branch-22.08 to branch-22.10 ([#11608](https://github.com/rapidsai/cudf/pull/11608)) [@bdice](https://github.com/bdice) -- Use stream in Java API. ([#11601](https://github.com/rapidsai/cudf/pull/11601)) [@bdice](https://github.com/bdice) -- Refactors of public/detail APIs, CUDF_FUNC_RANGE, stream handling. ([#11600](https://github.com/rapidsai/cudf/pull/11600)) [@bdice](https://github.com/bdice) -- Improve ORC writer benchmark with nvbench ([#11598](https://github.com/rapidsai/cudf/pull/11598)) [@PointKernel](https://github.com/PointKernel) -- Tune multibyte_split kernel ([#11587](https://github.com/rapidsai/cudf/pull/11587)) [@upsj](https://github.com/upsj) -- Move split_utils.cuh to strings/detail ([#11585](https://github.com/rapidsai/cudf/pull/11585)) [@davidwendt](https://github.com/davidwendt) -- Fix warnings due to compiler regression with `if constexpr` ([#11581](https://github.com/rapidsai/cudf/pull/11581)) [@ttnghia](https://github.com/ttnghia) -- Add full 24-bit dictionary support to Parquet writer ([#11580](https://github.com/rapidsai/cudf/pull/11580)) [@etseidl](https://github.com/etseidl) -- Expose "explicit-comms" option in shuffle-based dask_cudf functions ([#11576](https://github.com/rapidsai/cudf/pull/11576)) [@rjzamora](https://github.com/rjzamora) -- Move cudf::strings::findall_record to cudf::strings::findall ([#11575](https://github.com/rapidsai/cudf/pull/11575)) [@davidwendt](https://github.com/davidwendt) -- Refactor dask_cudf groupby to use apply_concat_apply ([#11571](https://github.com/rapidsai/cudf/pull/11571)) [@rjzamora](https://github.com/rjzamora) -- Add ability to write `list(struct)` columns as `map` type in orc writer ([#11568](https://github.com/rapidsai/cudf/pull/11568)) [@galipremsagar](https://github.com/galipremsagar) -- Add byte_range to multibyte_split benchmark + NVBench refactor ([#11562](https://github.com/rapidsai/cudf/pull/11562)) [@upsj](https://github.com/upsj) -- JNI support for writing binary columns in parquet ([#11556](https://github.com/rapidsai/cudf/pull/11556)) [@revans2](https://github.com/revans2) -- Support additional dictionary bit widths in Parquet writer ([#11547](https://github.com/rapidsai/cudf/pull/11547)) [@etseidl](https://github.com/etseidl) -- Refactor string/numeric conversion utilities ([#11545](https://github.com/rapidsai/cudf/pull/11545)) [@davidwendt](https://github.com/davidwendt) -- Removing unnecessary asserts in parquet tests ([#11544](https://github.com/rapidsai/cudf/pull/11544)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Clean up ORC reader benchmarks with NVBench ([#11543](https://github.com/rapidsai/cudf/pull/11543)) [@PointKernel](https://github.com/PointKernel) -- Reuse MurmurHash3_32 in Parquet page data. ([#11528](https://github.com/rapidsai/cudf/pull/11528)) [@bdice](https://github.com/bdice) -- Add hexadecimal value separators ([#11527](https://github.com/rapidsai/cudf/pull/11527)) [@bdice](https://github.com/bdice) -- Deprecate `skiprows` and `num_rows` in `read_orc` ([#11522](https://github.com/rapidsai/cudf/pull/11522)) [@galipremsagar](https://github.com/galipremsagar) -- Struct support for `NULL_EQUALS` binary operation ([#11520](https://github.com/rapidsai/cudf/pull/11520)) [@rwlee](https://github.com/rwlee) -- Bump hadoop-common from 3.2.3 to 3.2.4 in /java ([#11516](https://github.com/rapidsai/cudf/pull/11516)) [@dependabot[bot]](https://github.com/dependabot[bot]) -- Fix Feather test warning. ([#11511](https://github.com/rapidsai/cudf/pull/11511)) [@bdice](https://github.com/bdice) -- copy_range ballot_syncs to have no execution dependency ([#11508](https://github.com/rapidsai/cudf/pull/11508)) [@robertmaynard](https://github.com/robertmaynard) -- Upgrade to `arrow-9.x` ([#11507](https://github.com/rapidsai/cudf/pull/11507)) [@galipremsagar](https://github.com/galipremsagar) -- Remove support for skip_rows / num_rows options in the parquet reader. ([#11503](https://github.com/rapidsai/cudf/pull/11503)) [@nvdbaranec](https://github.com/nvdbaranec) -- Single-pass `multibyte_split` ([#11500](https://github.com/rapidsai/cudf/pull/11500)) [@upsj](https://github.com/upsj) -- Sanitize percentile_approx() output for empty input ([#11498](https://github.com/rapidsai/cudf/pull/11498)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Unpin `dask` and `distributed` for development ([#11492](https://github.com/rapidsai/cudf/pull/11492)) [@galipremsagar](https://github.com/galipremsagar) -- Move SparkMurmurHash3_32 functor. ([#11489](https://github.com/rapidsai/cudf/pull/11489)) [@bdice](https://github.com/bdice) -- Refactor group_nunique.cu to use nullate::DYNAMIC for reduce-by-key functor ([#11482](https://github.com/rapidsai/cudf/pull/11482)) [@davidwendt](https://github.com/davidwendt) -- Drop support for `skiprows` and `num_rows` in `cudf.read_parquet` ([#11480](https://github.com/rapidsai/cudf/pull/11480)) [@galipremsagar](https://github.com/galipremsagar) -- Add reduction `distinct_count` benchmark ([#11473](https://github.com/rapidsai/cudf/pull/11473)) [@ttnghia](https://github.com/ttnghia) -- Add groupby `nunique` aggregation benchmark ([#11472](https://github.com/rapidsai/cudf/pull/11472)) [@ttnghia](https://github.com/ttnghia) -- Disable Arrow S3 support by default. ([#11470](https://github.com/rapidsai/cudf/pull/11470)) [@bdice](https://github.com/bdice) -- Add groupby `max` aggregation benchmark ([#11464](https://github.com/rapidsai/cudf/pull/11464)) [@ttnghia](https://github.com/ttnghia) -- Extract Dremel encoding code from Parquet ([#11461](https://github.com/rapidsai/cudf/pull/11461)) [@vyasr](https://github.com/vyasr) -- Add missing Thrust #includes. ([#11457](https://github.com/rapidsai/cudf/pull/11457)) [@bdice](https://github.com/bdice) -- Make CMake hooks verbose ([#11456](https://github.com/rapidsai/cudf/pull/11456)) [@vyasr](https://github.com/vyasr) -- Control Parquet page size through Python API ([#11454](https://github.com/rapidsai/cudf/pull/11454)) [@etseidl](https://github.com/etseidl) -- Add control of Parquet column index creation to python ([#11453](https://github.com/rapidsai/cudf/pull/11453)) [@etseidl](https://github.com/etseidl) -- Remove unused is_struct trait. ([#11450](https://github.com/rapidsai/cudf/pull/11450)) [@bdice](https://github.com/bdice) -- Refactor the `Buffer` class ([#11447](https://github.com/rapidsai/cudf/pull/11447)) [@madsbk](https://github.com/madsbk) -- Refactor pad_side and strip_type enums into side_type enum ([#11438](https://github.com/rapidsai/cudf/pull/11438)) [@davidwendt](https://github.com/davidwendt) -- Update to Thrust 1.17.0 ([#11437](https://github.com/rapidsai/cudf/pull/11437)) [@bdice](https://github.com/bdice) -- Add in JNI for parsing JSON data and getting the metadata back too. ([#11431](https://github.com/rapidsai/cudf/pull/11431)) [@revans2](https://github.com/revans2) -- Convert byte_array_view to use std::byte ([#11424](https://github.com/rapidsai/cudf/pull/11424)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Deprecate unflatten_nested_columns ([#11421](https://github.com/rapidsai/cudf/pull/11421)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Remove HASH_SERIAL_MURMUR3 / serial32BitMurmurHash3 ([#11383](https://github.com/rapidsai/cudf/pull/11383)) [@bdice](https://github.com/bdice) -- Add Spark list hashing Java tests ([#11379](https://github.com/rapidsai/cudf/pull/11379)) [@bdice](https://github.com/bdice) -- Move cmake to the build section. ([#11376](https://github.com/rapidsai/cudf/pull/11376)) [@vyasr](https://github.com/vyasr) -- Remove use of CUDA driver API calls from libcudf ([#11370](https://github.com/rapidsai/cudf/pull/11370)) [@shwina](https://github.com/shwina) -- Add column constructor from device_uvector&& ([#11356](https://github.com/rapidsai/cudf/pull/11356)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Remove unused custreamz thirdparty directory ([#11343](https://github.com/rapidsai/cudf/pull/11343)) [@vyasr](https://github.com/vyasr) -- Update jni version to 22.10.0-SNAPSHOT ([#11338](https://github.com/rapidsai/cudf/pull/11338)) [@pxLi](https://github.com/pxLi) -- Enable using upstream jitify2 ([#11287](https://github.com/rapidsai/cudf/pull/11287)) [@shwina](https://github.com/shwina) -- Cache cudf.Scalar ([#11246](https://github.com/rapidsai/cudf/pull/11246)) [@shwina](https://github.com/shwina) -- Remove deprecated Series.applymap. ([#11031](https://github.com/rapidsai/cudf/pull/11031)) [@bdice](https://github.com/bdice) -- Remove deprecated expand parameter from str.findall. ([#11030](https://github.com/rapidsai/cudf/pull/11030)) [@bdice](https://github.com/bdice) +- Pin `dask` and `distributed` for release ([#11822](https://github.com/NVIDIA/cudf/pull/11822)) [@galipremsagar](https://github.com/galipremsagar) +- Add examples for Nested JSON reader ([#11814](https://github.com/NVIDIA/cudf/pull/11814)) [@GregoryKimball](https://github.com/GregoryKimball) +- Support shuffle-based groupby aggregations in dask_cudf ([#11800](https://github.com/NVIDIA/cudf/pull/11800)) [@rjzamora](https://github.com/rjzamora) +- Update strings udf version updater script ([#11772](https://github.com/NVIDIA/cudf/pull/11772)) [@galipremsagar](https://github.com/galipremsagar) +- Remove `kwargs` in `read_csv` & `to_csv` ([#11762](https://github.com/NVIDIA/cudf/pull/11762)) [@galipremsagar](https://github.com/galipremsagar) +- Pass `dtype` param to avoid `pd.Series` warnings ([#11761](https://github.com/NVIDIA/cudf/pull/11761)) [@galipremsagar](https://github.com/galipremsagar) +- Enable `schema_element` & `keep_quotes` support in json reader ([#11746](https://github.com/NVIDIA/cudf/pull/11746)) [@galipremsagar](https://github.com/galipremsagar) +- Add ability to construct `ListColumn` when size is `None` ([#11745](https://github.com/NVIDIA/cudf/pull/11745)) [@galipremsagar](https://github.com/galipremsagar) +- Reduces memory requirements in JSON parser and adds bytes/s and peak memory usage to benchmarks ([#11732](https://github.com/NVIDIA/cudf/pull/11732)) [@elstehle](https://github.com/elstehle) +- Add missing copyright headers. ([#11712](https://github.com/NVIDIA/cudf/pull/11712)) [@bdice](https://github.com/bdice) +- Fix copyright check issues in pre-commit ([#11711](https://github.com/NVIDIA/cudf/pull/11711)) [@bdice](https://github.com/bdice) +- Include decimal in supported types for range window order-by columns ([#11710](https://github.com/NVIDIA/cudf/pull/11710)) [@mythrocks](https://github.com/mythrocks) +- Disable very large column gtest for contiguous-split ([#11706](https://github.com/NVIDIA/cudf/pull/11706)) [@davidwendt](https://github.com/davidwendt) +- Drop split_out=None test from groupby.agg ([#11704](https://github.com/NVIDIA/cudf/pull/11704)) [@wence-](https://github.com/wence-) +- Use CubinLinker for CUDA Minor Version Compatibility ([#11701](https://github.com/NVIDIA/cudf/pull/11701)) [@gmarkall](https://github.com/gmarkall) +- Add regex capture-group parameter to auto convert to non-capture groups ([#11695](https://github.com/NVIDIA/cudf/pull/11695)) [@davidwendt](https://github.com/davidwendt) +- Add a `__dataframe__` method to the protocol dataframe object ([#11692](https://github.com/NVIDIA/cudf/pull/11692)) [@rgommers](https://github.com/rgommers) +- Special-case multibyte_split for single-byte delimiter ([#11681](https://github.com/NVIDIA/cudf/pull/11681)) [@upsj](https://github.com/upsj) +- Remove isort exclusions ([#11680](https://github.com/NVIDIA/cudf/pull/11680)) [@bdice](https://github.com/bdice) +- Refactor CSV reader benchmarks with nvbench ([#11678](https://github.com/NVIDIA/cudf/pull/11678)) [@PointKernel](https://github.com/PointKernel) +- Check conda recipe headers with pre-commit ([#11669](https://github.com/NVIDIA/cudf/pull/11669)) [@bdice](https://github.com/bdice) +- Remove redundant style check for clang-format. ([#11668](https://github.com/NVIDIA/cudf/pull/11668)) [@bdice](https://github.com/bdice) +- Add support for `group_keys` in `groupby` ([#11659](https://github.com/NVIDIA/cudf/pull/11659)) [@galipremsagar](https://github.com/galipremsagar) +- Fix pandoc pinning. ([#11658](https://github.com/NVIDIA/cudf/pull/11658)) [@bdice](https://github.com/bdice) +- Revert removal of skip_rows / num_rows options from the Parquet reader. ([#11657](https://github.com/NVIDIA/cudf/pull/11657)) [@nvdbaranec](https://github.com/nvdbaranec) +- Update git metadata ([#11647](https://github.com/NVIDIA/cudf/pull/11647)) [@bdice](https://github.com/bdice) +- Call set_null_count on a returning column if null-count is known ([#11646](https://github.com/NVIDIA/cudf/pull/11646)) [@davidwendt](https://github.com/davidwendt) +- Fix some libcudf detail calls not passing the stream variable ([#11642](https://github.com/NVIDIA/cudf/pull/11642)) [@davidwendt](https://github.com/davidwendt) +- Update to mypy 0.971 ([#11640](https://github.com/NVIDIA/cudf/pull/11640)) [@wence-](https://github.com/wence-) +- Refactor strings strip functor to details header ([#11635](https://github.com/NVIDIA/cudf/pull/11635)) [@davidwendt](https://github.com/davidwendt) +- Fix incorrect `nullCount` in `get_json_object` ([#11633](https://github.com/NVIDIA/cudf/pull/11633)) [@trxcllnt](https://github.com/trxcllnt) +- Simplify `hostdevice_vector` ([#11631](https://github.com/NVIDIA/cudf/pull/11631)) [@upsj](https://github.com/upsj) +- Refactor parquet writer benchmarks with nvbench ([#11623](https://github.com/NVIDIA/cudf/pull/11623)) [@PointKernel](https://github.com/PointKernel) +- Rework contains_scalar to check nulls at runtime ([#11622](https://github.com/NVIDIA/cudf/pull/11622)) [@davidwendt](https://github.com/davidwendt) +- Fix incorrect memory resource used in rolling temp columns ([#11618](https://github.com/NVIDIA/cudf/pull/11618)) [@mythrocks](https://github.com/mythrocks) +- Upgrade `pandas` to `1.5` ([#11617](https://github.com/NVIDIA/cudf/pull/11617)) [@galipremsagar](https://github.com/galipremsagar) +- Move type-dispatcher calls from traits.hpp to traits.cpp ([#11616](https://github.com/NVIDIA/cudf/pull/11616)) [@davidwendt](https://github.com/davidwendt) +- Refactor parquet reader benchmarks with nvbench ([#11611](https://github.com/NVIDIA/cudf/pull/11611)) [@PointKernel](https://github.com/PointKernel) +- Forward-merge branch-22.08 to branch-22.10 ([#11608](https://github.com/NVIDIA/cudf/pull/11608)) [@bdice](https://github.com/bdice) +- Use stream in Java API. ([#11601](https://github.com/NVIDIA/cudf/pull/11601)) [@bdice](https://github.com/bdice) +- Refactors of public/detail APIs, CUDF_FUNC_RANGE, stream handling. ([#11600](https://github.com/NVIDIA/cudf/pull/11600)) [@bdice](https://github.com/bdice) +- Improve ORC writer benchmark with nvbench ([#11598](https://github.com/NVIDIA/cudf/pull/11598)) [@PointKernel](https://github.com/PointKernel) +- Tune multibyte_split kernel ([#11587](https://github.com/NVIDIA/cudf/pull/11587)) [@upsj](https://github.com/upsj) +- Move split_utils.cuh to strings/detail ([#11585](https://github.com/NVIDIA/cudf/pull/11585)) [@davidwendt](https://github.com/davidwendt) +- Fix warnings due to compiler regression with `if constexpr` ([#11581](https://github.com/NVIDIA/cudf/pull/11581)) [@ttnghia](https://github.com/ttnghia) +- Add full 24-bit dictionary support to Parquet writer ([#11580](https://github.com/NVIDIA/cudf/pull/11580)) [@etseidl](https://github.com/etseidl) +- Expose "explicit-comms" option in shuffle-based dask_cudf functions ([#11576](https://github.com/NVIDIA/cudf/pull/11576)) [@rjzamora](https://github.com/rjzamora) +- Move cudf::strings::findall_record to cudf::strings::findall ([#11575](https://github.com/NVIDIA/cudf/pull/11575)) [@davidwendt](https://github.com/davidwendt) +- Refactor dask_cudf groupby to use apply_concat_apply ([#11571](https://github.com/NVIDIA/cudf/pull/11571)) [@rjzamora](https://github.com/rjzamora) +- Add ability to write `list(struct)` columns as `map` type in orc writer ([#11568](https://github.com/NVIDIA/cudf/pull/11568)) [@galipremsagar](https://github.com/galipremsagar) +- Add byte_range to multibyte_split benchmark + NVBench refactor ([#11562](https://github.com/NVIDIA/cudf/pull/11562)) [@upsj](https://github.com/upsj) +- JNI support for writing binary columns in parquet ([#11556](https://github.com/NVIDIA/cudf/pull/11556)) [@revans2](https://github.com/revans2) +- Support additional dictionary bit widths in Parquet writer ([#11547](https://github.com/NVIDIA/cudf/pull/11547)) [@etseidl](https://github.com/etseidl) +- Refactor string/numeric conversion utilities ([#11545](https://github.com/NVIDIA/cudf/pull/11545)) [@davidwendt](https://github.com/davidwendt) +- Removing unnecessary asserts in parquet tests ([#11544](https://github.com/NVIDIA/cudf/pull/11544)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Clean up ORC reader benchmarks with NVBench ([#11543](https://github.com/NVIDIA/cudf/pull/11543)) [@PointKernel](https://github.com/PointKernel) +- Reuse MurmurHash3_32 in Parquet page data. ([#11528](https://github.com/NVIDIA/cudf/pull/11528)) [@bdice](https://github.com/bdice) +- Add hexadecimal value separators ([#11527](https://github.com/NVIDIA/cudf/pull/11527)) [@bdice](https://github.com/bdice) +- Deprecate `skiprows` and `num_rows` in `read_orc` ([#11522](https://github.com/NVIDIA/cudf/pull/11522)) [@galipremsagar](https://github.com/galipremsagar) +- Struct support for `NULL_EQUALS` binary operation ([#11520](https://github.com/NVIDIA/cudf/pull/11520)) [@rwlee](https://github.com/rwlee) +- Bump hadoop-common from 3.2.3 to 3.2.4 in /java ([#11516](https://github.com/NVIDIA/cudf/pull/11516)) [@dependabot[bot]](https://github.com/dependabot[bot]) +- Fix Feather test warning. ([#11511](https://github.com/NVIDIA/cudf/pull/11511)) [@bdice](https://github.com/bdice) +- copy_range ballot_syncs to have no execution dependency ([#11508](https://github.com/NVIDIA/cudf/pull/11508)) [@robertmaynard](https://github.com/robertmaynard) +- Upgrade to `arrow-9.x` ([#11507](https://github.com/NVIDIA/cudf/pull/11507)) [@galipremsagar](https://github.com/galipremsagar) +- Remove support for skip_rows / num_rows options in the parquet reader. ([#11503](https://github.com/NVIDIA/cudf/pull/11503)) [@nvdbaranec](https://github.com/nvdbaranec) +- Single-pass `multibyte_split` ([#11500](https://github.com/NVIDIA/cudf/pull/11500)) [@upsj](https://github.com/upsj) +- Sanitize percentile_approx() output for empty input ([#11498](https://github.com/NVIDIA/cudf/pull/11498)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Unpin `dask` and `distributed` for development ([#11492](https://github.com/NVIDIA/cudf/pull/11492)) [@galipremsagar](https://github.com/galipremsagar) +- Move SparkMurmurHash3_32 functor. ([#11489](https://github.com/NVIDIA/cudf/pull/11489)) [@bdice](https://github.com/bdice) +- Refactor group_nunique.cu to use nullate::DYNAMIC for reduce-by-key functor ([#11482](https://github.com/NVIDIA/cudf/pull/11482)) [@davidwendt](https://github.com/davidwendt) +- Drop support for `skiprows` and `num_rows` in `cudf.read_parquet` ([#11480](https://github.com/NVIDIA/cudf/pull/11480)) [@galipremsagar](https://github.com/galipremsagar) +- Add reduction `distinct_count` benchmark ([#11473](https://github.com/NVIDIA/cudf/pull/11473)) [@ttnghia](https://github.com/ttnghia) +- Add groupby `nunique` aggregation benchmark ([#11472](https://github.com/NVIDIA/cudf/pull/11472)) [@ttnghia](https://github.com/ttnghia) +- Disable Arrow S3 support by default. ([#11470](https://github.com/NVIDIA/cudf/pull/11470)) [@bdice](https://github.com/bdice) +- Add groupby `max` aggregation benchmark ([#11464](https://github.com/NVIDIA/cudf/pull/11464)) [@ttnghia](https://github.com/ttnghia) +- Extract Dremel encoding code from Parquet ([#11461](https://github.com/NVIDIA/cudf/pull/11461)) [@vyasr](https://github.com/vyasr) +- Add missing Thrust #includes. ([#11457](https://github.com/NVIDIA/cudf/pull/11457)) [@bdice](https://github.com/bdice) +- Make CMake hooks verbose ([#11456](https://github.com/NVIDIA/cudf/pull/11456)) [@vyasr](https://github.com/vyasr) +- Control Parquet page size through Python API ([#11454](https://github.com/NVIDIA/cudf/pull/11454)) [@etseidl](https://github.com/etseidl) +- Add control of Parquet column index creation to python ([#11453](https://github.com/NVIDIA/cudf/pull/11453)) [@etseidl](https://github.com/etseidl) +- Remove unused is_struct trait. ([#11450](https://github.com/NVIDIA/cudf/pull/11450)) [@bdice](https://github.com/bdice) +- Refactor the `Buffer` class ([#11447](https://github.com/NVIDIA/cudf/pull/11447)) [@madsbk](https://github.com/madsbk) +- Refactor pad_side and strip_type enums into side_type enum ([#11438](https://github.com/NVIDIA/cudf/pull/11438)) [@davidwendt](https://github.com/davidwendt) +- Update to Thrust 1.17.0 ([#11437](https://github.com/NVIDIA/cudf/pull/11437)) [@bdice](https://github.com/bdice) +- Add in JNI for parsing JSON data and getting the metadata back too. ([#11431](https://github.com/NVIDIA/cudf/pull/11431)) [@revans2](https://github.com/revans2) +- Convert byte_array_view to use std::byte ([#11424](https://github.com/NVIDIA/cudf/pull/11424)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Deprecate unflatten_nested_columns ([#11421](https://github.com/NVIDIA/cudf/pull/11421)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Remove HASH_SERIAL_MURMUR3 / serial32BitMurmurHash3 ([#11383](https://github.com/NVIDIA/cudf/pull/11383)) [@bdice](https://github.com/bdice) +- Add Spark list hashing Java tests ([#11379](https://github.com/NVIDIA/cudf/pull/11379)) [@bdice](https://github.com/bdice) +- Move cmake to the build section. ([#11376](https://github.com/NVIDIA/cudf/pull/11376)) [@vyasr](https://github.com/vyasr) +- Remove use of CUDA driver API calls from libcudf ([#11370](https://github.com/NVIDIA/cudf/pull/11370)) [@shwina](https://github.com/shwina) +- Add column constructor from device_uvector&& ([#11356](https://github.com/NVIDIA/cudf/pull/11356)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Remove unused custreamz thirdparty directory ([#11343](https://github.com/NVIDIA/cudf/pull/11343)) [@vyasr](https://github.com/vyasr) +- Update jni version to 22.10.0-SNAPSHOT ([#11338](https://github.com/NVIDIA/cudf/pull/11338)) [@pxLi](https://github.com/pxLi) +- Enable using upstream jitify2 ([#11287](https://github.com/NVIDIA/cudf/pull/11287)) [@shwina](https://github.com/shwina) +- Cache cudf.Scalar ([#11246](https://github.com/NVIDIA/cudf/pull/11246)) [@shwina](https://github.com/shwina) +- Remove deprecated Series.applymap. ([#11031](https://github.com/NVIDIA/cudf/pull/11031)) [@bdice](https://github.com/bdice) +- Remove deprecated expand parameter from str.findall. ([#11030](https://github.com/NVIDIA/cudf/pull/11030)) [@bdice](https://github.com/bdice) # cuDF 22.08.00 (17 Aug 2022) ## 🚨 Breaking Changes -- Remove legacy join APIs ([#11274](https://github.com/rapidsai/cudf/pull/11274)) [@vyasr](https://github.com/vyasr) -- Remove `lists::drop_list_duplicates` ([#11236](https://github.com/rapidsai/cudf/pull/11236)) [@ttnghia](https://github.com/ttnghia) -- Remove Index.replace API ([#11131](https://github.com/rapidsai/cudf/pull/11131)) [@vyasr](https://github.com/vyasr) -- Remove deprecated Index methods from Frame ([#11073](https://github.com/rapidsai/cudf/pull/11073)) [@vyasr](https://github.com/vyasr) -- Remove public API of cudf.merge_sorted. ([#11032](https://github.com/rapidsai/cudf/pull/11032)) [@bdice](https://github.com/bdice) -- Drop python `3.7` in code-base ([#11029](https://github.com/rapidsai/cudf/pull/11029)) [@galipremsagar](https://github.com/galipremsagar) -- Return empty dataframe when reading a Parquet file using empty `columns` option ([#11018](https://github.com/rapidsai/cudf/pull/11018)) [@vuule](https://github.com/vuule) -- Remove Arrow CUDA IPC code ([#10995](https://github.com/rapidsai/cudf/pull/10995)) [@shwina](https://github.com/shwina) -- Buffer: make `.ptr` read-only ([#10872](https://github.com/rapidsai/cudf/pull/10872)) [@madsbk](https://github.com/madsbk) +- Remove legacy join APIs ([#11274](https://github.com/NVIDIA/cudf/pull/11274)) [@vyasr](https://github.com/vyasr) +- Remove `lists::drop_list_duplicates` ([#11236](https://github.com/NVIDIA/cudf/pull/11236)) [@ttnghia](https://github.com/ttnghia) +- Remove Index.replace API ([#11131](https://github.com/NVIDIA/cudf/pull/11131)) [@vyasr](https://github.com/vyasr) +- Remove deprecated Index methods from Frame ([#11073](https://github.com/NVIDIA/cudf/pull/11073)) [@vyasr](https://github.com/vyasr) +- Remove public API of cudf.merge_sorted. ([#11032](https://github.com/NVIDIA/cudf/pull/11032)) [@bdice](https://github.com/bdice) +- Drop python `3.7` in code-base ([#11029](https://github.com/NVIDIA/cudf/pull/11029)) [@galipremsagar](https://github.com/galipremsagar) +- Return empty dataframe when reading a Parquet file using empty `columns` option ([#11018](https://github.com/NVIDIA/cudf/pull/11018)) [@vuule](https://github.com/vuule) +- Remove Arrow CUDA IPC code ([#10995](https://github.com/NVIDIA/cudf/pull/10995)) [@shwina](https://github.com/shwina) +- Buffer: make `.ptr` read-only ([#10872](https://github.com/NVIDIA/cudf/pull/10872)) [@madsbk](https://github.com/madsbk) ## 🐛 Bug Fixes -- Fix `distributed` error related to `loop_in_thread` ([#11428](https://github.com/rapidsai/cudf/pull/11428)) [@galipremsagar](https://github.com/galipremsagar) -- Relax arrow pinning to just 8.x and remove cuda build dependency from cudf recipe ([#11412](https://github.com/rapidsai/cudf/pull/11412)) [@kkraus14](https://github.com/kkraus14) -- Revert "Allow CuPy 11" ([#11409](https://github.com/rapidsai/cudf/pull/11409)) [@jakirkham](https://github.com/jakirkham) -- Fix `moto` timeouts ([#11369](https://github.com/rapidsai/cudf/pull/11369)) [@galipremsagar](https://github.com/galipremsagar) -- Set `+/-infinity` as the `identity` values for floating-point numbers in device operators `min` and `max` ([#11357](https://github.com/rapidsai/cudf/pull/11357)) [@ttnghia](https://github.com/ttnghia) -- Fix memory_usage() for `ListSeries` ([#11355](https://github.com/rapidsai/cudf/pull/11355)) [@thomcom](https://github.com/thomcom) -- Fix constructing Column from column_view with expired mask ([#11354](https://github.com/rapidsai/cudf/pull/11354)) [@shwina](https://github.com/shwina) -- Handle parquet corner case: Columns with more rows than are in the row group. ([#11353](https://github.com/rapidsai/cudf/pull/11353)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix `DatetimeIndex` & `TimedeltaIndex` constructors ([#11342](https://github.com/rapidsai/cudf/pull/11342)) [@galipremsagar](https://github.com/galipremsagar) -- Fix unsigned-compare compile warning in IntPow binops ([#11339](https://github.com/rapidsai/cudf/pull/11339)) [@davidwendt](https://github.com/davidwendt) -- Fix performance issue and add a new code path to `cudf::detail::contains` ([#11330](https://github.com/rapidsai/cudf/pull/11330)) [@ttnghia](https://github.com/ttnghia) -- Pin `pytorch` to temporarily unblock from `libcupti` errors ([#11289](https://github.com/rapidsai/cudf/pull/11289)) [@galipremsagar](https://github.com/galipremsagar) -- Workaround for nvcomp zstd overwriting blocks for orc due to underestimate of sizes ([#11288](https://github.com/rapidsai/cudf/pull/11288)) [@jbrennan333](https://github.com/jbrennan333) -- Fix inconsistency when hashing two tables in `cudf::detail::contains` ([#11284](https://github.com/rapidsai/cudf/pull/11284)) [@ttnghia](https://github.com/ttnghia) -- Fix issue related to numpy array and `category` dtype ([#11282](https://github.com/rapidsai/cudf/pull/11282)) [@galipremsagar](https://github.com/galipremsagar) -- Add NotImplementedError when on is specified in DataFrame.join. ([#11275](https://github.com/rapidsai/cudf/pull/11275)) [@vyasr](https://github.com/vyasr) -- Fix invalid allocate_like() and empty_like() tests. ([#11268](https://github.com/rapidsai/cudf/pull/11268)) [@nvdbaranec](https://github.com/nvdbaranec) -- Returns DataFrame When Concatenating Along Axis 1 ([#11263](https://github.com/rapidsai/cudf/pull/11263)) [@isVoid](https://github.com/isVoid) -- Fix compile error due to missing header ([#11257](https://github.com/rapidsai/cudf/pull/11257)) [@ttnghia](https://github.com/ttnghia) -- Fix a memory aliasing/crash issue in scatter for lists. ([#11254](https://github.com/rapidsai/cudf/pull/11254)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix `tests/rolling/empty_input_test` ([#11238](https://github.com/rapidsai/cudf/pull/11238)) [@ttnghia](https://github.com/ttnghia) -- Fix const qualifier when using `host_span<bitmask_type const*>` ([#11220](https://github.com/rapidsai/cudf/pull/11220)) [@ttnghia](https://github.com/ttnghia) -- Avoid using `nvcompBatchedDeflateDecompressGetTempSizeEx` in cuIO ([#11213](https://github.com/rapidsai/cudf/pull/11213)) [@vuule](https://github.com/vuule) -- Generate benchmark data with correct run length regardless of cardinality ([#11205](https://github.com/rapidsai/cudf/pull/11205)) [@vuule](https://github.com/vuule) -- Fix cumulative count index behavior ([#11188](https://github.com/rapidsai/cudf/pull/11188)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix assertion in dask_cudf test_struct_explode ([#11170](https://github.com/rapidsai/cudf/pull/11170)) [@rjzamora](https://github.com/rjzamora) -- Provides a method for the user to remove the hook and re-register the hook in a custom shutdown hook manager ([#11161](https://github.com/rapidsai/cudf/pull/11161)) [@res-life](https://github.com/res-life) -- Fix compatibility issues with pandas 1.4.3 ([#11152](https://github.com/rapidsai/cudf/pull/11152)) [@vyasr](https://github.com/vyasr) -- Ensure cuco export set is installed in cmake build ([#11147](https://github.com/rapidsai/cudf/pull/11147)) [@jlowe](https://github.com/jlowe) -- Avoid redundant deepcopy in `cudf.from_pandas` ([#11142](https://github.com/rapidsai/cudf/pull/11142)) [@galipremsagar](https://github.com/galipremsagar) -- Fix compile error due to missing header ([#11126](https://github.com/rapidsai/cudf/pull/11126)) [@ttnghia](https://github.com/ttnghia) -- Fix `__cuda_array_interface__` failures ([#11113](https://github.com/rapidsai/cudf/pull/11113)) [@galipremsagar](https://github.com/galipremsagar) -- Support octal and hex within regex character class pattern ([#11112](https://github.com/rapidsai/cudf/pull/11112)) [@davidwendt](https://github.com/davidwendt) -- Fix split_re matching logic for word boundaries ([#11106](https://github.com/rapidsai/cudf/pull/11106)) [@davidwendt](https://github.com/davidwendt) -- Handle multiple files metadata in `read_parquet` ([#11105](https://github.com/rapidsai/cudf/pull/11105)) [@galipremsagar](https://github.com/galipremsagar) -- Fix index alignment for Series objects with repeated index ([#11103](https://github.com/rapidsai/cudf/pull/11103)) [@shwina](https://github.com/shwina) -- FindcuFile now searches in the current CUDA Toolkit location ([#11101](https://github.com/rapidsai/cudf/pull/11101)) [@robertmaynard](https://github.com/robertmaynard) -- Fix regex word boundary logic to include underline ([#11099](https://github.com/rapidsai/cudf/pull/11099)) [@davidwendt](https://github.com/davidwendt) -- Exclude CudaFatalTest when selecting all Java tests ([#11083](https://github.com/rapidsai/cudf/pull/11083)) [@jlowe](https://github.com/jlowe) -- Fix duplicate `cudatoolkit` pinning issue ([#11070](https://github.com/rapidsai/cudf/pull/11070)) [@galipremsagar](https://github.com/galipremsagar) -- Maintain the input index in the result of a groupby-transform ([#11068](https://github.com/rapidsai/cudf/pull/11068)) [@shwina](https://github.com/shwina) -- Fix bug with row count comparison for expect_columns_equivalent(). ([#11059](https://github.com/rapidsai/cudf/pull/11059)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix BPE uninitialized size value for null and empty input strings ([#11054](https://github.com/rapidsai/cudf/pull/11054)) [@davidwendt](https://github.com/davidwendt) -- Include missing header for usage of `get_current_device_resource()` ([#11047](https://github.com/rapidsai/cudf/pull/11047)) [@AtlantaPepsi](https://github.com/AtlantaPepsi) -- Fix warn_unused_result error in parquet test ([#11026](https://github.com/rapidsai/cudf/pull/11026)) [@karthikeyann](https://github.com/karthikeyann) -- Return empty dataframe when reading a Parquet file using empty `columns` option ([#11018](https://github.com/rapidsai/cudf/pull/11018)) [@vuule](https://github.com/vuule) -- Fix small error in page row count limiting ([#10991](https://github.com/rapidsai/cudf/pull/10991)) [@etseidl](https://github.com/etseidl) -- Fix a row index entry error in ORC writer issue ([#10989](https://github.com/rapidsai/cudf/pull/10989)) [@vuule](https://github.com/vuule) -- Fix grouped covariance to require both values to be convertible to double. ([#10891](https://github.com/rapidsai/cudf/pull/10891)) [@bdice](https://github.com/bdice) +- Fix `distributed` error related to `loop_in_thread` ([#11428](https://github.com/NVIDIA/cudf/pull/11428)) [@galipremsagar](https://github.com/galipremsagar) +- Relax arrow pinning to just 8.x and remove cuda build dependency from cudf recipe ([#11412](https://github.com/NVIDIA/cudf/pull/11412)) [@kkraus14](https://github.com/kkraus14) +- Revert "Allow CuPy 11" ([#11409](https://github.com/NVIDIA/cudf/pull/11409)) [@jakirkham](https://github.com/jakirkham) +- Fix `moto` timeouts ([#11369](https://github.com/NVIDIA/cudf/pull/11369)) [@galipremsagar](https://github.com/galipremsagar) +- Set `+/-infinity` as the `identity` values for floating-point numbers in device operators `min` and `max` ([#11357](https://github.com/NVIDIA/cudf/pull/11357)) [@ttnghia](https://github.com/ttnghia) +- Fix memory_usage() for `ListSeries` ([#11355](https://github.com/NVIDIA/cudf/pull/11355)) [@thomcom](https://github.com/thomcom) +- Fix constructing Column from column_view with expired mask ([#11354](https://github.com/NVIDIA/cudf/pull/11354)) [@shwina](https://github.com/shwina) +- Handle parquet corner case: Columns with more rows than are in the row group. ([#11353](https://github.com/NVIDIA/cudf/pull/11353)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix `DatetimeIndex` & `TimedeltaIndex` constructors ([#11342](https://github.com/NVIDIA/cudf/pull/11342)) [@galipremsagar](https://github.com/galipremsagar) +- Fix unsigned-compare compile warning in IntPow binops ([#11339](https://github.com/NVIDIA/cudf/pull/11339)) [@davidwendt](https://github.com/davidwendt) +- Fix performance issue and add a new code path to `cudf::detail::contains` ([#11330](https://github.com/NVIDIA/cudf/pull/11330)) [@ttnghia](https://github.com/ttnghia) +- Pin `pytorch` to temporarily unblock from `libcupti` errors ([#11289](https://github.com/NVIDIA/cudf/pull/11289)) [@galipremsagar](https://github.com/galipremsagar) +- Workaround for nvcomp zstd overwriting blocks for orc due to underestimate of sizes ([#11288](https://github.com/NVIDIA/cudf/pull/11288)) [@jbrennan333](https://github.com/jbrennan333) +- Fix inconsistency when hashing two tables in `cudf::detail::contains` ([#11284](https://github.com/NVIDIA/cudf/pull/11284)) [@ttnghia](https://github.com/ttnghia) +- Fix issue related to numpy array and `category` dtype ([#11282](https://github.com/NVIDIA/cudf/pull/11282)) [@galipremsagar](https://github.com/galipremsagar) +- Add NotImplementedError when on is specified in DataFrame.join. ([#11275](https://github.com/NVIDIA/cudf/pull/11275)) [@vyasr](https://github.com/vyasr) +- Fix invalid allocate_like() and empty_like() tests. ([#11268](https://github.com/NVIDIA/cudf/pull/11268)) [@nvdbaranec](https://github.com/nvdbaranec) +- Returns DataFrame When Concatenating Along Axis 1 ([#11263](https://github.com/NVIDIA/cudf/pull/11263)) [@isVoid](https://github.com/isVoid) +- Fix compile error due to missing header ([#11257](https://github.com/NVIDIA/cudf/pull/11257)) [@ttnghia](https://github.com/ttnghia) +- Fix a memory aliasing/crash issue in scatter for lists. ([#11254](https://github.com/NVIDIA/cudf/pull/11254)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix `tests/rolling/empty_input_test` ([#11238](https://github.com/NVIDIA/cudf/pull/11238)) [@ttnghia](https://github.com/ttnghia) +- Fix const qualifier when using `host_span<bitmask_type const*>` ([#11220](https://github.com/NVIDIA/cudf/pull/11220)) [@ttnghia](https://github.com/ttnghia) +- Avoid using `nvcompBatchedDeflateDecompressGetTempSizeEx` in cuIO ([#11213](https://github.com/NVIDIA/cudf/pull/11213)) [@vuule](https://github.com/vuule) +- Generate benchmark data with correct run length regardless of cardinality ([#11205](https://github.com/NVIDIA/cudf/pull/11205)) [@vuule](https://github.com/vuule) +- Fix cumulative count index behavior ([#11188](https://github.com/NVIDIA/cudf/pull/11188)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix assertion in dask_cudf test_struct_explode ([#11170](https://github.com/NVIDIA/cudf/pull/11170)) [@rjzamora](https://github.com/rjzamora) +- Provides a method for the user to remove the hook and re-register the hook in a custom shutdown hook manager ([#11161](https://github.com/NVIDIA/cudf/pull/11161)) [@res-life](https://github.com/res-life) +- Fix compatibility issues with pandas 1.4.3 ([#11152](https://github.com/NVIDIA/cudf/pull/11152)) [@vyasr](https://github.com/vyasr) +- Ensure cuco export set is installed in cmake build ([#11147](https://github.com/NVIDIA/cudf/pull/11147)) [@jlowe](https://github.com/jlowe) +- Avoid redundant deepcopy in `cudf.from_pandas` ([#11142](https://github.com/NVIDIA/cudf/pull/11142)) [@galipremsagar](https://github.com/galipremsagar) +- Fix compile error due to missing header ([#11126](https://github.com/NVIDIA/cudf/pull/11126)) [@ttnghia](https://github.com/ttnghia) +- Fix `__cuda_array_interface__` failures ([#11113](https://github.com/NVIDIA/cudf/pull/11113)) [@galipremsagar](https://github.com/galipremsagar) +- Support octal and hex within regex character class pattern ([#11112](https://github.com/NVIDIA/cudf/pull/11112)) [@davidwendt](https://github.com/davidwendt) +- Fix split_re matching logic for word boundaries ([#11106](https://github.com/NVIDIA/cudf/pull/11106)) [@davidwendt](https://github.com/davidwendt) +- Handle multiple files metadata in `read_parquet` ([#11105](https://github.com/NVIDIA/cudf/pull/11105)) [@galipremsagar](https://github.com/galipremsagar) +- Fix index alignment for Series objects with repeated index ([#11103](https://github.com/NVIDIA/cudf/pull/11103)) [@shwina](https://github.com/shwina) +- FindcuFile now searches in the current CUDA Toolkit location ([#11101](https://github.com/NVIDIA/cudf/pull/11101)) [@robertmaynard](https://github.com/robertmaynard) +- Fix regex word boundary logic to include underline ([#11099](https://github.com/NVIDIA/cudf/pull/11099)) [@davidwendt](https://github.com/davidwendt) +- Exclude CudaFatalTest when selecting all Java tests ([#11083](https://github.com/NVIDIA/cudf/pull/11083)) [@jlowe](https://github.com/jlowe) +- Fix duplicate `cudatoolkit` pinning issue ([#11070](https://github.com/NVIDIA/cudf/pull/11070)) [@galipremsagar](https://github.com/galipremsagar) +- Maintain the input index in the result of a groupby-transform ([#11068](https://github.com/NVIDIA/cudf/pull/11068)) [@shwina](https://github.com/shwina) +- Fix bug with row count comparison for expect_columns_equivalent(). ([#11059](https://github.com/NVIDIA/cudf/pull/11059)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix BPE uninitialized size value for null and empty input strings ([#11054](https://github.com/NVIDIA/cudf/pull/11054)) [@davidwendt](https://github.com/davidwendt) +- Include missing header for usage of `get_current_device_resource()` ([#11047](https://github.com/NVIDIA/cudf/pull/11047)) [@AtlantaPepsi](https://github.com/AtlantaPepsi) +- Fix warn_unused_result error in parquet test ([#11026](https://github.com/NVIDIA/cudf/pull/11026)) [@karthikeyann](https://github.com/karthikeyann) +- Return empty dataframe when reading a Parquet file using empty `columns` option ([#11018](https://github.com/NVIDIA/cudf/pull/11018)) [@vuule](https://github.com/vuule) +- Fix small error in page row count limiting ([#10991](https://github.com/NVIDIA/cudf/pull/10991)) [@etseidl](https://github.com/etseidl) +- Fix a row index entry error in ORC writer issue ([#10989](https://github.com/NVIDIA/cudf/pull/10989)) [@vuule](https://github.com/vuule) +- Fix grouped covariance to require both values to be convertible to double. ([#10891](https://github.com/NVIDIA/cudf/pull/10891)) [@bdice](https://github.com/bdice) ## 📖 Documentation -- Fix issues with day & night modes in python docs ([#11400](https://github.com/rapidsai/cudf/pull/11400)) [@galipremsagar](https://github.com/galipremsagar) -- Update missing data handling APIs in docs ([#11345](https://github.com/rapidsai/cudf/pull/11345)) [@galipremsagar](https://github.com/galipremsagar) -- Add lists filtering APIs to doxygen group. ([#11336](https://github.com/rapidsai/cudf/pull/11336)) [@bdice](https://github.com/bdice) -- Remove unused import in README sample ([#11318](https://github.com/rapidsai/cudf/pull/11318)) [@vyasr](https://github.com/vyasr) -- Note null behavior in `where` docs ([#11276](https://github.com/rapidsai/cudf/pull/11276)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Update docstring for spans in `get_row_data_range` ([#11271](https://github.com/rapidsai/cudf/pull/11271)) [@vyasr](https://github.com/vyasr) -- Update nvCOMP integration table ([#11231](https://github.com/rapidsai/cudf/pull/11231)) [@vuule](https://github.com/vuule) -- Add dev docs for documentation writing ([#11217](https://github.com/rapidsai/cudf/pull/11217)) [@vyasr](https://github.com/vyasr) -- Documentation fix for concatenate ([#11187](https://github.com/rapidsai/cudf/pull/11187)) [@dagardner-nv](https://github.com/dagardner-nv) -- Fix unresolved links in markdown ([#11173](https://github.com/rapidsai/cudf/pull/11173)) [@karthikeyann](https://github.com/karthikeyann) -- Fix cudf version in README.md install commands ([#11164](https://github.com/rapidsai/cudf/pull/11164)) [@jvanstraten](https://github.com/jvanstraten) -- Switch `language` from `None` to `"en"` in docs build ([#11133](https://github.com/rapidsai/cudf/pull/11133)) [@galipremsagar](https://github.com/galipremsagar) -- Remove docs mentioning scalar_view since no such class exists. ([#11132](https://github.com/rapidsai/cudf/pull/11132)) [@bdice](https://github.com/bdice) -- Add docstring entry for `DataFrame.value_counts` ([#11039](https://github.com/rapidsai/cudf/pull/11039)) [@galipremsagar](https://github.com/galipremsagar) -- Add docs to rolling var, std, count. ([#11035](https://github.com/rapidsai/cudf/pull/11035)) [@bdice](https://github.com/bdice) -- Fix docs for Numba UDFs. ([#11020](https://github.com/rapidsai/cudf/pull/11020)) [@bdice](https://github.com/bdice) -- Replace column comparison utilities functions with macros ([#11007](https://github.com/rapidsai/cudf/pull/11007)) [@karthikeyann](https://github.com/karthikeyann) -- Fix Doxygen warnings in multiple headers files ([#11003](https://github.com/rapidsai/cudf/pull/11003)) [@karthikeyann](https://github.com/karthikeyann) -- Fix doxygen warnings in utilities/ headers ([#10974](https://github.com/rapidsai/cudf/pull/10974)) [@karthikeyann](https://github.com/karthikeyann) -- Fix Doxygen warnings in table header files ([#10964](https://github.com/rapidsai/cudf/pull/10964)) [@karthikeyann](https://github.com/karthikeyann) -- Fix Doxygen warnings in column header files ([#10963](https://github.com/rapidsai/cudf/pull/10963)) [@karthikeyann](https://github.com/karthikeyann) -- Fix Doxygen warnings in strings / header files ([#10937](https://github.com/rapidsai/cudf/pull/10937)) [@karthikeyann](https://github.com/karthikeyann) -- Generate Doxygen Tag File for Libcudf ([#10932](https://github.com/rapidsai/cudf/pull/10932)) [@isVoid](https://github.com/isVoid) -- Fix doxygen warnings in structs, lists headers ([#10923](https://github.com/rapidsai/cudf/pull/10923)) [@karthikeyann](https://github.com/karthikeyann) -- Fix doxygen warnings in fixed_point.hpp ([#10922](https://github.com/rapidsai/cudf/pull/10922)) [@karthikeyann](https://github.com/karthikeyann) -- Fix doxygen warnings in ast/, rolling, tdigest/, wrappers/, dictionary/ headers ([#10921](https://github.com/rapidsai/cudf/pull/10921)) [@karthikeyann](https://github.com/karthikeyann) -- fix doxygen warnings in cudf/io/types.hpp, other header files ([#10913](https://github.com/rapidsai/cudf/pull/10913)) [@karthikeyann](https://github.com/karthikeyann) -- fix doxygen warnings in cudf/io/ avro, csv, json, orc, parquet header files ([#10912](https://github.com/rapidsai/cudf/pull/10912)) [@karthikeyann](https://github.com/karthikeyann) -- Fix doxygen warnings in cudf/*.hpp ([#10896](https://github.com/rapidsai/cudf/pull/10896)) [@karthikeyann](https://github.com/karthikeyann) -- Add missing documentation in aggregation.hpp ([#10887](https://github.com/rapidsai/cudf/pull/10887)) [@karthikeyann](https://github.com/karthikeyann) -- Revise PR template. ([#10774](https://github.com/rapidsai/cudf/pull/10774)) [@bdice](https://github.com/bdice) +- Fix issues with day & night modes in python docs ([#11400](https://github.com/NVIDIA/cudf/pull/11400)) [@galipremsagar](https://github.com/galipremsagar) +- Update missing data handling APIs in docs ([#11345](https://github.com/NVIDIA/cudf/pull/11345)) [@galipremsagar](https://github.com/galipremsagar) +- Add lists filtering APIs to doxygen group. ([#11336](https://github.com/NVIDIA/cudf/pull/11336)) [@bdice](https://github.com/bdice) +- Remove unused import in README sample ([#11318](https://github.com/NVIDIA/cudf/pull/11318)) [@vyasr](https://github.com/vyasr) +- Note null behavior in `where` docs ([#11276](https://github.com/NVIDIA/cudf/pull/11276)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Update docstring for spans in `get_row_data_range` ([#11271](https://github.com/NVIDIA/cudf/pull/11271)) [@vyasr](https://github.com/vyasr) +- Update nvCOMP integration table ([#11231](https://github.com/NVIDIA/cudf/pull/11231)) [@vuule](https://github.com/vuule) +- Add dev docs for documentation writing ([#11217](https://github.com/NVIDIA/cudf/pull/11217)) [@vyasr](https://github.com/vyasr) +- Documentation fix for concatenate ([#11187](https://github.com/NVIDIA/cudf/pull/11187)) [@dagardner-nv](https://github.com/dagardner-nv) +- Fix unresolved links in markdown ([#11173](https://github.com/NVIDIA/cudf/pull/11173)) [@karthikeyann](https://github.com/karthikeyann) +- Fix cudf version in README.md install commands ([#11164](https://github.com/NVIDIA/cudf/pull/11164)) [@jvanstraten](https://github.com/jvanstraten) +- Switch `language` from `None` to `"en"` in docs build ([#11133](https://github.com/NVIDIA/cudf/pull/11133)) [@galipremsagar](https://github.com/galipremsagar) +- Remove docs mentioning scalar_view since no such class exists. ([#11132](https://github.com/NVIDIA/cudf/pull/11132)) [@bdice](https://github.com/bdice) +- Add docstring entry for `DataFrame.value_counts` ([#11039](https://github.com/NVIDIA/cudf/pull/11039)) [@galipremsagar](https://github.com/galipremsagar) +- Add docs to rolling var, std, count. ([#11035](https://github.com/NVIDIA/cudf/pull/11035)) [@bdice](https://github.com/bdice) +- Fix docs for Numba UDFs. ([#11020](https://github.com/NVIDIA/cudf/pull/11020)) [@bdice](https://github.com/bdice) +- Replace column comparison utilities functions with macros ([#11007](https://github.com/NVIDIA/cudf/pull/11007)) [@karthikeyann](https://github.com/karthikeyann) +- Fix Doxygen warnings in multiple headers files ([#11003](https://github.com/NVIDIA/cudf/pull/11003)) [@karthikeyann](https://github.com/karthikeyann) +- Fix doxygen warnings in utilities/ headers ([#10974](https://github.com/NVIDIA/cudf/pull/10974)) [@karthikeyann](https://github.com/karthikeyann) +- Fix Doxygen warnings in table header files ([#10964](https://github.com/NVIDIA/cudf/pull/10964)) [@karthikeyann](https://github.com/karthikeyann) +- Fix Doxygen warnings in column header files ([#10963](https://github.com/NVIDIA/cudf/pull/10963)) [@karthikeyann](https://github.com/karthikeyann) +- Fix Doxygen warnings in strings / header files ([#10937](https://github.com/NVIDIA/cudf/pull/10937)) [@karthikeyann](https://github.com/karthikeyann) +- Generate Doxygen Tag File for Libcudf ([#10932](https://github.com/NVIDIA/cudf/pull/10932)) [@isVoid](https://github.com/isVoid) +- Fix doxygen warnings in structs, lists headers ([#10923](https://github.com/NVIDIA/cudf/pull/10923)) [@karthikeyann](https://github.com/karthikeyann) +- Fix doxygen warnings in fixed_point.hpp ([#10922](https://github.com/NVIDIA/cudf/pull/10922)) [@karthikeyann](https://github.com/karthikeyann) +- Fix doxygen warnings in ast/, rolling, tdigest/, wrappers/, dictionary/ headers ([#10921](https://github.com/NVIDIA/cudf/pull/10921)) [@karthikeyann](https://github.com/karthikeyann) +- fix doxygen warnings in cudf/io/types.hpp, other header files ([#10913](https://github.com/NVIDIA/cudf/pull/10913)) [@karthikeyann](https://github.com/karthikeyann) +- fix doxygen warnings in cudf/io/ avro, csv, json, orc, parquet header files ([#10912](https://github.com/NVIDIA/cudf/pull/10912)) [@karthikeyann](https://github.com/karthikeyann) +- Fix doxygen warnings in cudf/*.hpp ([#10896](https://github.com/NVIDIA/cudf/pull/10896)) [@karthikeyann](https://github.com/karthikeyann) +- Add missing documentation in aggregation.hpp ([#10887](https://github.com/NVIDIA/cudf/pull/10887)) [@karthikeyann](https://github.com/karthikeyann) +- Revise PR template. ([#10774](https://github.com/NVIDIA/cudf/pull/10774)) [@bdice](https://github.com/bdice) ## 🚀 New Features -- Change cmake to allow controlling Arrow version via cmake variable ([#11429](https://github.com/rapidsai/cudf/pull/11429)) [@kkraus14](https://github.com/kkraus14) -- Adding support for list<int8> columns to be written as byte arrays in parquet ([#11328](https://github.com/rapidsai/cudf/pull/11328)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Adding byte array view structure ([#11322](https://github.com/rapidsai/cudf/pull/11322)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Adding byte_array statistics ([#11303](https://github.com/rapidsai/cudf/pull/11303)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add column indexes to Parquet writer ([#11302](https://github.com/rapidsai/cudf/pull/11302)) [@etseidl](https://github.com/etseidl) -- Provide an Option for Default Integer and Floating Bitwidth ([#11272](https://github.com/rapidsai/cudf/pull/11272)) [@isVoid](https://github.com/isVoid) -- FST benchmark ([#11243](https://github.com/rapidsai/cudf/pull/11243)) [@karthikeyann](https://github.com/karthikeyann) -- Adds the Finite-State Transducer algorithm ([#11242](https://github.com/rapidsai/cudf/pull/11242)) [@elstehle](https://github.com/elstehle) -- Refactor `collect_set` to use `cudf::distinct` and `cudf::lists::distinct` ([#11228](https://github.com/rapidsai/cudf/pull/11228)) [@ttnghia](https://github.com/ttnghia) -- Treat zstd as stable in nvcomp releases 2.3.2 and later ([#11226](https://github.com/rapidsai/cudf/pull/11226)) [@jbrennan333](https://github.com/jbrennan333) -- Add 24 bit dictionary support to Parquet writer ([#11216](https://github.com/rapidsai/cudf/pull/11216)) [@devavret](https://github.com/devavret) -- Enable positive group indices for extractAllRecord on JNI ([#11215](https://github.com/rapidsai/cudf/pull/11215)) [@anthony-chang](https://github.com/anthony-chang) -- JNI bindings for NTH_ELEMENT window aggregation ([#11201](https://github.com/rapidsai/cudf/pull/11201)) [@mythrocks](https://github.com/mythrocks) -- Add JNI bindings for extractAllRecord ([#11196](https://github.com/rapidsai/cudf/pull/11196)) [@anthony-chang](https://github.com/anthony-chang) -- Add `cudf.options` ([#11193](https://github.com/rapidsai/cudf/pull/11193)) [@isVoid](https://github.com/isVoid) -- Add thrift support for parquet column and offset indexes ([#11178](https://github.com/rapidsai/cudf/pull/11178)) [@etseidl](https://github.com/etseidl) -- Adding binary read/write as options for parquet ([#11160](https://github.com/rapidsai/cudf/pull/11160)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Support `nth_element` for window functions ([#11158](https://github.com/rapidsai/cudf/pull/11158)) [@mythrocks](https://github.com/mythrocks) -- Implement `lists::distinct` and `cudf::detail::stable_distinct` ([#11149](https://github.com/rapidsai/cudf/pull/11149)) [@ttnghia](https://github.com/ttnghia) -- Implement Groupby pct_change ([#11144](https://github.com/rapidsai/cudf/pull/11144)) [@skirui-source](https://github.com/skirui-source) -- Add JNI for set operations ([#11143](https://github.com/rapidsai/cudf/pull/11143)) [@ttnghia](https://github.com/ttnghia) -- Remove deprecated PER_THREAD_DEFAULT_STREAM ([#11134](https://github.com/rapidsai/cudf/pull/11134)) [@jbrennan333](https://github.com/jbrennan333) -- Added a Java method to check the existence of a list of keys in a map ([#11128](https://github.com/rapidsai/cudf/pull/11128)) [@razajafri](https://github.com/razajafri) -- Feature/python benchmarking ([#11125](https://github.com/rapidsai/cudf/pull/11125)) [@vyasr](https://github.com/vyasr) -- Support `nan_equality` in `cudf::distinct` ([#11118](https://github.com/rapidsai/cudf/pull/11118)) [@ttnghia](https://github.com/ttnghia) -- Added JNI for getMapValueForKeys ([#11104](https://github.com/rapidsai/cudf/pull/11104)) [@razajafri](https://github.com/razajafri) -- Refactor `semi_anti_join` ([#11100](https://github.com/rapidsai/cudf/pull/11100)) [@ttnghia](https://github.com/ttnghia) -- Replace remaining instances of rmm::cuda_stream_default with cudf::default_stream_value ([#11082](https://github.com/rapidsai/cudf/pull/11082)) [@jbrennan333](https://github.com/jbrennan333) -- Adds the Logical Stack algorithm ([#11078](https://github.com/rapidsai/cudf/pull/11078)) [@elstehle](https://github.com/elstehle) -- Add doxygen-check pre-commit hook ([#11076](https://github.com/rapidsai/cudf/pull/11076)) [@karthikeyann](https://github.com/karthikeyann) -- Use new nvCOMP API to optimize the decompression temp memory size ([#11064](https://github.com/rapidsai/cudf/pull/11064)) [@vuule](https://github.com/vuule) -- Add Doxygen CI check ([#11057](https://github.com/rapidsai/cudf/pull/11057)) [@karthikeyann](https://github.com/karthikeyann) -- Support `duplicate_keep_option` in `cudf::distinct` ([#11052](https://github.com/rapidsai/cudf/pull/11052)) [@ttnghia](https://github.com/ttnghia) -- Support set operations ([#11043](https://github.com/rapidsai/cudf/pull/11043)) [@ttnghia](https://github.com/ttnghia) -- Support for ZLIB compression in ORC writer ([#11036](https://github.com/rapidsai/cudf/pull/11036)) [@vuule](https://github.com/vuule) -- Adding feature swaplevels ([#11027](https://github.com/rapidsai/cudf/pull/11027)) [@VamsiTallam95](https://github.com/VamsiTallam95) -- Use nvCOMP for ZLIB decompression in ORC reader ([#11024](https://github.com/rapidsai/cudf/pull/11024)) [@vuule](https://github.com/vuule) -- Function for bfill, ffill #9591 ([#11022](https://github.com/rapidsai/cudf/pull/11022)) [@Sreekiran096](https://github.com/Sreekiran096) -- Generate group offsets from element labels ([#11017](https://github.com/rapidsai/cudf/pull/11017)) [@ttnghia](https://github.com/ttnghia) -- Feature axes ([#10979](https://github.com/rapidsai/cudf/pull/10979)) [@VamsiTallam95](https://github.com/VamsiTallam95) -- Generate group labels from offsets ([#10945](https://github.com/rapidsai/cudf/pull/10945)) [@ttnghia](https://github.com/ttnghia) -- Add missing cuIO benchmark coverage for duration types ([#10933](https://github.com/rapidsai/cudf/pull/10933)) [@vuule](https://github.com/vuule) -- Dask-cuDF cumulative groupby ops ([#10889](https://github.com/rapidsai/cudf/pull/10889)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Reindex Improvements ([#10815](https://github.com/rapidsai/cudf/pull/10815)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Implement value_counts for DataFrame ([#10813](https://github.com/rapidsai/cudf/pull/10813)) [@martinfalisse](https://github.com/martinfalisse) +- Change cmake to allow controlling Arrow version via cmake variable ([#11429](https://github.com/NVIDIA/cudf/pull/11429)) [@kkraus14](https://github.com/kkraus14) +- Adding support for list<int8> columns to be written as byte arrays in parquet ([#11328](https://github.com/NVIDIA/cudf/pull/11328)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Adding byte array view structure ([#11322](https://github.com/NVIDIA/cudf/pull/11322)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Adding byte_array statistics ([#11303](https://github.com/NVIDIA/cudf/pull/11303)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add column indexes to Parquet writer ([#11302](https://github.com/NVIDIA/cudf/pull/11302)) [@etseidl](https://github.com/etseidl) +- Provide an Option for Default Integer and Floating Bitwidth ([#11272](https://github.com/NVIDIA/cudf/pull/11272)) [@isVoid](https://github.com/isVoid) +- FST benchmark ([#11243](https://github.com/NVIDIA/cudf/pull/11243)) [@karthikeyann](https://github.com/karthikeyann) +- Adds the Finite-State Transducer algorithm ([#11242](https://github.com/NVIDIA/cudf/pull/11242)) [@elstehle](https://github.com/elstehle) +- Refactor `collect_set` to use `cudf::distinct` and `cudf::lists::distinct` ([#11228](https://github.com/NVIDIA/cudf/pull/11228)) [@ttnghia](https://github.com/ttnghia) +- Treat zstd as stable in nvcomp releases 2.3.2 and later ([#11226](https://github.com/NVIDIA/cudf/pull/11226)) [@jbrennan333](https://github.com/jbrennan333) +- Add 24 bit dictionary support to Parquet writer ([#11216](https://github.com/NVIDIA/cudf/pull/11216)) [@devavret](https://github.com/devavret) +- Enable positive group indices for extractAllRecord on JNI ([#11215](https://github.com/NVIDIA/cudf/pull/11215)) [@anthony-chang](https://github.com/anthony-chang) +- JNI bindings for NTH_ELEMENT window aggregation ([#11201](https://github.com/NVIDIA/cudf/pull/11201)) [@mythrocks](https://github.com/mythrocks) +- Add JNI bindings for extractAllRecord ([#11196](https://github.com/NVIDIA/cudf/pull/11196)) [@anthony-chang](https://github.com/anthony-chang) +- Add `cudf.options` ([#11193](https://github.com/NVIDIA/cudf/pull/11193)) [@isVoid](https://github.com/isVoid) +- Add thrift support for parquet column and offset indexes ([#11178](https://github.com/NVIDIA/cudf/pull/11178)) [@etseidl](https://github.com/etseidl) +- Adding binary read/write as options for parquet ([#11160](https://github.com/NVIDIA/cudf/pull/11160)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Support `nth_element` for window functions ([#11158](https://github.com/NVIDIA/cudf/pull/11158)) [@mythrocks](https://github.com/mythrocks) +- Implement `lists::distinct` and `cudf::detail::stable_distinct` ([#11149](https://github.com/NVIDIA/cudf/pull/11149)) [@ttnghia](https://github.com/ttnghia) +- Implement Groupby pct_change ([#11144](https://github.com/NVIDIA/cudf/pull/11144)) [@skirui-source](https://github.com/skirui-source) +- Add JNI for set operations ([#11143](https://github.com/NVIDIA/cudf/pull/11143)) [@ttnghia](https://github.com/ttnghia) +- Remove deprecated PER_THREAD_DEFAULT_STREAM ([#11134](https://github.com/NVIDIA/cudf/pull/11134)) [@jbrennan333](https://github.com/jbrennan333) +- Added a Java method to check the existence of a list of keys in a map ([#11128](https://github.com/NVIDIA/cudf/pull/11128)) [@razajafri](https://github.com/razajafri) +- Feature/python benchmarking ([#11125](https://github.com/NVIDIA/cudf/pull/11125)) [@vyasr](https://github.com/vyasr) +- Support `nan_equality` in `cudf::distinct` ([#11118](https://github.com/NVIDIA/cudf/pull/11118)) [@ttnghia](https://github.com/ttnghia) +- Added JNI for getMapValueForKeys ([#11104](https://github.com/NVIDIA/cudf/pull/11104)) [@razajafri](https://github.com/razajafri) +- Refactor `semi_anti_join` ([#11100](https://github.com/NVIDIA/cudf/pull/11100)) [@ttnghia](https://github.com/ttnghia) +- Replace remaining instances of rmm::cuda_stream_default with cudf::default_stream_value ([#11082](https://github.com/NVIDIA/cudf/pull/11082)) [@jbrennan333](https://github.com/jbrennan333) +- Adds the Logical Stack algorithm ([#11078](https://github.com/NVIDIA/cudf/pull/11078)) [@elstehle](https://github.com/elstehle) +- Add doxygen-check pre-commit hook ([#11076](https://github.com/NVIDIA/cudf/pull/11076)) [@karthikeyann](https://github.com/karthikeyann) +- Use new nvCOMP API to optimize the decompression temp memory size ([#11064](https://github.com/NVIDIA/cudf/pull/11064)) [@vuule](https://github.com/vuule) +- Add Doxygen CI check ([#11057](https://github.com/NVIDIA/cudf/pull/11057)) [@karthikeyann](https://github.com/karthikeyann) +- Support `duplicate_keep_option` in `cudf::distinct` ([#11052](https://github.com/NVIDIA/cudf/pull/11052)) [@ttnghia](https://github.com/ttnghia) +- Support set operations ([#11043](https://github.com/NVIDIA/cudf/pull/11043)) [@ttnghia](https://github.com/ttnghia) +- Support for ZLIB compression in ORC writer ([#11036](https://github.com/NVIDIA/cudf/pull/11036)) [@vuule](https://github.com/vuule) +- Adding feature swaplevels ([#11027](https://github.com/NVIDIA/cudf/pull/11027)) [@VamsiTallam95](https://github.com/VamsiTallam95) +- Use nvCOMP for ZLIB decompression in ORC reader ([#11024](https://github.com/NVIDIA/cudf/pull/11024)) [@vuule](https://github.com/vuule) +- Function for bfill, ffill #9591 ([#11022](https://github.com/NVIDIA/cudf/pull/11022)) [@Sreekiran096](https://github.com/Sreekiran096) +- Generate group offsets from element labels ([#11017](https://github.com/NVIDIA/cudf/pull/11017)) [@ttnghia](https://github.com/ttnghia) +- Feature axes ([#10979](https://github.com/NVIDIA/cudf/pull/10979)) [@VamsiTallam95](https://github.com/VamsiTallam95) +- Generate group labels from offsets ([#10945](https://github.com/NVIDIA/cudf/pull/10945)) [@ttnghia](https://github.com/ttnghia) +- Add missing cuIO benchmark coverage for duration types ([#10933](https://github.com/NVIDIA/cudf/pull/10933)) [@vuule](https://github.com/vuule) +- Dask-cuDF cumulative groupby ops ([#10889](https://github.com/NVIDIA/cudf/pull/10889)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Reindex Improvements ([#10815](https://github.com/NVIDIA/cudf/pull/10815)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Implement value_counts for DataFrame ([#10813](https://github.com/NVIDIA/cudf/pull/10813)) [@martinfalisse](https://github.com/martinfalisse) ## 🛠️ Improvements -- Pin `dask` & `distributed` for release ([#11433](https://github.com/rapidsai/cudf/pull/11433)) [@galipremsagar](https://github.com/galipremsagar) -- Use documented header template for `doxygen` ([#11430](https://github.com/rapidsai/cudf/pull/11430)) [@galipremsagar](https://github.com/galipremsagar) -- Relax arrow version in dev env ([#11418](https://github.com/rapidsai/cudf/pull/11418)) [@galipremsagar](https://github.com/galipremsagar) -- Allow CuPy 11 ([#11393](https://github.com/rapidsai/cudf/pull/11393)) [@jakirkham](https://github.com/jakirkham) -- Improve multibyte_split performance ([#11347](https://github.com/rapidsai/cudf/pull/11347)) [@cwharris](https://github.com/cwharris) -- Switch death test to use explicit trap. ([#11326](https://github.com/rapidsai/cudf/pull/11326)) [@vyasr](https://github.com/vyasr) -- Add --output-on-failure to ctest args. ([#11321](https://github.com/rapidsai/cudf/pull/11321)) [@vyasr](https://github.com/vyasr) -- Consolidate remaining DataFrame/Series APIs ([#11315](https://github.com/rapidsai/cudf/pull/11315)) [@vyasr](https://github.com/vyasr) -- Add JNI support for the join_strings API ([#11309](https://github.com/rapidsai/cudf/pull/11309)) [@revans2](https://github.com/revans2) -- Add cupy version to setup.py install_requires ([#11306](https://github.com/rapidsai/cudf/pull/11306)) [@vyasr](https://github.com/vyasr) -- removing some unused code ([#11305](https://github.com/rapidsai/cudf/pull/11305)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add test of wildcard selection ([#11300](https://github.com/rapidsai/cudf/pull/11300)) [@vyasr](https://github.com/vyasr) -- Update parquet reader to take stream parameter ([#11294](https://github.com/rapidsai/cudf/pull/11294)) [@PointKernel](https://github.com/PointKernel) -- Spark list hashing ([#11292](https://github.com/rapidsai/cudf/pull/11292)) [@bdice](https://github.com/bdice) -- Remove legacy join APIs ([#11274](https://github.com/rapidsai/cudf/pull/11274)) [@vyasr](https://github.com/vyasr) -- Fix `cudf` recipes syntax ([#11273](https://github.com/rapidsai/cudf/pull/11273)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix `cudf` recipe ([#11267](https://github.com/rapidsai/cudf/pull/11267)) [@ajschmidt8](https://github.com/ajschmidt8) -- Cleanup config files ([#11266](https://github.com/rapidsai/cudf/pull/11266)) [@vyasr](https://github.com/vyasr) -- Run mypy on all packages ([#11265](https://github.com/rapidsai/cudf/pull/11265)) [@vyasr](https://github.com/vyasr) -- Update to isort 5.10.1. ([#11262](https://github.com/rapidsai/cudf/pull/11262)) [@vyasr](https://github.com/vyasr) -- Consolidate flake8 and pydocstyle configuration ([#11260](https://github.com/rapidsai/cudf/pull/11260)) [@vyasr](https://github.com/vyasr) -- Remove redundant black config specifications. ([#11258](https://github.com/rapidsai/cudf/pull/11258)) [@vyasr](https://github.com/vyasr) -- Ensure DeprecationWarnings are not introduced via pre-commit ([#11255](https://github.com/rapidsai/cudf/pull/11255)) [@wence-](https://github.com/wence-) -- Optimization to gpu::PreprocessColumnData in parquet reader. ([#11252](https://github.com/rapidsai/cudf/pull/11252)) [@nvdbaranec](https://github.com/nvdbaranec) -- Move rolling impl details to detail/ directory. ([#11250](https://github.com/rapidsai/cudf/pull/11250)) [@mythrocks](https://github.com/mythrocks) -- Remove `lists::drop_list_duplicates` ([#11236](https://github.com/rapidsai/cudf/pull/11236)) [@ttnghia](https://github.com/ttnghia) -- Use `cudf::lists::distinct` in Python binding ([#11234](https://github.com/rapidsai/cudf/pull/11234)) [@ttnghia](https://github.com/ttnghia) -- Use `cudf::lists::distinct` in Java binding ([#11233](https://github.com/rapidsai/cudf/pull/11233)) [@ttnghia](https://github.com/ttnghia) -- Use `cudf::distinct` in Java binding ([#11232](https://github.com/rapidsai/cudf/pull/11232)) [@ttnghia](https://github.com/ttnghia) -- Pin `dask-cuda` in dev environment ([#11229](https://github.com/rapidsai/cudf/pull/11229)) [@galipremsagar](https://github.com/galipremsagar) -- Remove cruft in map_lookup ([#11221](https://github.com/rapidsai/cudf/pull/11221)) [@mythrocks](https://github.com/mythrocks) -- Deprecate `skiprows` & `num_rows` in parquet reader ([#11218](https://github.com/rapidsai/cudf/pull/11218)) [@galipremsagar](https://github.com/galipremsagar) -- Remove Frame._index ([#11210](https://github.com/rapidsai/cudf/pull/11210)) [@vyasr](https://github.com/vyasr) -- Improve performance for `cudf::contains` when searching for a scalar ([#11202](https://github.com/rapidsai/cudf/pull/11202)) [@ttnghia](https://github.com/ttnghia) -- Document why Development component is needing for CMake. ([#11200](https://github.com/rapidsai/cudf/pull/11200)) [@vyasr](https://github.com/vyasr) -- cleanup unused code in rolling_test.hpp ([#11195](https://github.com/rapidsai/cudf/pull/11195)) [@karthikeyann](https://github.com/karthikeyann) -- Standardize join internals around DataFrame ([#11184](https://github.com/rapidsai/cudf/pull/11184)) [@vyasr](https://github.com/vyasr) -- Move character case table declarations from src to detail ([#11183](https://github.com/rapidsai/cudf/pull/11183)) [@davidwendt](https://github.com/davidwendt) -- Remove usage of Frame in StringMethods ([#11181](https://github.com/rapidsai/cudf/pull/11181)) [@vyasr](https://github.com/vyasr) -- Expose get_json_object_options to Python ([#11180](https://github.com/rapidsai/cudf/pull/11180)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Fix decimal128 stats in parquet writer ([#11179](https://github.com/rapidsai/cudf/pull/11179)) [@etseidl](https://github.com/etseidl) -- Modify CheckPageRows in parquet_test to use datasources ([#11177](https://github.com/rapidsai/cudf/pull/11177)) [@etseidl](https://github.com/etseidl) -- Pin max version of `cuda-python` to `11.7.0` ([#11174](https://github.com/rapidsai/cudf/pull/11174)) [@Ethyling](https://github.com/Ethyling) -- Refactor and optimize Frame.where ([#11168](https://github.com/rapidsai/cudf/pull/11168)) [@vyasr](https://github.com/vyasr) -- Add npos const static member to cudf::string_view ([#11166](https://github.com/rapidsai/cudf/pull/11166)) [@davidwendt](https://github.com/davidwendt) -- Move _drop_rows_by_label from Frame to IndexedFrame ([#11157](https://github.com/rapidsai/cudf/pull/11157)) [@vyasr](https://github.com/vyasr) -- Clean up _copy_type_metadata ([#11156](https://github.com/rapidsai/cudf/pull/11156)) [@vyasr](https://github.com/vyasr) -- Add `nvcc` conda package in dev environment ([#11154](https://github.com/rapidsai/cudf/pull/11154)) [@galipremsagar](https://github.com/galipremsagar) -- Struct binary comparison op functionality for spark rapids ([#11153](https://github.com/rapidsai/cudf/pull/11153)) [@rwlee](https://github.com/rwlee) -- Refactor inline conditionals. ([#11151](https://github.com/rapidsai/cudf/pull/11151)) [@bdice](https://github.com/bdice) -- Refactor Spark hashing tests ([#11145](https://github.com/rapidsai/cudf/pull/11145)) [@bdice](https://github.com/bdice) -- Add new `_from_data_like_self` factory ([#11140](https://github.com/rapidsai/cudf/pull/11140)) [@vyasr](https://github.com/vyasr) -- Update get_cucollections to use rapids-cmake ([#11139](https://github.com/rapidsai/cudf/pull/11139)) [@vyasr](https://github.com/vyasr) -- Remove unnecessary extra function for libcudacxx detection ([#11138](https://github.com/rapidsai/cudf/pull/11138)) [@vyasr](https://github.com/vyasr) -- Allow initial value for cudf::reduce and cudf::segmented_reduce. ([#11137](https://github.com/rapidsai/cudf/pull/11137)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Remove Index.replace API ([#11131](https://github.com/rapidsai/cudf/pull/11131)) [@vyasr](https://github.com/vyasr) -- Move char-type table function declarations from src to detail ([#11127](https://github.com/rapidsai/cudf/pull/11127)) [@davidwendt](https://github.com/davidwendt) -- Clean up repo root ([#11124](https://github.com/rapidsai/cudf/pull/11124)) [@bdice](https://github.com/bdice) -- Improve print formatting of strings containing newline characters. ([#11108](https://github.com/rapidsai/cudf/pull/11108)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix cudf::string_view::find() to return pos for empty string argument ([#11107](https://github.com/rapidsai/cudf/pull/11107)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-22.06 to branch-22.08 ([#11086](https://github.com/rapidsai/cudf/pull/11086)) [@bdice](https://github.com/bdice) -- Take iterators by value in clamp.cu. ([#11084](https://github.com/rapidsai/cudf/pull/11084)) [@bdice](https://github.com/bdice) -- Performance improvements for row to column conversions ([#11075](https://github.com/rapidsai/cudf/pull/11075)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Remove deprecated Index methods from Frame ([#11073](https://github.com/rapidsai/cudf/pull/11073)) [@vyasr](https://github.com/vyasr) -- Use per-page max compressed size estimate for compression ([#11066](https://github.com/rapidsai/cudf/pull/11066)) [@devavret](https://github.com/devavret) -- column to row refactor for performance ([#11063](https://github.com/rapidsai/cudf/pull/11063)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Include `skbuild` directory into `build.sh` `clean` operation ([#11060](https://github.com/rapidsai/cudf/pull/11060)) [@galipremsagar](https://github.com/galipremsagar) -- Unpin `dask` & `distributed` for development ([#11058](https://github.com/rapidsai/cudf/pull/11058)) [@galipremsagar](https://github.com/galipremsagar) -- Add support for `Series.between` ([#11051](https://github.com/rapidsai/cudf/pull/11051)) [@galipremsagar](https://github.com/galipremsagar) -- Fix groupby include ([#11046](https://github.com/rapidsai/cudf/pull/11046)) [@bwyogatama](https://github.com/bwyogatama) -- Regex cleanup internal reclass and reclass_device classes ([#11045](https://github.com/rapidsai/cudf/pull/11045)) [@davidwendt](https://github.com/davidwendt) -- Remove public API of cudf.merge_sorted. ([#11032](https://github.com/rapidsai/cudf/pull/11032)) [@bdice](https://github.com/bdice) -- Drop python `3.7` in code-base ([#11029](https://github.com/rapidsai/cudf/pull/11029)) [@galipremsagar](https://github.com/galipremsagar) -- Addition & integration of the integer power operator ([#11025](https://github.com/rapidsai/cudf/pull/11025)) [@AtlantaPepsi](https://github.com/AtlantaPepsi) -- Refactor `lists::contains` ([#11019](https://github.com/rapidsai/cudf/pull/11019)) [@ttnghia](https://github.com/ttnghia) -- Change build.sh to find C++ library by default and avoid shadowing CMAKE_ARGS ([#11013](https://github.com/rapidsai/cudf/pull/11013)) [@vyasr](https://github.com/vyasr) -- Clean up parquet unit test ([#11005](https://github.com/rapidsai/cudf/pull/11005)) [@PointKernel](https://github.com/PointKernel) -- Add missing #pragma once to header files ([#11004](https://github.com/rapidsai/cudf/pull/11004)) [@karthikeyann](https://github.com/karthikeyann) -- Cleanup `iterator.cuh` and add fixed point support for `scalar_optional_accessor` ([#10999](https://github.com/rapidsai/cudf/pull/10999)) [@ttnghia](https://github.com/ttnghia) -- Refactor `cudf::contains` ([#10997](https://github.com/rapidsai/cudf/pull/10997)) [@ttnghia](https://github.com/ttnghia) -- Remove Arrow CUDA IPC code ([#10995](https://github.com/rapidsai/cudf/pull/10995)) [@shwina](https://github.com/shwina) -- Change file extension for groupby benchmark ([#10985](https://github.com/rapidsai/cudf/pull/10985)) [@ttnghia](https://github.com/ttnghia) -- Sort recipe include checks. ([#10984](https://github.com/rapidsai/cudf/pull/10984)) [@bdice](https://github.com/bdice) -- Update cuCollections for thrust upgrade ([#10983](https://github.com/rapidsai/cudf/pull/10983)) [@PointKernel](https://github.com/PointKernel) -- Expose row-group size options in cudf ParquetWriter ([#10980](https://github.com/rapidsai/cudf/pull/10980)) [@rjzamora](https://github.com/rjzamora) -- Cleanup cudf::strings::detail::regex_parser class source ([#10975](https://github.com/rapidsai/cudf/pull/10975)) [@davidwendt](https://github.com/davidwendt) -- Handle missing fields as nulls in get_json_object() ([#10970](https://github.com/rapidsai/cudf/pull/10970)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) -- Fix license families to match all-caps expected by conda-verify. ([#10931](https://github.com/rapidsai/cudf/pull/10931)) [@bdice](https://github.com/bdice) -- Include <optional> for GCC 11 compatibility. ([#10927](https://github.com/rapidsai/cudf/pull/10927)) [@bdice](https://github.com/bdice) -- Enable builds with scikit-build ([#10919](https://github.com/rapidsai/cudf/pull/10919)) [@vyasr](https://github.com/vyasr) -- Improve `distinct` by using `cuco::static_map::retrieve_all` ([#10916](https://github.com/rapidsai/cudf/pull/10916)) [@PointKernel](https://github.com/PointKernel) -- update cudfjni to 22.08.0-SNAPSHOT ([#10910](https://github.com/rapidsai/cudf/pull/10910)) [@pxLi](https://github.com/pxLi) -- Improve the capture of fatal cuda error ([#10884](https://github.com/rapidsai/cudf/pull/10884)) [@sperlingxx](https://github.com/sperlingxx) -- Cleanup regex compiler operators and operands source ([#10879](https://github.com/rapidsai/cudf/pull/10879)) [@davidwendt](https://github.com/davidwendt) -- Buffer: make `.ptr` read-only ([#10872](https://github.com/rapidsai/cudf/pull/10872)) [@madsbk](https://github.com/madsbk) -- Configurable NaN handling in device_row_comparators ([#10870](https://github.com/rapidsai/cudf/pull/10870)) [@rwlee](https://github.com/rwlee) -- Register `cudf.core.groupby.Grouper` objects to dask `grouper_dispatch` ([#10838](https://github.com/rapidsai/cudf/pull/10838)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Upgrade to `arrow-8` ([#10816](https://github.com/rapidsai/cudf/pull/10816)) [@galipremsagar](https://github.com/galipremsagar) -- Remove _getattr_ method in RangeIndex class ([#10538](https://github.com/rapidsai/cudf/pull/10538)) [@skirui-source](https://github.com/skirui-source) -- Adding bins to value counts ([#8247](https://github.com/rapidsai/cudf/pull/8247)) [@marlenezw](https://github.com/marlenezw) +- Pin `dask` & `distributed` for release ([#11433](https://github.com/NVIDIA/cudf/pull/11433)) [@galipremsagar](https://github.com/galipremsagar) +- Use documented header template for `doxygen` ([#11430](https://github.com/NVIDIA/cudf/pull/11430)) [@galipremsagar](https://github.com/galipremsagar) +- Relax arrow version in dev env ([#11418](https://github.com/NVIDIA/cudf/pull/11418)) [@galipremsagar](https://github.com/galipremsagar) +- Allow CuPy 11 ([#11393](https://github.com/NVIDIA/cudf/pull/11393)) [@jakirkham](https://github.com/jakirkham) +- Improve multibyte_split performance ([#11347](https://github.com/NVIDIA/cudf/pull/11347)) [@cwharris](https://github.com/cwharris) +- Switch death test to use explicit trap. ([#11326](https://github.com/NVIDIA/cudf/pull/11326)) [@vyasr](https://github.com/vyasr) +- Add --output-on-failure to ctest args. ([#11321](https://github.com/NVIDIA/cudf/pull/11321)) [@vyasr](https://github.com/vyasr) +- Consolidate remaining DataFrame/Series APIs ([#11315](https://github.com/NVIDIA/cudf/pull/11315)) [@vyasr](https://github.com/vyasr) +- Add JNI support for the join_strings API ([#11309](https://github.com/NVIDIA/cudf/pull/11309)) [@revans2](https://github.com/revans2) +- Add cupy version to setup.py install_requires ([#11306](https://github.com/NVIDIA/cudf/pull/11306)) [@vyasr](https://github.com/vyasr) +- removing some unused code ([#11305](https://github.com/NVIDIA/cudf/pull/11305)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add test of wildcard selection ([#11300](https://github.com/NVIDIA/cudf/pull/11300)) [@vyasr](https://github.com/vyasr) +- Update parquet reader to take stream parameter ([#11294](https://github.com/NVIDIA/cudf/pull/11294)) [@PointKernel](https://github.com/PointKernel) +- Spark list hashing ([#11292](https://github.com/NVIDIA/cudf/pull/11292)) [@bdice](https://github.com/bdice) +- Remove legacy join APIs ([#11274](https://github.com/NVIDIA/cudf/pull/11274)) [@vyasr](https://github.com/vyasr) +- Fix `cudf` recipes syntax ([#11273](https://github.com/NVIDIA/cudf/pull/11273)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix `cudf` recipe ([#11267](https://github.com/NVIDIA/cudf/pull/11267)) [@ajschmidt8](https://github.com/ajschmidt8) +- Cleanup config files ([#11266](https://github.com/NVIDIA/cudf/pull/11266)) [@vyasr](https://github.com/vyasr) +- Run mypy on all packages ([#11265](https://github.com/NVIDIA/cudf/pull/11265)) [@vyasr](https://github.com/vyasr) +- Update to isort 5.10.1. ([#11262](https://github.com/NVIDIA/cudf/pull/11262)) [@vyasr](https://github.com/vyasr) +- Consolidate flake8 and pydocstyle configuration ([#11260](https://github.com/NVIDIA/cudf/pull/11260)) [@vyasr](https://github.com/vyasr) +- Remove redundant black config specifications. ([#11258](https://github.com/NVIDIA/cudf/pull/11258)) [@vyasr](https://github.com/vyasr) +- Ensure DeprecationWarnings are not introduced via pre-commit ([#11255](https://github.com/NVIDIA/cudf/pull/11255)) [@wence-](https://github.com/wence-) +- Optimization to gpu::PreprocessColumnData in parquet reader. ([#11252](https://github.com/NVIDIA/cudf/pull/11252)) [@nvdbaranec](https://github.com/nvdbaranec) +- Move rolling impl details to detail/ directory. ([#11250](https://github.com/NVIDIA/cudf/pull/11250)) [@mythrocks](https://github.com/mythrocks) +- Remove `lists::drop_list_duplicates` ([#11236](https://github.com/NVIDIA/cudf/pull/11236)) [@ttnghia](https://github.com/ttnghia) +- Use `cudf::lists::distinct` in Python binding ([#11234](https://github.com/NVIDIA/cudf/pull/11234)) [@ttnghia](https://github.com/ttnghia) +- Use `cudf::lists::distinct` in Java binding ([#11233](https://github.com/NVIDIA/cudf/pull/11233)) [@ttnghia](https://github.com/ttnghia) +- Use `cudf::distinct` in Java binding ([#11232](https://github.com/NVIDIA/cudf/pull/11232)) [@ttnghia](https://github.com/ttnghia) +- Pin `dask-cuda` in dev environment ([#11229](https://github.com/NVIDIA/cudf/pull/11229)) [@galipremsagar](https://github.com/galipremsagar) +- Remove cruft in map_lookup ([#11221](https://github.com/NVIDIA/cudf/pull/11221)) [@mythrocks](https://github.com/mythrocks) +- Deprecate `skiprows` & `num_rows` in parquet reader ([#11218](https://github.com/NVIDIA/cudf/pull/11218)) [@galipremsagar](https://github.com/galipremsagar) +- Remove Frame._index ([#11210](https://github.com/NVIDIA/cudf/pull/11210)) [@vyasr](https://github.com/vyasr) +- Improve performance for `cudf::contains` when searching for a scalar ([#11202](https://github.com/NVIDIA/cudf/pull/11202)) [@ttnghia](https://github.com/ttnghia) +- Document why Development component is needing for CMake. ([#11200](https://github.com/NVIDIA/cudf/pull/11200)) [@vyasr](https://github.com/vyasr) +- cleanup unused code in rolling_test.hpp ([#11195](https://github.com/NVIDIA/cudf/pull/11195)) [@karthikeyann](https://github.com/karthikeyann) +- Standardize join internals around DataFrame ([#11184](https://github.com/NVIDIA/cudf/pull/11184)) [@vyasr](https://github.com/vyasr) +- Move character case table declarations from src to detail ([#11183](https://github.com/NVIDIA/cudf/pull/11183)) [@davidwendt](https://github.com/davidwendt) +- Remove usage of Frame in StringMethods ([#11181](https://github.com/NVIDIA/cudf/pull/11181)) [@vyasr](https://github.com/vyasr) +- Expose get_json_object_options to Python ([#11180](https://github.com/NVIDIA/cudf/pull/11180)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Fix decimal128 stats in parquet writer ([#11179](https://github.com/NVIDIA/cudf/pull/11179)) [@etseidl](https://github.com/etseidl) +- Modify CheckPageRows in parquet_test to use datasources ([#11177](https://github.com/NVIDIA/cudf/pull/11177)) [@etseidl](https://github.com/etseidl) +- Pin max version of `cuda-python` to `11.7.0` ([#11174](https://github.com/NVIDIA/cudf/pull/11174)) [@Ethyling](https://github.com/Ethyling) +- Refactor and optimize Frame.where ([#11168](https://github.com/NVIDIA/cudf/pull/11168)) [@vyasr](https://github.com/vyasr) +- Add npos const static member to cudf::string_view ([#11166](https://github.com/NVIDIA/cudf/pull/11166)) [@davidwendt](https://github.com/davidwendt) +- Move _drop_rows_by_label from Frame to IndexedFrame ([#11157](https://github.com/NVIDIA/cudf/pull/11157)) [@vyasr](https://github.com/vyasr) +- Clean up _copy_type_metadata ([#11156](https://github.com/NVIDIA/cudf/pull/11156)) [@vyasr](https://github.com/vyasr) +- Add `nvcc` conda package in dev environment ([#11154](https://github.com/NVIDIA/cudf/pull/11154)) [@galipremsagar](https://github.com/galipremsagar) +- Struct binary comparison op functionality for spark rapids ([#11153](https://github.com/NVIDIA/cudf/pull/11153)) [@rwlee](https://github.com/rwlee) +- Refactor inline conditionals. ([#11151](https://github.com/NVIDIA/cudf/pull/11151)) [@bdice](https://github.com/bdice) +- Refactor Spark hashing tests ([#11145](https://github.com/NVIDIA/cudf/pull/11145)) [@bdice](https://github.com/bdice) +- Add new `_from_data_like_self` factory ([#11140](https://github.com/NVIDIA/cudf/pull/11140)) [@vyasr](https://github.com/vyasr) +- Update get_cucollections to use rapids-cmake ([#11139](https://github.com/NVIDIA/cudf/pull/11139)) [@vyasr](https://github.com/vyasr) +- Remove unnecessary extra function for libcudacxx detection ([#11138](https://github.com/NVIDIA/cudf/pull/11138)) [@vyasr](https://github.com/vyasr) +- Allow initial value for cudf::reduce and cudf::segmented_reduce. ([#11137](https://github.com/NVIDIA/cudf/pull/11137)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Remove Index.replace API ([#11131](https://github.com/NVIDIA/cudf/pull/11131)) [@vyasr](https://github.com/vyasr) +- Move char-type table function declarations from src to detail ([#11127](https://github.com/NVIDIA/cudf/pull/11127)) [@davidwendt](https://github.com/davidwendt) +- Clean up repo root ([#11124](https://github.com/NVIDIA/cudf/pull/11124)) [@bdice](https://github.com/bdice) +- Improve print formatting of strings containing newline characters. ([#11108](https://github.com/NVIDIA/cudf/pull/11108)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix cudf::string_view::find() to return pos for empty string argument ([#11107](https://github.com/NVIDIA/cudf/pull/11107)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-22.06 to branch-22.08 ([#11086](https://github.com/NVIDIA/cudf/pull/11086)) [@bdice](https://github.com/bdice) +- Take iterators by value in clamp.cu. ([#11084](https://github.com/NVIDIA/cudf/pull/11084)) [@bdice](https://github.com/bdice) +- Performance improvements for row to column conversions ([#11075](https://github.com/NVIDIA/cudf/pull/11075)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Remove deprecated Index methods from Frame ([#11073](https://github.com/NVIDIA/cudf/pull/11073)) [@vyasr](https://github.com/vyasr) +- Use per-page max compressed size estimate for compression ([#11066](https://github.com/NVIDIA/cudf/pull/11066)) [@devavret](https://github.com/devavret) +- column to row refactor for performance ([#11063](https://github.com/NVIDIA/cudf/pull/11063)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Include `skbuild` directory into `build.sh` `clean` operation ([#11060](https://github.com/NVIDIA/cudf/pull/11060)) [@galipremsagar](https://github.com/galipremsagar) +- Unpin `dask` & `distributed` for development ([#11058](https://github.com/NVIDIA/cudf/pull/11058)) [@galipremsagar](https://github.com/galipremsagar) +- Add support for `Series.between` ([#11051](https://github.com/NVIDIA/cudf/pull/11051)) [@galipremsagar](https://github.com/galipremsagar) +- Fix groupby include ([#11046](https://github.com/NVIDIA/cudf/pull/11046)) [@bwyogatama](https://github.com/bwyogatama) +- Regex cleanup internal reclass and reclass_device classes ([#11045](https://github.com/NVIDIA/cudf/pull/11045)) [@davidwendt](https://github.com/davidwendt) +- Remove public API of cudf.merge_sorted. ([#11032](https://github.com/NVIDIA/cudf/pull/11032)) [@bdice](https://github.com/bdice) +- Drop python `3.7` in code-base ([#11029](https://github.com/NVIDIA/cudf/pull/11029)) [@galipremsagar](https://github.com/galipremsagar) +- Addition & integration of the integer power operator ([#11025](https://github.com/NVIDIA/cudf/pull/11025)) [@AtlantaPepsi](https://github.com/AtlantaPepsi) +- Refactor `lists::contains` ([#11019](https://github.com/NVIDIA/cudf/pull/11019)) [@ttnghia](https://github.com/ttnghia) +- Change build.sh to find C++ library by default and avoid shadowing CMAKE_ARGS ([#11013](https://github.com/NVIDIA/cudf/pull/11013)) [@vyasr](https://github.com/vyasr) +- Clean up parquet unit test ([#11005](https://github.com/NVIDIA/cudf/pull/11005)) [@PointKernel](https://github.com/PointKernel) +- Add missing #pragma once to header files ([#11004](https://github.com/NVIDIA/cudf/pull/11004)) [@karthikeyann](https://github.com/karthikeyann) +- Cleanup `iterator.cuh` and add fixed point support for `scalar_optional_accessor` ([#10999](https://github.com/NVIDIA/cudf/pull/10999)) [@ttnghia](https://github.com/ttnghia) +- Refactor `cudf::contains` ([#10997](https://github.com/NVIDIA/cudf/pull/10997)) [@ttnghia](https://github.com/ttnghia) +- Remove Arrow CUDA IPC code ([#10995](https://github.com/NVIDIA/cudf/pull/10995)) [@shwina](https://github.com/shwina) +- Change file extension for groupby benchmark ([#10985](https://github.com/NVIDIA/cudf/pull/10985)) [@ttnghia](https://github.com/ttnghia) +- Sort recipe include checks. ([#10984](https://github.com/NVIDIA/cudf/pull/10984)) [@bdice](https://github.com/bdice) +- Update cuCollections for thrust upgrade ([#10983](https://github.com/NVIDIA/cudf/pull/10983)) [@PointKernel](https://github.com/PointKernel) +- Expose row-group size options in cudf ParquetWriter ([#10980](https://github.com/NVIDIA/cudf/pull/10980)) [@rjzamora](https://github.com/rjzamora) +- Cleanup cudf::strings::detail::regex_parser class source ([#10975](https://github.com/NVIDIA/cudf/pull/10975)) [@davidwendt](https://github.com/davidwendt) +- Handle missing fields as nulls in get_json_object() ([#10970](https://github.com/NVIDIA/cudf/pull/10970)) [@SrikarVanavasam](https://github.com/SrikarVanavasam) +- Fix license families to match all-caps expected by conda-verify. ([#10931](https://github.com/NVIDIA/cudf/pull/10931)) [@bdice](https://github.com/bdice) +- Include <optional> for GCC 11 compatibility. ([#10927](https://github.com/NVIDIA/cudf/pull/10927)) [@bdice](https://github.com/bdice) +- Enable builds with scikit-build ([#10919](https://github.com/NVIDIA/cudf/pull/10919)) [@vyasr](https://github.com/vyasr) +- Improve `distinct` by using `cuco::static_map::retrieve_all` ([#10916](https://github.com/NVIDIA/cudf/pull/10916)) [@PointKernel](https://github.com/PointKernel) +- update cudfjni to 22.08.0-SNAPSHOT ([#10910](https://github.com/NVIDIA/cudf/pull/10910)) [@pxLi](https://github.com/pxLi) +- Improve the capture of fatal cuda error ([#10884](https://github.com/NVIDIA/cudf/pull/10884)) [@sperlingxx](https://github.com/sperlingxx) +- Cleanup regex compiler operators and operands source ([#10879](https://github.com/NVIDIA/cudf/pull/10879)) [@davidwendt](https://github.com/davidwendt) +- Buffer: make `.ptr` read-only ([#10872](https://github.com/NVIDIA/cudf/pull/10872)) [@madsbk](https://github.com/madsbk) +- Configurable NaN handling in device_row_comparators ([#10870](https://github.com/NVIDIA/cudf/pull/10870)) [@rwlee](https://github.com/rwlee) +- Register `cudf.core.groupby.Grouper` objects to dask `grouper_dispatch` ([#10838](https://github.com/NVIDIA/cudf/pull/10838)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Upgrade to `arrow-8` ([#10816](https://github.com/NVIDIA/cudf/pull/10816)) [@galipremsagar](https://github.com/galipremsagar) +- Remove _getattr_ method in RangeIndex class ([#10538](https://github.com/NVIDIA/cudf/pull/10538)) [@skirui-source](https://github.com/skirui-source) +- Adding bins to value counts ([#8247](https://github.com/NVIDIA/cudf/pull/8247)) [@marlenezw](https://github.com/marlenezw) # cuDF 22.06.00 (7 Jun 2022) ## 🚨 Breaking Changes -- Enable Zstandard decompression only when all nvcomp integrations are enabled ([#10944](https://github.com/rapidsai/cudf/pull/10944)) [@vuule](https://github.com/vuule) -- Rename `sliced_child` to `get_sliced_child`. ([#10885](https://github.com/rapidsai/cudf/pull/10885)) [@bdice](https://github.com/bdice) -- Add parameters to control page size in Parquet writer ([#10882](https://github.com/rapidsai/cudf/pull/10882)) [@etseidl](https://github.com/etseidl) -- Make cudf::test::expect_columns_equal() to fail when comparing unsanitary lists. ([#10880](https://github.com/rapidsai/cudf/pull/10880)) [@nvdbaranec](https://github.com/nvdbaranec) -- Cleanup regex compiler fixed quantifiers source ([#10843](https://github.com/rapidsai/cudf/pull/10843)) [@davidwendt](https://github.com/davidwendt) -- Refactor `cudf::contains`, renaming and switching parameters role ([#10802](https://github.com/rapidsai/cudf/pull/10802)) [@ttnghia](https://github.com/ttnghia) -- Generic serialization of all column types ([#10784](https://github.com/rapidsai/cudf/pull/10784)) [@wence-](https://github.com/wence-) -- Return per-file metadata from readers ([#10782](https://github.com/rapidsai/cudf/pull/10782)) [@vuule](https://github.com/vuule) -- HostColumnVectoreCore#isNull should return true for out-of-range rows ([#10779](https://github.com/rapidsai/cudf/pull/10779)) [@gerashegalov](https://github.com/gerashegalov) -- Update `groupby::hash` to use new row operators for keys ([#10770](https://github.com/rapidsai/cudf/pull/10770)) [@PointKernel](https://github.com/PointKernel) -- update mangle_dupe_cols behavior in csv reader to match pandas 1.4.0 behavior ([#10749](https://github.com/rapidsai/cudf/pull/10749)) [@karthikeyann](https://github.com/karthikeyann) -- Rename CUDA_TRY macro to CUDF_CUDA_TRY, rename CHECK_CUDA macro to CUDF_CHECK_CUDA. ([#10589](https://github.com/rapidsai/cudf/pull/10589)) [@bdice](https://github.com/bdice) -- Upgrade `cudf` to support `pandas` 1.4.x versions ([#10584](https://github.com/rapidsai/cudf/pull/10584)) [@galipremsagar](https://github.com/galipremsagar) -- Move binop methods from Frame to IndexedFrame and standardize the docstring ([#10576](https://github.com/rapidsai/cudf/pull/10576)) [@vyasr](https://github.com/vyasr) -- Add default= kwarg to .list.get() accessor method ([#10547](https://github.com/rapidsai/cudf/pull/10547)) [@shwina](https://github.com/shwina) -- Remove deprecated `decimal_cols_as_float` in the ORC reader ([#10515](https://github.com/rapidsai/cudf/pull/10515)) [@vuule](https://github.com/vuule) -- Support nvComp 2.3 if local, otherwise use nvcomp 2.2 ([#10513](https://github.com/rapidsai/cudf/pull/10513)) [@robertmaynard](https://github.com/robertmaynard) -- Fix findall_record to return empty list for no matches ([#10491](https://github.com/rapidsai/cudf/pull/10491)) [@davidwendt](https://github.com/davidwendt) -- Namespace/Docstring Fixes for Reduction ([#10471](https://github.com/rapidsai/cudf/pull/10471)) [@isVoid](https://github.com/isVoid) -- Additional refactoring of hash functions ([#10462](https://github.com/rapidsai/cudf/pull/10462)) [@bdice](https://github.com/bdice) -- Fix default value of str.split expand parameter. ([#10457](https://github.com/rapidsai/cudf/pull/10457)) [@bdice](https://github.com/bdice) -- Remove deprecated code. ([#10450](https://github.com/rapidsai/cudf/pull/10450)) [@vyasr](https://github.com/vyasr) +- Enable Zstandard decompression only when all nvcomp integrations are enabled ([#10944](https://github.com/NVIDIA/cudf/pull/10944)) [@vuule](https://github.com/vuule) +- Rename `sliced_child` to `get_sliced_child`. ([#10885](https://github.com/NVIDIA/cudf/pull/10885)) [@bdice](https://github.com/bdice) +- Add parameters to control page size in Parquet writer ([#10882](https://github.com/NVIDIA/cudf/pull/10882)) [@etseidl](https://github.com/etseidl) +- Make cudf::test::expect_columns_equal() to fail when comparing unsanitary lists. ([#10880](https://github.com/NVIDIA/cudf/pull/10880)) [@nvdbaranec](https://github.com/nvdbaranec) +- Cleanup regex compiler fixed quantifiers source ([#10843](https://github.com/NVIDIA/cudf/pull/10843)) [@davidwendt](https://github.com/davidwendt) +- Refactor `cudf::contains`, renaming and switching parameters role ([#10802](https://github.com/NVIDIA/cudf/pull/10802)) [@ttnghia](https://github.com/ttnghia) +- Generic serialization of all column types ([#10784](https://github.com/NVIDIA/cudf/pull/10784)) [@wence-](https://github.com/wence-) +- Return per-file metadata from readers ([#10782](https://github.com/NVIDIA/cudf/pull/10782)) [@vuule](https://github.com/vuule) +- HostColumnVectoreCore#isNull should return true for out-of-range rows ([#10779](https://github.com/NVIDIA/cudf/pull/10779)) [@gerashegalov](https://github.com/gerashegalov) +- Update `groupby::hash` to use new row operators for keys ([#10770](https://github.com/NVIDIA/cudf/pull/10770)) [@PointKernel](https://github.com/PointKernel) +- update mangle_dupe_cols behavior in csv reader to match pandas 1.4.0 behavior ([#10749](https://github.com/NVIDIA/cudf/pull/10749)) [@karthikeyann](https://github.com/karthikeyann) +- Rename CUDA_TRY macro to CUDF_CUDA_TRY, rename CHECK_CUDA macro to CUDF_CHECK_CUDA. ([#10589](https://github.com/NVIDIA/cudf/pull/10589)) [@bdice](https://github.com/bdice) +- Upgrade `cudf` to support `pandas` 1.4.x versions ([#10584](https://github.com/NVIDIA/cudf/pull/10584)) [@galipremsagar](https://github.com/galipremsagar) +- Move binop methods from Frame to IndexedFrame and standardize the docstring ([#10576](https://github.com/NVIDIA/cudf/pull/10576)) [@vyasr](https://github.com/vyasr) +- Add default= kwarg to .list.get() accessor method ([#10547](https://github.com/NVIDIA/cudf/pull/10547)) [@shwina](https://github.com/shwina) +- Remove deprecated `decimal_cols_as_float` in the ORC reader ([#10515](https://github.com/NVIDIA/cudf/pull/10515)) [@vuule](https://github.com/vuule) +- Support nvComp 2.3 if local, otherwise use nvcomp 2.2 ([#10513](https://github.com/NVIDIA/cudf/pull/10513)) [@robertmaynard](https://github.com/robertmaynard) +- Fix findall_record to return empty list for no matches ([#10491](https://github.com/NVIDIA/cudf/pull/10491)) [@davidwendt](https://github.com/davidwendt) +- Namespace/Docstring Fixes for Reduction ([#10471](https://github.com/NVIDIA/cudf/pull/10471)) [@isVoid](https://github.com/isVoid) +- Additional refactoring of hash functions ([#10462](https://github.com/NVIDIA/cudf/pull/10462)) [@bdice](https://github.com/bdice) +- Fix default value of str.split expand parameter. ([#10457](https://github.com/NVIDIA/cudf/pull/10457)) [@bdice](https://github.com/bdice) +- Remove deprecated code. ([#10450](https://github.com/NVIDIA/cudf/pull/10450)) [@vyasr](https://github.com/vyasr) ## 🐛 Bug Fixes -- Fix single column `MultiIndex` issue in `sort_index` ([#10957](https://github.com/rapidsai/cudf/pull/10957)) [@galipremsagar](https://github.com/galipremsagar) -- Make SerializedTableHeader(numRows) public ([#10949](https://github.com/rapidsai/cudf/pull/10949)) [@gerashegalov](https://github.com/gerashegalov) -- Fix `gcc_linux` version pinning in dev environment ([#10943](https://github.com/rapidsai/cudf/pull/10943)) [@galipremsagar](https://github.com/galipremsagar) -- Fix an issue with reading raw string in `cudf.read_json` ([#10924](https://github.com/rapidsai/cudf/pull/10924)) [@galipremsagar](https://github.com/galipremsagar) -- Make cudf::test::expect_columns_equal() to fail when comparing unsanitary lists. ([#10880](https://github.com/rapidsai/cudf/pull/10880)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix segmented_reduce on empty column with non-empty offsets ([#10876](https://github.com/rapidsai/cudf/pull/10876)) [@davidwendt](https://github.com/davidwendt) -- Fix dask-cudf groupby handling when grouping by all columns ([#10866](https://github.com/rapidsai/cudf/pull/10866)) [@charlesbluca](https://github.com/charlesbluca) -- Fix a bug in `distinct`: using nested nulls logic ([#10848](https://github.com/rapidsai/cudf/pull/10848)) [@PointKernel](https://github.com/PointKernel) -- Fix constness / references in weak ordering operator() signatures. ([#10846](https://github.com/rapidsai/cudf/pull/10846)) [@bdice](https://github.com/bdice) -- Suppress sizeof-array-div warnings in thrust found by gcc-11 ([#10840](https://github.com/rapidsai/cudf/pull/10840)) [@robertmaynard](https://github.com/robertmaynard) -- Add handling for string by-columns in dask-cudf groupby ([#10830](https://github.com/rapidsai/cudf/pull/10830)) [@charlesbluca](https://github.com/charlesbluca) -- Fix compile warning in search.cu ([#10827](https://github.com/rapidsai/cudf/pull/10827)) [@davidwendt](https://github.com/davidwendt) -- Fix element access const correctness in `hostdevice_vector` ([#10804](https://github.com/rapidsai/cudf/pull/10804)) [@vuule](https://github.com/vuule) -- Update `cuco` git tag ([#10788](https://github.com/rapidsai/cudf/pull/10788)) [@PointKernel](https://github.com/PointKernel) -- HostColumnVectoreCore#isNull should return true for out-of-range rows ([#10779](https://github.com/rapidsai/cudf/pull/10779)) [@gerashegalov](https://github.com/gerashegalov) -- Fixing deprecation warnings in test_orc.py ([#10772](https://github.com/rapidsai/cudf/pull/10772)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Enable writing to `s3` storage in chunked parquet writer ([#10769](https://github.com/rapidsai/cudf/pull/10769)) [@galipremsagar](https://github.com/galipremsagar) -- Fix construction of nested structs with EMPTY child ([#10761](https://github.com/rapidsai/cudf/pull/10761)) [@shwina](https://github.com/shwina) -- Fix replace error when regex has only zero match quantifiers ([#10760](https://github.com/rapidsai/cudf/pull/10760)) [@davidwendt](https://github.com/davidwendt) -- Fix an issue with one_level_list schemas in parquet reader. ([#10750](https://github.com/rapidsai/cudf/pull/10750)) [@nvdbaranec](https://github.com/nvdbaranec) -- update mangle_dupe_cols behavior in csv reader to match pandas 1.4.0 behavior ([#10749](https://github.com/rapidsai/cudf/pull/10749)) [@karthikeyann](https://github.com/karthikeyann) -- Fix `cupy` function in notebook ([#10737](https://github.com/rapidsai/cudf/pull/10737)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix `fillna` to retain `columns` when it is `MultiIndex` ([#10729](https://github.com/rapidsai/cudf/pull/10729)) [@galipremsagar](https://github.com/galipremsagar) -- Fix scatter for all-empty-string column case ([#10724](https://github.com/rapidsai/cudf/pull/10724)) [@davidwendt](https://github.com/davidwendt) -- Retain series name in `Series.apply` ([#10716](https://github.com/rapidsai/cudf/pull/10716)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Correct build dir `cudf-config` dependency issues for static builds ([#10704](https://github.com/rapidsai/cudf/pull/10704)) [@robertmaynard](https://github.com/robertmaynard) -- Fix list of testing requirements in setup.py. ([#10678](https://github.com/rapidsai/cudf/pull/10678)) [@bdice](https://github.com/bdice) -- Fix rounding to zero error in stod on very small float numbers ([#10672](https://github.com/rapidsai/cudf/pull/10672)) [@davidwendt](https://github.com/davidwendt) -- cuco isn't a cudf dependency when we are built shared ([#10662](https://github.com/rapidsai/cudf/pull/10662)) [@robertmaynard](https://github.com/robertmaynard) -- Fix to_timestamps to support Z for %z format specifier ([#10617](https://github.com/rapidsai/cudf/pull/10617)) [@davidwendt](https://github.com/davidwendt) -- Verify compression type in Parquet reader ([#10610](https://github.com/rapidsai/cudf/pull/10610)) [@vuule](https://github.com/vuule) -- Fix struct row comparator's exception on empty structs ([#10604](https://github.com/rapidsai/cudf/pull/10604)) [@sperlingxx](https://github.com/sperlingxx) -- Fix strings strip() to accept only str Scalar for to_strip parameter ([#10597](https://github.com/rapidsai/cudf/pull/10597)) [@davidwendt](https://github.com/davidwendt) -- Fix has_atomic_support check in can_use_hash_groupby() ([#10588](https://github.com/rapidsai/cudf/pull/10588)) [@jbrennan333](https://github.com/jbrennan333) -- Revert Thrust 1.16 to Thrust 1.15 ([#10586](https://github.com/rapidsai/cudf/pull/10586)) [@bdice](https://github.com/bdice) -- Fix missing RMM_STATIC_CUDART define when compiling JNI with static CUDA runtime ([#10585](https://github.com/rapidsai/cudf/pull/10585)) [@jlowe](https://github.com/jlowe) -- pin more cmake versions ([#10570](https://github.com/rapidsai/cudf/pull/10570)) [@robertmaynard](https://github.com/robertmaynard) -- Re-enable Build Metrics Report ([#10562](https://github.com/rapidsai/cudf/pull/10562)) [@davidwendt](https://github.com/davidwendt) -- Remove statically linked CUDA runtime check in Java build ([#10532](https://github.com/rapidsai/cudf/pull/10532)) [@jlowe](https://github.com/jlowe) -- Fix temp data cleanup in `test_text.py` ([#10524](https://github.com/rapidsai/cudf/pull/10524)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Update pre-commit to run black 22.3.0 ([#10523](https://github.com/rapidsai/cudf/pull/10523)) [@vyasr](https://github.com/vyasr) -- Remove deprecated `decimal_cols_as_float` in the ORC reader ([#10515](https://github.com/rapidsai/cudf/pull/10515)) [@vuule](https://github.com/vuule) -- Fix findall_record to return empty list for no matches ([#10491](https://github.com/rapidsai/cudf/pull/10491)) [@davidwendt](https://github.com/davidwendt) -- Allow users to specify data types for a subset of columns in `read_csv` ([#10484](https://github.com/rapidsai/cudf/pull/10484)) [@vuule](https://github.com/vuule) -- Fix default value of str.split expand parameter. ([#10457](https://github.com/rapidsai/cudf/pull/10457)) [@bdice](https://github.com/bdice) -- Improve coverage of dask-cudf's groupby aggregation, add tests for `dropna` support ([#10449](https://github.com/rapidsai/cudf/pull/10449)) [@charlesbluca](https://github.com/charlesbluca) -- Allow string aggs for `dask_cudf.CudfDataFrameGroupBy.aggregate` ([#10222](https://github.com/rapidsai/cudf/pull/10222)) [@charlesbluca](https://github.com/charlesbluca) -- In-place updates with loc or iloc don't work correctly when the LHS has more than one column ([#9918](https://github.com/rapidsai/cudf/pull/9918)) [@skirui-source](https://github.com/skirui-source) +- Fix single column `MultiIndex` issue in `sort_index` ([#10957](https://github.com/NVIDIA/cudf/pull/10957)) [@galipremsagar](https://github.com/galipremsagar) +- Make SerializedTableHeader(numRows) public ([#10949](https://github.com/NVIDIA/cudf/pull/10949)) [@gerashegalov](https://github.com/gerashegalov) +- Fix `gcc_linux` version pinning in dev environment ([#10943](https://github.com/NVIDIA/cudf/pull/10943)) [@galipremsagar](https://github.com/galipremsagar) +- Fix an issue with reading raw string in `cudf.read_json` ([#10924](https://github.com/NVIDIA/cudf/pull/10924)) [@galipremsagar](https://github.com/galipremsagar) +- Make cudf::test::expect_columns_equal() to fail when comparing unsanitary lists. ([#10880](https://github.com/NVIDIA/cudf/pull/10880)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix segmented_reduce on empty column with non-empty offsets ([#10876](https://github.com/NVIDIA/cudf/pull/10876)) [@davidwendt](https://github.com/davidwendt) +- Fix dask-cudf groupby handling when grouping by all columns ([#10866](https://github.com/NVIDIA/cudf/pull/10866)) [@charlesbluca](https://github.com/charlesbluca) +- Fix a bug in `distinct`: using nested nulls logic ([#10848](https://github.com/NVIDIA/cudf/pull/10848)) [@PointKernel](https://github.com/PointKernel) +- Fix constness / references in weak ordering operator() signatures. ([#10846](https://github.com/NVIDIA/cudf/pull/10846)) [@bdice](https://github.com/bdice) +- Suppress sizeof-array-div warnings in thrust found by gcc-11 ([#10840](https://github.com/NVIDIA/cudf/pull/10840)) [@robertmaynard](https://github.com/robertmaynard) +- Add handling for string by-columns in dask-cudf groupby ([#10830](https://github.com/NVIDIA/cudf/pull/10830)) [@charlesbluca](https://github.com/charlesbluca) +- Fix compile warning in search.cu ([#10827](https://github.com/NVIDIA/cudf/pull/10827)) [@davidwendt](https://github.com/davidwendt) +- Fix element access const correctness in `hostdevice_vector` ([#10804](https://github.com/NVIDIA/cudf/pull/10804)) [@vuule](https://github.com/vuule) +- Update `cuco` git tag ([#10788](https://github.com/NVIDIA/cudf/pull/10788)) [@PointKernel](https://github.com/PointKernel) +- HostColumnVectoreCore#isNull should return true for out-of-range rows ([#10779](https://github.com/NVIDIA/cudf/pull/10779)) [@gerashegalov](https://github.com/gerashegalov) +- Fixing deprecation warnings in test_orc.py ([#10772](https://github.com/NVIDIA/cudf/pull/10772)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Enable writing to `s3` storage in chunked parquet writer ([#10769](https://github.com/NVIDIA/cudf/pull/10769)) [@galipremsagar](https://github.com/galipremsagar) +- Fix construction of nested structs with EMPTY child ([#10761](https://github.com/NVIDIA/cudf/pull/10761)) [@shwina](https://github.com/shwina) +- Fix replace error when regex has only zero match quantifiers ([#10760](https://github.com/NVIDIA/cudf/pull/10760)) [@davidwendt](https://github.com/davidwendt) +- Fix an issue with one_level_list schemas in parquet reader. ([#10750](https://github.com/NVIDIA/cudf/pull/10750)) [@nvdbaranec](https://github.com/nvdbaranec) +- update mangle_dupe_cols behavior in csv reader to match pandas 1.4.0 behavior ([#10749](https://github.com/NVIDIA/cudf/pull/10749)) [@karthikeyann](https://github.com/karthikeyann) +- Fix `cupy` function in notebook ([#10737](https://github.com/NVIDIA/cudf/pull/10737)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix `fillna` to retain `columns` when it is `MultiIndex` ([#10729](https://github.com/NVIDIA/cudf/pull/10729)) [@galipremsagar](https://github.com/galipremsagar) +- Fix scatter for all-empty-string column case ([#10724](https://github.com/NVIDIA/cudf/pull/10724)) [@davidwendt](https://github.com/davidwendt) +- Retain series name in `Series.apply` ([#10716](https://github.com/NVIDIA/cudf/pull/10716)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Correct build dir `cudf-config` dependency issues for static builds ([#10704](https://github.com/NVIDIA/cudf/pull/10704)) [@robertmaynard](https://github.com/robertmaynard) +- Fix list of testing requirements in setup.py. ([#10678](https://github.com/NVIDIA/cudf/pull/10678)) [@bdice](https://github.com/bdice) +- Fix rounding to zero error in stod on very small float numbers ([#10672](https://github.com/NVIDIA/cudf/pull/10672)) [@davidwendt](https://github.com/davidwendt) +- cuco isn't a cudf dependency when we are built shared ([#10662](https://github.com/NVIDIA/cudf/pull/10662)) [@robertmaynard](https://github.com/robertmaynard) +- Fix to_timestamps to support Z for %z format specifier ([#10617](https://github.com/NVIDIA/cudf/pull/10617)) [@davidwendt](https://github.com/davidwendt) +- Verify compression type in Parquet reader ([#10610](https://github.com/NVIDIA/cudf/pull/10610)) [@vuule](https://github.com/vuule) +- Fix struct row comparator's exception on empty structs ([#10604](https://github.com/NVIDIA/cudf/pull/10604)) [@sperlingxx](https://github.com/sperlingxx) +- Fix strings strip() to accept only str Scalar for to_strip parameter ([#10597](https://github.com/NVIDIA/cudf/pull/10597)) [@davidwendt](https://github.com/davidwendt) +- Fix has_atomic_support check in can_use_hash_groupby() ([#10588](https://github.com/NVIDIA/cudf/pull/10588)) [@jbrennan333](https://github.com/jbrennan333) +- Revert Thrust 1.16 to Thrust 1.15 ([#10586](https://github.com/NVIDIA/cudf/pull/10586)) [@bdice](https://github.com/bdice) +- Fix missing RMM_STATIC_CUDART define when compiling JNI with static CUDA runtime ([#10585](https://github.com/NVIDIA/cudf/pull/10585)) [@jlowe](https://github.com/jlowe) +- pin more cmake versions ([#10570](https://github.com/NVIDIA/cudf/pull/10570)) [@robertmaynard](https://github.com/robertmaynard) +- Re-enable Build Metrics Report ([#10562](https://github.com/NVIDIA/cudf/pull/10562)) [@davidwendt](https://github.com/davidwendt) +- Remove statically linked CUDA runtime check in Java build ([#10532](https://github.com/NVIDIA/cudf/pull/10532)) [@jlowe](https://github.com/jlowe) +- Fix temp data cleanup in `test_text.py` ([#10524](https://github.com/NVIDIA/cudf/pull/10524)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Update pre-commit to run black 22.3.0 ([#10523](https://github.com/NVIDIA/cudf/pull/10523)) [@vyasr](https://github.com/vyasr) +- Remove deprecated `decimal_cols_as_float` in the ORC reader ([#10515](https://github.com/NVIDIA/cudf/pull/10515)) [@vuule](https://github.com/vuule) +- Fix findall_record to return empty list for no matches ([#10491](https://github.com/NVIDIA/cudf/pull/10491)) [@davidwendt](https://github.com/davidwendt) +- Allow users to specify data types for a subset of columns in `read_csv` ([#10484](https://github.com/NVIDIA/cudf/pull/10484)) [@vuule](https://github.com/vuule) +- Fix default value of str.split expand parameter. ([#10457](https://github.com/NVIDIA/cudf/pull/10457)) [@bdice](https://github.com/bdice) +- Improve coverage of dask-cudf's groupby aggregation, add tests for `dropna` support ([#10449](https://github.com/NVIDIA/cudf/pull/10449)) [@charlesbluca](https://github.com/charlesbluca) +- Allow string aggs for `dask_cudf.CudfDataFrameGroupBy.aggregate` ([#10222](https://github.com/NVIDIA/cudf/pull/10222)) [@charlesbluca](https://github.com/charlesbluca) +- In-place updates with loc or iloc don't work correctly when the LHS has more than one column ([#9918](https://github.com/NVIDIA/cudf/pull/9918)) [@skirui-source](https://github.com/skirui-source) ## 📖 Documentation -- Clarify append deprecation notice. ([#10930](https://github.com/rapidsai/cudf/pull/10930)) [@bdice](https://github.com/bdice) -- Use full name of GPUDirect Storage SDK in docs ([#10904](https://github.com/rapidsai/cudf/pull/10904)) [@vuule](https://github.com/vuule) -- Update Dask + Pandas to Dask + cuDF path ([#10897](https://github.com/rapidsai/cudf/pull/10897)) [@miguelusque](https://github.com/miguelusque) -- Add missing documentation in cudf/types.hpp ([#10895](https://github.com/rapidsai/cudf/pull/10895)) [@karthikeyann](https://github.com/karthikeyann) -- Add strong index iterator docs. ([#10888](https://github.com/rapidsai/cudf/pull/10888)) [@bdice](https://github.com/bdice) -- spell check fixes ([#10865](https://github.com/rapidsai/cudf/pull/10865)) [@karthikeyann](https://github.com/karthikeyann) -- Add missing documentation in scalar/ headers ([#10861](https://github.com/rapidsai/cudf/pull/10861)) [@karthikeyann](https://github.com/karthikeyann) -- Remove typo in ngram documentation ([#10859](https://github.com/rapidsai/cudf/pull/10859)) [@miguelusque](https://github.com/miguelusque) -- fix doxygen warnings ([#10842](https://github.com/rapidsai/cudf/pull/10842)) [@karthikeyann](https://github.com/karthikeyann) -- Add a library_design.md file documenting the core Python data structures and their relationship ([#10817](https://github.com/rapidsai/cudf/pull/10817)) [@vyasr](https://github.com/vyasr) -- Add NumPy to intersphinx references. ([#10809](https://github.com/rapidsai/cudf/pull/10809)) [@bdice](https://github.com/bdice) -- Add a section to the docs that compares cuDF with Pandas ([#10796](https://github.com/rapidsai/cudf/pull/10796)) [@shwina](https://github.com/shwina) -- Mention 2 cpp-reviewer requirement in pull request template ([#10768](https://github.com/rapidsai/cudf/pull/10768)) [@davidwendt](https://github.com/davidwendt) -- Enable pydocstyle for all packages. ([#10759](https://github.com/rapidsai/cudf/pull/10759)) [@bdice](https://github.com/bdice) -- Enable pydocstyle rules involving quotes ([#10748](https://github.com/rapidsai/cudf/pull/10748)) [@vyasr](https://github.com/vyasr) -- Revise 10 minutes notebook. ([#10738](https://github.com/rapidsai/cudf/pull/10738)) [@bdice](https://github.com/bdice) -- Reorganize cuDF Python docs ([#10691](https://github.com/rapidsai/cudf/pull/10691)) [@shwina](https://github.com/shwina) -- Fix sphinx/jupyter heading issue in UDF notebook ([#10690](https://github.com/rapidsai/cudf/pull/10690)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Migrated user guide notebooks to MyST-NB and added sphinx extension ([#10685](https://github.com/rapidsai/cudf/pull/10685)) [@mmccarty](https://github.com/mmccarty) -- add data generation to benchmark documentation ([#10677](https://github.com/rapidsai/cudf/pull/10677)) [@karthikeyann](https://github.com/karthikeyann) -- Fix some docs build warnings ([#10674](https://github.com/rapidsai/cudf/pull/10674)) [@galipremsagar](https://github.com/galipremsagar) -- Update UDF notebook in User Guide. ([#10668](https://github.com/rapidsai/cudf/pull/10668)) [@bdice](https://github.com/bdice) -- Improve User Guide docs ([#10663](https://github.com/rapidsai/cudf/pull/10663)) [@bdice](https://github.com/bdice) -- Fix some docstrings formatting ([#10660](https://github.com/rapidsai/cudf/pull/10660)) [@galipremsagar](https://github.com/galipremsagar) -- Remove implementation details from `apply` docstrings ([#10651](https://github.com/rapidsai/cudf/pull/10651)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Revise CONTRIBUTING.md ([#10644](https://github.com/rapidsai/cudf/pull/10644)) [@bdice](https://github.com/bdice) -- Add missing APIs to documentation. ([#10643](https://github.com/rapidsai/cudf/pull/10643)) [@bdice](https://github.com/bdice) -- Use cudf.read_json as documented API name. ([#10640](https://github.com/rapidsai/cudf/pull/10640)) [@bdice](https://github.com/bdice) -- Fix docstring section headings. ([#10639](https://github.com/rapidsai/cudf/pull/10639)) [@bdice](https://github.com/bdice) -- Document cudf.read_text and cudf.read_avro. ([#10638](https://github.com/rapidsai/cudf/pull/10638)) [@bdice](https://github.com/bdice) -- Fix type-o in docstring for json_reader_options ([#10627](https://github.com/rapidsai/cudf/pull/10627)) [@dagardner-nv](https://github.com/dagardner-nv) -- Update guide to UDFs with notes about `Series.applymap` deprecation and related changes ([#10607](https://github.com/rapidsai/cudf/pull/10607)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix doxygen Modules page for cudf::lists::sequences ([#10561](https://github.com/rapidsai/cudf/pull/10561)) [@davidwendt](https://github.com/davidwendt) -- Add Replace Backreferences section to Regex Features page ([#10560](https://github.com/rapidsai/cudf/pull/10560)) [@davidwendt](https://github.com/davidwendt) -- Introduce deprecation policy to developer guide. ([#10252](https://github.com/rapidsai/cudf/pull/10252)) [@vyasr](https://github.com/vyasr) +- Clarify append deprecation notice. ([#10930](https://github.com/NVIDIA/cudf/pull/10930)) [@bdice](https://github.com/bdice) +- Use full name of GPUDirect Storage SDK in docs ([#10904](https://github.com/NVIDIA/cudf/pull/10904)) [@vuule](https://github.com/vuule) +- Update Dask + Pandas to Dask + cuDF path ([#10897](https://github.com/NVIDIA/cudf/pull/10897)) [@miguelusque](https://github.com/miguelusque) +- Add missing documentation in cudf/types.hpp ([#10895](https://github.com/NVIDIA/cudf/pull/10895)) [@karthikeyann](https://github.com/karthikeyann) +- Add strong index iterator docs. ([#10888](https://github.com/NVIDIA/cudf/pull/10888)) [@bdice](https://github.com/bdice) +- spell check fixes ([#10865](https://github.com/NVIDIA/cudf/pull/10865)) [@karthikeyann](https://github.com/karthikeyann) +- Add missing documentation in scalar/ headers ([#10861](https://github.com/NVIDIA/cudf/pull/10861)) [@karthikeyann](https://github.com/karthikeyann) +- Remove typo in ngram documentation ([#10859](https://github.com/NVIDIA/cudf/pull/10859)) [@miguelusque](https://github.com/miguelusque) +- fix doxygen warnings ([#10842](https://github.com/NVIDIA/cudf/pull/10842)) [@karthikeyann](https://github.com/karthikeyann) +- Add a library_design.md file documenting the core Python data structures and their relationship ([#10817](https://github.com/NVIDIA/cudf/pull/10817)) [@vyasr](https://github.com/vyasr) +- Add NumPy to intersphinx references. ([#10809](https://github.com/NVIDIA/cudf/pull/10809)) [@bdice](https://github.com/bdice) +- Add a section to the docs that compares cuDF with Pandas ([#10796](https://github.com/NVIDIA/cudf/pull/10796)) [@shwina](https://github.com/shwina) +- Mention 2 cpp-reviewer requirement in pull request template ([#10768](https://github.com/NVIDIA/cudf/pull/10768)) [@davidwendt](https://github.com/davidwendt) +- Enable pydocstyle for all packages. ([#10759](https://github.com/NVIDIA/cudf/pull/10759)) [@bdice](https://github.com/bdice) +- Enable pydocstyle rules involving quotes ([#10748](https://github.com/NVIDIA/cudf/pull/10748)) [@vyasr](https://github.com/vyasr) +- Revise 10 minutes notebook. ([#10738](https://github.com/NVIDIA/cudf/pull/10738)) [@bdice](https://github.com/bdice) +- Reorganize cuDF Python docs ([#10691](https://github.com/NVIDIA/cudf/pull/10691)) [@shwina](https://github.com/shwina) +- Fix sphinx/jupyter heading issue in UDF notebook ([#10690](https://github.com/NVIDIA/cudf/pull/10690)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Migrated user guide notebooks to MyST-NB and added sphinx extension ([#10685](https://github.com/NVIDIA/cudf/pull/10685)) [@mmccarty](https://github.com/mmccarty) +- add data generation to benchmark documentation ([#10677](https://github.com/NVIDIA/cudf/pull/10677)) [@karthikeyann](https://github.com/karthikeyann) +- Fix some docs build warnings ([#10674](https://github.com/NVIDIA/cudf/pull/10674)) [@galipremsagar](https://github.com/galipremsagar) +- Update UDF notebook in User Guide. ([#10668](https://github.com/NVIDIA/cudf/pull/10668)) [@bdice](https://github.com/bdice) +- Improve User Guide docs ([#10663](https://github.com/NVIDIA/cudf/pull/10663)) [@bdice](https://github.com/bdice) +- Fix some docstrings formatting ([#10660](https://github.com/NVIDIA/cudf/pull/10660)) [@galipremsagar](https://github.com/galipremsagar) +- Remove implementation details from `apply` docstrings ([#10651](https://github.com/NVIDIA/cudf/pull/10651)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Revise CONTRIBUTING.md ([#10644](https://github.com/NVIDIA/cudf/pull/10644)) [@bdice](https://github.com/bdice) +- Add missing APIs to documentation. ([#10643](https://github.com/NVIDIA/cudf/pull/10643)) [@bdice](https://github.com/bdice) +- Use cudf.read_json as documented API name. ([#10640](https://github.com/NVIDIA/cudf/pull/10640)) [@bdice](https://github.com/bdice) +- Fix docstring section headings. ([#10639](https://github.com/NVIDIA/cudf/pull/10639)) [@bdice](https://github.com/bdice) +- Document cudf.read_text and cudf.read_avro. ([#10638](https://github.com/NVIDIA/cudf/pull/10638)) [@bdice](https://github.com/bdice) +- Fix type-o in docstring for json_reader_options ([#10627](https://github.com/NVIDIA/cudf/pull/10627)) [@dagardner-nv](https://github.com/dagardner-nv) +- Update guide to UDFs with notes about `Series.applymap` deprecation and related changes ([#10607](https://github.com/NVIDIA/cudf/pull/10607)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix doxygen Modules page for cudf::lists::sequences ([#10561](https://github.com/NVIDIA/cudf/pull/10561)) [@davidwendt](https://github.com/davidwendt) +- Add Replace Backreferences section to Regex Features page ([#10560](https://github.com/NVIDIA/cudf/pull/10560)) [@davidwendt](https://github.com/davidwendt) +- Introduce deprecation policy to developer guide. ([#10252](https://github.com/NVIDIA/cudf/pull/10252)) [@vyasr](https://github.com/vyasr) ## 🚀 New Features -- Enable Zstandard decompression only when all nvcomp integrations are enabled ([#10944](https://github.com/rapidsai/cudf/pull/10944)) [@vuule](https://github.com/vuule) -- Handle nested types in cudf::concatenate_rows() ([#10890](https://github.com/rapidsai/cudf/pull/10890)) [@nvdbaranec](https://github.com/nvdbaranec) -- Strong index types for equality comparator ([#10883](https://github.com/rapidsai/cudf/pull/10883)) [@ttnghia](https://github.com/ttnghia) -- Add parameters to control page size in Parquet writer ([#10882](https://github.com/rapidsai/cudf/pull/10882)) [@etseidl](https://github.com/etseidl) -- Support for Zstandard decompression in ORC reader ([#10873](https://github.com/rapidsai/cudf/pull/10873)) [@vuule](https://github.com/vuule) -- Use pre-built nvcomp 2.3 binaries by default ([#10851](https://github.com/rapidsai/cudf/pull/10851)) [@robertmaynard](https://github.com/robertmaynard) -- Support for Zstandard decompression in Parquet reader ([#10847](https://github.com/rapidsai/cudf/pull/10847)) [@vuule](https://github.com/vuule) -- Add JNI support for apply_boolean_mask ([#10812](https://github.com/rapidsai/cudf/pull/10812)) [@res-life](https://github.com/res-life) -- Segmented Min/Max for Fixed Point Types ([#10794](https://github.com/rapidsai/cudf/pull/10794)) [@isVoid](https://github.com/isVoid) -- Return per-file metadata from readers ([#10782](https://github.com/rapidsai/cudf/pull/10782)) [@vuule](https://github.com/vuule) -- Segmented `apply_boolean_mask` for `LIST` columns ([#10773](https://github.com/rapidsai/cudf/pull/10773)) [@mythrocks](https://github.com/mythrocks) -- Update `groupby::hash` to use new row operators for keys ([#10770](https://github.com/rapidsai/cudf/pull/10770)) [@PointKernel](https://github.com/PointKernel) -- Support purging non-empty null elements from LIST/STRING columns ([#10701](https://github.com/rapidsai/cudf/pull/10701)) [@mythrocks](https://github.com/mythrocks) -- Add `detail::hash_join` ([#10695](https://github.com/rapidsai/cudf/pull/10695)) [@PointKernel](https://github.com/PointKernel) -- Persist string statistics data across multiple calls to orc chunked write ([#10694](https://github.com/rapidsai/cudf/pull/10694)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add `.list.astype()` to cast list leaves to specified dtype ([#10693](https://github.com/rapidsai/cudf/pull/10693)) [@shwina](https://github.com/shwina) -- JNI: Add generateListOffsets API ([#10683](https://github.com/rapidsai/cudf/pull/10683)) [@sperlingxx](https://github.com/sperlingxx) -- Support `args` in groupby apply ([#10682](https://github.com/rapidsai/cudf/pull/10682)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Enable segmented_gather in Java package ([#10669](https://github.com/rapidsai/cudf/pull/10669)) [@sperlingxx](https://github.com/sperlingxx) -- Add row hasher with nested column support ([#10641](https://github.com/rapidsai/cudf/pull/10641)) [@devavret](https://github.com/devavret) -- Add support for numeric_only in DataFrame._reduce ([#10629](https://github.com/rapidsai/cudf/pull/10629)) [@martinfalisse](https://github.com/martinfalisse) -- First step toward statistics in ORC files with chunked writes ([#10567](https://github.com/rapidsai/cudf/pull/10567)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add support for struct columns to the random table generator ([#10566](https://github.com/rapidsai/cudf/pull/10566)) [@vuule](https://github.com/vuule) -- Enable passing a sequence for the `index` argument to `.list.get()` ([#10564](https://github.com/rapidsai/cudf/pull/10564)) [@shwina](https://github.com/shwina) -- Add python bindings for cudf::list::index_of ([#10549](https://github.com/rapidsai/cudf/pull/10549)) [@ChrisJar](https://github.com/ChrisJar) -- Add default= kwarg to .list.get() accessor method ([#10547](https://github.com/rapidsai/cudf/pull/10547)) [@shwina](https://github.com/shwina) -- Add `cudf.DataFrame.applymap` ([#10542](https://github.com/rapidsai/cudf/pull/10542)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Support nvComp 2.3 if local, otherwise use nvcomp 2.2 ([#10513](https://github.com/rapidsai/cudf/pull/10513)) [@robertmaynard](https://github.com/robertmaynard) -- Add column field ID control in parquet writer ([#10504](https://github.com/rapidsai/cudf/pull/10504)) [@PointKernel](https://github.com/PointKernel) -- Deprecate `Series.applymap` ([#10497](https://github.com/rapidsai/cudf/pull/10497)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add option to drop cache in cuIO benchmarks ([#10488](https://github.com/rapidsai/cudf/pull/10488)) [@vuule](https://github.com/vuule) -- move benchmark input generation in device in reduction nvbench ([#10486](https://github.com/rapidsai/cudf/pull/10486)) [@karthikeyann](https://github.com/karthikeyann) -- Support Segmented Min/Max Reduction on String Type ([#10447](https://github.com/rapidsai/cudf/pull/10447)) [@isVoid](https://github.com/isVoid) -- List element Equality comparator ([#10289](https://github.com/rapidsai/cudf/pull/10289)) [@devavret](https://github.com/devavret) -- Implement all methods of groupby rank aggregation in libcudf, python ([#9569](https://github.com/rapidsai/cudf/pull/9569)) [@karthikeyann](https://github.com/karthikeyann) -- Implement DataFrame.eval using libcudf ASTs ([#8022](https://github.com/rapidsai/cudf/pull/8022)) [@vyasr](https://github.com/vyasr) +- Enable Zstandard decompression only when all nvcomp integrations are enabled ([#10944](https://github.com/NVIDIA/cudf/pull/10944)) [@vuule](https://github.com/vuule) +- Handle nested types in cudf::concatenate_rows() ([#10890](https://github.com/NVIDIA/cudf/pull/10890)) [@nvdbaranec](https://github.com/nvdbaranec) +- Strong index types for equality comparator ([#10883](https://github.com/NVIDIA/cudf/pull/10883)) [@ttnghia](https://github.com/ttnghia) +- Add parameters to control page size in Parquet writer ([#10882](https://github.com/NVIDIA/cudf/pull/10882)) [@etseidl](https://github.com/etseidl) +- Support for Zstandard decompression in ORC reader ([#10873](https://github.com/NVIDIA/cudf/pull/10873)) [@vuule](https://github.com/vuule) +- Use pre-built nvcomp 2.3 binaries by default ([#10851](https://github.com/NVIDIA/cudf/pull/10851)) [@robertmaynard](https://github.com/robertmaynard) +- Support for Zstandard decompression in Parquet reader ([#10847](https://github.com/NVIDIA/cudf/pull/10847)) [@vuule](https://github.com/vuule) +- Add JNI support for apply_boolean_mask ([#10812](https://github.com/NVIDIA/cudf/pull/10812)) [@res-life](https://github.com/res-life) +- Segmented Min/Max for Fixed Point Types ([#10794](https://github.com/NVIDIA/cudf/pull/10794)) [@isVoid](https://github.com/isVoid) +- Return per-file metadata from readers ([#10782](https://github.com/NVIDIA/cudf/pull/10782)) [@vuule](https://github.com/vuule) +- Segmented `apply_boolean_mask` for `LIST` columns ([#10773](https://github.com/NVIDIA/cudf/pull/10773)) [@mythrocks](https://github.com/mythrocks) +- Update `groupby::hash` to use new row operators for keys ([#10770](https://github.com/NVIDIA/cudf/pull/10770)) [@PointKernel](https://github.com/PointKernel) +- Support purging non-empty null elements from LIST/STRING columns ([#10701](https://github.com/NVIDIA/cudf/pull/10701)) [@mythrocks](https://github.com/mythrocks) +- Add `detail::hash_join` ([#10695](https://github.com/NVIDIA/cudf/pull/10695)) [@PointKernel](https://github.com/PointKernel) +- Persist string statistics data across multiple calls to orc chunked write ([#10694](https://github.com/NVIDIA/cudf/pull/10694)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add `.list.astype()` to cast list leaves to specified dtype ([#10693](https://github.com/NVIDIA/cudf/pull/10693)) [@shwina](https://github.com/shwina) +- JNI: Add generateListOffsets API ([#10683](https://github.com/NVIDIA/cudf/pull/10683)) [@sperlingxx](https://github.com/sperlingxx) +- Support `args` in groupby apply ([#10682](https://github.com/NVIDIA/cudf/pull/10682)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Enable segmented_gather in Java package ([#10669](https://github.com/NVIDIA/cudf/pull/10669)) [@sperlingxx](https://github.com/sperlingxx) +- Add row hasher with nested column support ([#10641](https://github.com/NVIDIA/cudf/pull/10641)) [@devavret](https://github.com/devavret) +- Add support for numeric_only in DataFrame._reduce ([#10629](https://github.com/NVIDIA/cudf/pull/10629)) [@martinfalisse](https://github.com/martinfalisse) +- First step toward statistics in ORC files with chunked writes ([#10567](https://github.com/NVIDIA/cudf/pull/10567)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add support for struct columns to the random table generator ([#10566](https://github.com/NVIDIA/cudf/pull/10566)) [@vuule](https://github.com/vuule) +- Enable passing a sequence for the `index` argument to `.list.get()` ([#10564](https://github.com/NVIDIA/cudf/pull/10564)) [@shwina](https://github.com/shwina) +- Add python bindings for cudf::list::index_of ([#10549](https://github.com/NVIDIA/cudf/pull/10549)) [@ChrisJar](https://github.com/ChrisJar) +- Add default= kwarg to .list.get() accessor method ([#10547](https://github.com/NVIDIA/cudf/pull/10547)) [@shwina](https://github.com/shwina) +- Add `cudf.DataFrame.applymap` ([#10542](https://github.com/NVIDIA/cudf/pull/10542)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Support nvComp 2.3 if local, otherwise use nvcomp 2.2 ([#10513](https://github.com/NVIDIA/cudf/pull/10513)) [@robertmaynard](https://github.com/robertmaynard) +- Add column field ID control in parquet writer ([#10504](https://github.com/NVIDIA/cudf/pull/10504)) [@PointKernel](https://github.com/PointKernel) +- Deprecate `Series.applymap` ([#10497](https://github.com/NVIDIA/cudf/pull/10497)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add option to drop cache in cuIO benchmarks ([#10488](https://github.com/NVIDIA/cudf/pull/10488)) [@vuule](https://github.com/vuule) +- move benchmark input generation in device in reduction nvbench ([#10486](https://github.com/NVIDIA/cudf/pull/10486)) [@karthikeyann](https://github.com/karthikeyann) +- Support Segmented Min/Max Reduction on String Type ([#10447](https://github.com/NVIDIA/cudf/pull/10447)) [@isVoid](https://github.com/isVoid) +- List element Equality comparator ([#10289](https://github.com/NVIDIA/cudf/pull/10289)) [@devavret](https://github.com/devavret) +- Implement all methods of groupby rank aggregation in libcudf, python ([#9569](https://github.com/NVIDIA/cudf/pull/9569)) [@karthikeyann](https://github.com/karthikeyann) +- Implement DataFrame.eval using libcudf ASTs ([#8022](https://github.com/NVIDIA/cudf/pull/8022)) [@vyasr](https://github.com/vyasr) ## 🛠️ Improvements -- Use `conda` compilers in env file ([#10915](https://github.com/rapidsai/cudf/pull/10915)) [@galipremsagar](https://github.com/galipremsagar) -- Remove C style artifacts in cuIO ([#10886](https://github.com/rapidsai/cudf/pull/10886)) [@vuule](https://github.com/vuule) -- Rename `sliced_child` to `get_sliced_child`. ([#10885](https://github.com/rapidsai/cudf/pull/10885)) [@bdice](https://github.com/bdice) -- Replace defaulted stream value for libcudf APIs that use NVCOMP ([#10877](https://github.com/rapidsai/cudf/pull/10877)) [@jbrennan333](https://github.com/jbrennan333) -- Add more unit tests for `cudf::distinct` for nested types with sliced input ([#10860](https://github.com/rapidsai/cudf/pull/10860)) [@ttnghia](https://github.com/ttnghia) -- Changing `list_view.cuh` to `list_view.hpp` ([#10854](https://github.com/rapidsai/cudf/pull/10854)) [@ttnghia](https://github.com/ttnghia) -- More error checking in `from_dlpack` ([#10850](https://github.com/rapidsai/cudf/pull/10850)) [@wence-](https://github.com/wence-) -- Cleanup regex compiler fixed quantifiers source ([#10843](https://github.com/rapidsai/cudf/pull/10843)) [@davidwendt](https://github.com/davidwendt) -- Adds the JNI call for Cuda.deviceSynchronize ([#10839](https://github.com/rapidsai/cudf/pull/10839)) [@abellina](https://github.com/abellina) -- Add missing cuda-python dependency to cudf ([#10833](https://github.com/rapidsai/cudf/pull/10833)) [@bdice](https://github.com/bdice) -- Change std::string parameters in cudf::strings APIs to std::string_view ([#10832](https://github.com/rapidsai/cudf/pull/10832)) [@davidwendt](https://github.com/davidwendt) -- Split up search.cu to improve compile time ([#10831](https://github.com/rapidsai/cudf/pull/10831)) [@davidwendt](https://github.com/davidwendt) -- Add tests for null scalar binaryops ([#10828](https://github.com/rapidsai/cudf/pull/10828)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Cleanup regex compile optimize functions ([#10825](https://github.com/rapidsai/cudf/pull/10825)) [@davidwendt](https://github.com/davidwendt) -- Use `ThreadedMotoServer` instead of `subprocess` in spinning up `s3` server ([#10822](https://github.com/rapidsai/cudf/pull/10822)) [@galipremsagar](https://github.com/galipremsagar) -- Import `NA` from `missing` rather than using `cudf.NA` everywhere ([#10821](https://github.com/rapidsai/cudf/pull/10821)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Refactor regex builtin character-class identifiers ([#10814](https://github.com/rapidsai/cudf/pull/10814)) [@davidwendt](https://github.com/davidwendt) -- Change pattern parameter for regex APIs from std::string to std::string_view ([#10810](https://github.com/rapidsai/cudf/pull/10810)) [@davidwendt](https://github.com/davidwendt) -- Make the JNI API to get list offsets as a view public. ([#10807](https://github.com/rapidsai/cudf/pull/10807)) [@revans2](https://github.com/revans2) -- Add cudf JNI docker build github action ([#10806](https://github.com/rapidsai/cudf/pull/10806)) [@pxLi](https://github.com/pxLi) -- Removed `mr` parameter from inplace bitmask operations ([#10805](https://github.com/rapidsai/cudf/pull/10805)) [@AtlantaPepsi](https://github.com/AtlantaPepsi) -- Refactor `cudf::contains`, renaming and switching parameters role ([#10802](https://github.com/rapidsai/cudf/pull/10802)) [@ttnghia](https://github.com/ttnghia) -- Handle closed property in IntervalDtype.from_pandas ([#10798](https://github.com/rapidsai/cudf/pull/10798)) [@wence-](https://github.com/wence-) -- Return weak orderings from `device_row_comparator`. ([#10793](https://github.com/rapidsai/cudf/pull/10793)) [@rwlee](https://github.com/rwlee) -- Rework `Scalar` imports ([#10791](https://github.com/rapidsai/cudf/pull/10791)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Enable ccache for cudfjni build in Docker ([#10790](https://github.com/rapidsai/cudf/pull/10790)) [@gerashegalov](https://github.com/gerashegalov) -- Generic serialization of all column types ([#10784](https://github.com/rapidsai/cudf/pull/10784)) [@wence-](https://github.com/wence-) -- simplifying skiprows test in test_orc.py ([#10783](https://github.com/rapidsai/cudf/pull/10783)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Use column_views instead of column_device_views in binary operations. ([#10780](https://github.com/rapidsai/cudf/pull/10780)) [@bdice](https://github.com/bdice) -- Add struct utility functions. ([#10776](https://github.com/rapidsai/cudf/pull/10776)) [@bdice](https://github.com/bdice) -- Add multiple rows to subword tokenizer benchmark ([#10767](https://github.com/rapidsai/cudf/pull/10767)) [@davidwendt](https://github.com/davidwendt) -- Refactor host decompression in ORC reader ([#10764](https://github.com/rapidsai/cudf/pull/10764)) [@vuule](https://github.com/vuule) -- Flush output streams before creating a process to drop caches ([#10762](https://github.com/rapidsai/cudf/pull/10762)) [@vuule](https://github.com/vuule) -- Refactor binaryop/compiled/util.cpp ([#10756](https://github.com/rapidsai/cudf/pull/10756)) [@bdice](https://github.com/bdice) -- Use warp per string for long strings in cudf::strings::contains() ([#10739](https://github.com/rapidsai/cudf/pull/10739)) [@davidwendt](https://github.com/davidwendt) -- Use generator expressions in any/all functions. ([#10736](https://github.com/rapidsai/cudf/pull/10736)) [@bdice](https://github.com/bdice) -- Use canonical "magic methods" (replace `x.__repr__()` with `repr(x)`). ([#10735](https://github.com/rapidsai/cudf/pull/10735)) [@bdice](https://github.com/bdice) -- Improve use of isinstance. ([#10734](https://github.com/rapidsai/cudf/pull/10734)) [@bdice](https://github.com/bdice) -- Rename tests from multiIndex to multiindex. ([#10732](https://github.com/rapidsai/cudf/pull/10732)) [@bdice](https://github.com/bdice) -- Two-table comparators with strong index types ([#10730](https://github.com/rapidsai/cudf/pull/10730)) [@bdice](https://github.com/bdice) -- Replace std::make_pair with std::pair (C++17 CTAD) ([#10727](https://github.com/rapidsai/cudf/pull/10727)) [@karthikeyann](https://github.com/karthikeyann) -- Use structured bindings instead of std::tie ([#10726](https://github.com/rapidsai/cudf/pull/10726)) [@karthikeyann](https://github.com/karthikeyann) -- Missing `f` prefix on f-strings fix ([#10721](https://github.com/rapidsai/cudf/pull/10721)) [@code-review-doctor](https://github.com/code-review-doctor) -- Add `max_file_size` parameter to chunked parquet dataset writer ([#10718](https://github.com/rapidsai/cudf/pull/10718)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `merge_sorted`, change dask cudf usage to internal method ([#10713](https://github.com/rapidsai/cudf/pull/10713)) [@isVoid](https://github.com/isVoid) -- Prepare dask_cudf test_parquet.py for upcoming API changes ([#10709](https://github.com/rapidsai/cudf/pull/10709)) [@rjzamora](https://github.com/rjzamora) -- Remove or simplify various utility functions ([#10705](https://github.com/rapidsai/cudf/pull/10705)) [@vyasr](https://github.com/vyasr) -- Allow building arrow with parquet and not python ([#10702](https://github.com/rapidsai/cudf/pull/10702)) [@revans2](https://github.com/revans2) -- Partial cuIO GPU decompression refactor ([#10699](https://github.com/rapidsai/cudf/pull/10699)) [@vuule](https://github.com/vuule) -- Cython API refactor: `merge.pyx` ([#10698](https://github.com/rapidsai/cudf/pull/10698)) [@isVoid](https://github.com/isVoid) -- Fix random string data length to become variable ([#10697](https://github.com/rapidsai/cudf/pull/10697)) [@galipremsagar](https://github.com/galipremsagar) -- Add bindings for index_of with column search key ([#10696](https://github.com/rapidsai/cudf/pull/10696)) [@ChrisJar](https://github.com/ChrisJar) -- Deprecate index merging ([#10689](https://github.com/rapidsai/cudf/pull/10689)) [@vyasr](https://github.com/vyasr) -- Remove cudf::strings::string namespace ([#10684](https://github.com/rapidsai/cudf/pull/10684)) [@davidwendt](https://github.com/davidwendt) -- Standardize imports. ([#10680](https://github.com/rapidsai/cudf/pull/10680)) [@bdice](https://github.com/bdice) -- Standardize usage of collections.abc. ([#10679](https://github.com/rapidsai/cudf/pull/10679)) [@bdice](https://github.com/bdice) -- Cython API Refactor: `transpose.pyx`, `sort.pyx` ([#10675](https://github.com/rapidsai/cudf/pull/10675)) [@isVoid](https://github.com/isVoid) -- Add device_memory_resource parameter to create_string_vector_from_column ([#10673](https://github.com/rapidsai/cudf/pull/10673)) [@davidwendt](https://github.com/davidwendt) -- Split up mixed-join kernels source files ([#10671](https://github.com/rapidsai/cudf/pull/10671)) [@davidwendt](https://github.com/davidwendt) -- Use `std::filesystem` for temporary directory location and deletion ([#10664](https://github.com/rapidsai/cudf/pull/10664)) [@vuule](https://github.com/vuule) -- cleanup benchmark includes ([#10661](https://github.com/rapidsai/cudf/pull/10661)) [@karthikeyann](https://github.com/karthikeyann) -- Use upstream clang-format pre-commit hook. ([#10659](https://github.com/rapidsai/cudf/pull/10659)) [@bdice](https://github.com/bdice) -- Clean up C++ includes to use <> instead of "". ([#10658](https://github.com/rapidsai/cudf/pull/10658)) [@bdice](https://github.com/bdice) -- Handle RuntimeError thrown by CUDA Python in `validate_setup` ([#10653](https://github.com/rapidsai/cudf/pull/10653)) [@shwina](https://github.com/shwina) -- Rework JNI CMake to leverage rapids_find_package ([#10649](https://github.com/rapidsai/cudf/pull/10649)) [@jlowe](https://github.com/jlowe) -- Use conda to build python packages during GPU tests ([#10648](https://github.com/rapidsai/cudf/pull/10648)) [@Ethyling](https://github.com/Ethyling) -- Deprecate various functions that don't need to be defined for Index. ([#10647](https://github.com/rapidsai/cudf/pull/10647)) [@vyasr](https://github.com/vyasr) -- Update pinning to allow newer CMake versions. ([#10646](https://github.com/rapidsai/cudf/pull/10646)) [@vyasr](https://github.com/vyasr) -- Bump hadoop-common from 3.1.4 to 3.2.3 in /java ([#10645](https://github.com/rapidsai/cudf/pull/10645)) [@dependabot[bot]](https://github.com/dependabot[bot]) -- Remove `concurrent_unordered_multimap`. ([#10642](https://github.com/rapidsai/cudf/pull/10642)) [@bdice](https://github.com/bdice) -- Improve parquet dictionary encoding ([#10635](https://github.com/rapidsai/cudf/pull/10635)) [@PointKernel](https://github.com/PointKernel) -- Improve cudf::cuda_error ([#10630](https://github.com/rapidsai/cudf/pull/10630)) [@sperlingxx](https://github.com/sperlingxx) -- Add support for null and non-numeric types in Series.diff and DataFrame.diff ([#10625](https://github.com/rapidsai/cudf/pull/10625)) [@Matt711](https://github.com/Matt711) -- Branch 22.06 merge 22.04 ([#10624](https://github.com/rapidsai/cudf/pull/10624)) [@vyasr](https://github.com/vyasr) -- Unpin `dask` & `distributed` for development ([#10623](https://github.com/rapidsai/cudf/pull/10623)) [@galipremsagar](https://github.com/galipremsagar) -- Slightly improve accuracy of stod in to_floats ([#10622](https://github.com/rapidsai/cudf/pull/10622)) [@davidwendt](https://github.com/davidwendt) -- Allow libcudfjni to be built as a static library ([#10619](https://github.com/rapidsai/cudf/pull/10619)) [@jlowe](https://github.com/jlowe) -- Change stack-based regex state data to use global memory ([#10600](https://github.com/rapidsai/cudf/pull/10600)) [@davidwendt](https://github.com/davidwendt) -- Resolve Forward merging of `branch-22.04` into `branch-22.06` ([#10598](https://github.com/rapidsai/cudf/pull/10598)) [@galipremsagar](https://github.com/galipremsagar) -- KvikIO as an alternative GDS backend ([#10593](https://github.com/rapidsai/cudf/pull/10593)) [@madsbk](https://github.com/madsbk) -- Rename CUDA_TRY macro to CUDF_CUDA_TRY, rename CHECK_CUDA macro to CUDF_CHECK_CUDA. ([#10589](https://github.com/rapidsai/cudf/pull/10589)) [@bdice](https://github.com/bdice) -- Upgrade `cudf` to support `pandas` 1.4.x versions ([#10584](https://github.com/rapidsai/cudf/pull/10584)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor binary ops for timedelta and datetime columns ([#10581](https://github.com/rapidsai/cudf/pull/10581)) [@vyasr](https://github.com/vyasr) -- Refactor cudf::strings::count_re API to use count_matches utility ([#10580](https://github.com/rapidsai/cudf/pull/10580)) [@davidwendt](https://github.com/davidwendt) -- Update `Programming Language :: Python` Versions to 3.8 & 3.9 ([#10579](https://github.com/rapidsai/cudf/pull/10579)) [@madsbk](https://github.com/madsbk) -- Automate Java cudf jar build with statically linked dependencies ([#10578](https://github.com/rapidsai/cudf/pull/10578)) [@gerashegalov](https://github.com/gerashegalov) -- Add patch for thrust-cub 1.16 to fix sort compile times ([#10577](https://github.com/rapidsai/cudf/pull/10577)) [@davidwendt](https://github.com/davidwendt) -- Move binop methods from Frame to IndexedFrame and standardize the docstring ([#10576](https://github.com/rapidsai/cudf/pull/10576)) [@vyasr](https://github.com/vyasr) -- Cleanup libcudf strings regex classes ([#10573](https://github.com/rapidsai/cudf/pull/10573)) [@davidwendt](https://github.com/davidwendt) -- Simplify preprocessing of arguments for DataFrame binops ([#10563](https://github.com/rapidsai/cudf/pull/10563)) [@vyasr](https://github.com/vyasr) -- Reduce kernel calls to build strings findall results ([#10559](https://github.com/rapidsai/cudf/pull/10559)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-22.04 to branch-22.06 ([#10557](https://github.com/rapidsai/cudf/pull/10557)) [@bdice](https://github.com/bdice) -- Update strings contains benchmark to measure varying match rates ([#10555](https://github.com/rapidsai/cudf/pull/10555)) [@davidwendt](https://github.com/davidwendt) -- JNI: throw CUDA errors more specifically ([#10551](https://github.com/rapidsai/cudf/pull/10551)) [@sperlingxx](https://github.com/sperlingxx) -- Enable building static libs ([#10545](https://github.com/rapidsai/cudf/pull/10545)) [@trxcllnt](https://github.com/trxcllnt) -- Remove pip requirements files. ([#10543](https://github.com/rapidsai/cudf/pull/10543)) [@bdice](https://github.com/bdice) -- Remove Click pinnings that are unnecessary after upgrading black. ([#10541](https://github.com/rapidsai/cudf/pull/10541)) [@vyasr](https://github.com/vyasr) -- Refactor `memory_usage` to improve performance ([#10537](https://github.com/rapidsai/cudf/pull/10537)) [@galipremsagar](https://github.com/galipremsagar) -- Adjust the valid range of group index for replace_with_backrefs ([#10530](https://github.com/rapidsai/cudf/pull/10530)) [@sperlingxx](https://github.com/sperlingxx) -- add accidentally removed comment. ([#10526](https://github.com/rapidsai/cudf/pull/10526)) [@vyasr](https://github.com/vyasr) -- Update conda environment. ([#10525](https://github.com/rapidsai/cudf/pull/10525)) [@vyasr](https://github.com/vyasr) -- Remove ColumnBase.__getitem__ ([#10516](https://github.com/rapidsai/cudf/pull/10516)) [@vyasr](https://github.com/vyasr) -- Optimize `left_semi_join` by materializing the gather mask ([#10511](https://github.com/rapidsai/cudf/pull/10511)) [@cheinger](https://github.com/cheinger) -- Define proper binary operation APIs for columns ([#10509](https://github.com/rapidsai/cudf/pull/10509)) [@vyasr](https://github.com/vyasr) -- Upgrade `arrow-cpp` & `pyarrow` to `7.0.0` ([#10503](https://github.com/rapidsai/cudf/pull/10503)) [@galipremsagar](https://github.com/galipremsagar) -- Update to Thrust 1.16 ([#10489](https://github.com/rapidsai/cudf/pull/10489)) [@bdice](https://github.com/bdice) -- Namespace/Docstring Fixes for Reduction ([#10471](https://github.com/rapidsai/cudf/pull/10471)) [@isVoid](https://github.com/isVoid) -- Update cudfjni 22.06.0-SNAPSHOT ([#10467](https://github.com/rapidsai/cudf/pull/10467)) [@pxLi](https://github.com/pxLi) -- Use Lists of Columns for Various Files ([#10463](https://github.com/rapidsai/cudf/pull/10463)) [@isVoid](https://github.com/isVoid) -- Additional refactoring of hash functions ([#10462](https://github.com/rapidsai/cudf/pull/10462)) [@bdice](https://github.com/bdice) -- Fix Series.str.findall behavior for expand=False. ([#10459](https://github.com/rapidsai/cudf/pull/10459)) [@bdice](https://github.com/bdice) -- Remove deprecated code. ([#10450](https://github.com/rapidsai/cudf/pull/10450)) [@vyasr](https://github.com/vyasr) -- Update cmake-format version. ([#10440](https://github.com/rapidsai/cudf/pull/10440)) [@vyasr](https://github.com/vyasr) -- Consolidate C++ `conda` recipes and add `libcudf-tests` package ([#10326](https://github.com/rapidsai/cudf/pull/10326)) [@ajschmidt8](https://github.com/ajschmidt8) -- Use conda compilers ([#10275](https://github.com/rapidsai/cudf/pull/10275)) [@Ethyling](https://github.com/Ethyling) -- Add row bitmask as a `detail::hash_join` member ([#10248](https://github.com/rapidsai/cudf/pull/10248)) [@PointKernel](https://github.com/PointKernel) +- Use `conda` compilers in env file ([#10915](https://github.com/NVIDIA/cudf/pull/10915)) [@galipremsagar](https://github.com/galipremsagar) +- Remove C style artifacts in cuIO ([#10886](https://github.com/NVIDIA/cudf/pull/10886)) [@vuule](https://github.com/vuule) +- Rename `sliced_child` to `get_sliced_child`. ([#10885](https://github.com/NVIDIA/cudf/pull/10885)) [@bdice](https://github.com/bdice) +- Replace defaulted stream value for libcudf APIs that use NVCOMP ([#10877](https://github.com/NVIDIA/cudf/pull/10877)) [@jbrennan333](https://github.com/jbrennan333) +- Add more unit tests for `cudf::distinct` for nested types with sliced input ([#10860](https://github.com/NVIDIA/cudf/pull/10860)) [@ttnghia](https://github.com/ttnghia) +- Changing `list_view.cuh` to `list_view.hpp` ([#10854](https://github.com/NVIDIA/cudf/pull/10854)) [@ttnghia](https://github.com/ttnghia) +- More error checking in `from_dlpack` ([#10850](https://github.com/NVIDIA/cudf/pull/10850)) [@wence-](https://github.com/wence-) +- Cleanup regex compiler fixed quantifiers source ([#10843](https://github.com/NVIDIA/cudf/pull/10843)) [@davidwendt](https://github.com/davidwendt) +- Adds the JNI call for Cuda.deviceSynchronize ([#10839](https://github.com/NVIDIA/cudf/pull/10839)) [@abellina](https://github.com/abellina) +- Add missing cuda-python dependency to cudf ([#10833](https://github.com/NVIDIA/cudf/pull/10833)) [@bdice](https://github.com/bdice) +- Change std::string parameters in cudf::strings APIs to std::string_view ([#10832](https://github.com/NVIDIA/cudf/pull/10832)) [@davidwendt](https://github.com/davidwendt) +- Split up search.cu to improve compile time ([#10831](https://github.com/NVIDIA/cudf/pull/10831)) [@davidwendt](https://github.com/davidwendt) +- Add tests for null scalar binaryops ([#10828](https://github.com/NVIDIA/cudf/pull/10828)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Cleanup regex compile optimize functions ([#10825](https://github.com/NVIDIA/cudf/pull/10825)) [@davidwendt](https://github.com/davidwendt) +- Use `ThreadedMotoServer` instead of `subprocess` in spinning up `s3` server ([#10822](https://github.com/NVIDIA/cudf/pull/10822)) [@galipremsagar](https://github.com/galipremsagar) +- Import `NA` from `missing` rather than using `cudf.NA` everywhere ([#10821](https://github.com/NVIDIA/cudf/pull/10821)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Refactor regex builtin character-class identifiers ([#10814](https://github.com/NVIDIA/cudf/pull/10814)) [@davidwendt](https://github.com/davidwendt) +- Change pattern parameter for regex APIs from std::string to std::string_view ([#10810](https://github.com/NVIDIA/cudf/pull/10810)) [@davidwendt](https://github.com/davidwendt) +- Make the JNI API to get list offsets as a view public. ([#10807](https://github.com/NVIDIA/cudf/pull/10807)) [@revans2](https://github.com/revans2) +- Add cudf JNI docker build github action ([#10806](https://github.com/NVIDIA/cudf/pull/10806)) [@pxLi](https://github.com/pxLi) +- Removed `mr` parameter from inplace bitmask operations ([#10805](https://github.com/NVIDIA/cudf/pull/10805)) [@AtlantaPepsi](https://github.com/AtlantaPepsi) +- Refactor `cudf::contains`, renaming and switching parameters role ([#10802](https://github.com/NVIDIA/cudf/pull/10802)) [@ttnghia](https://github.com/ttnghia) +- Handle closed property in IntervalDtype.from_pandas ([#10798](https://github.com/NVIDIA/cudf/pull/10798)) [@wence-](https://github.com/wence-) +- Return weak orderings from `device_row_comparator`. ([#10793](https://github.com/NVIDIA/cudf/pull/10793)) [@rwlee](https://github.com/rwlee) +- Rework `Scalar` imports ([#10791](https://github.com/NVIDIA/cudf/pull/10791)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Enable ccache for cudfjni build in Docker ([#10790](https://github.com/NVIDIA/cudf/pull/10790)) [@gerashegalov](https://github.com/gerashegalov) +- Generic serialization of all column types ([#10784](https://github.com/NVIDIA/cudf/pull/10784)) [@wence-](https://github.com/wence-) +- simplifying skiprows test in test_orc.py ([#10783](https://github.com/NVIDIA/cudf/pull/10783)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Use column_views instead of column_device_views in binary operations. ([#10780](https://github.com/NVIDIA/cudf/pull/10780)) [@bdice](https://github.com/bdice) +- Add struct utility functions. ([#10776](https://github.com/NVIDIA/cudf/pull/10776)) [@bdice](https://github.com/bdice) +- Add multiple rows to subword tokenizer benchmark ([#10767](https://github.com/NVIDIA/cudf/pull/10767)) [@davidwendt](https://github.com/davidwendt) +- Refactor host decompression in ORC reader ([#10764](https://github.com/NVIDIA/cudf/pull/10764)) [@vuule](https://github.com/vuule) +- Flush output streams before creating a process to drop caches ([#10762](https://github.com/NVIDIA/cudf/pull/10762)) [@vuule](https://github.com/vuule) +- Refactor binaryop/compiled/util.cpp ([#10756](https://github.com/NVIDIA/cudf/pull/10756)) [@bdice](https://github.com/bdice) +- Use warp per string for long strings in cudf::strings::contains() ([#10739](https://github.com/NVIDIA/cudf/pull/10739)) [@davidwendt](https://github.com/davidwendt) +- Use generator expressions in any/all functions. ([#10736](https://github.com/NVIDIA/cudf/pull/10736)) [@bdice](https://github.com/bdice) +- Use canonical "magic methods" (replace `x.__repr__()` with `repr(x)`). ([#10735](https://github.com/NVIDIA/cudf/pull/10735)) [@bdice](https://github.com/bdice) +- Improve use of isinstance. ([#10734](https://github.com/NVIDIA/cudf/pull/10734)) [@bdice](https://github.com/bdice) +- Rename tests from multiIndex to multiindex. ([#10732](https://github.com/NVIDIA/cudf/pull/10732)) [@bdice](https://github.com/bdice) +- Two-table comparators with strong index types ([#10730](https://github.com/NVIDIA/cudf/pull/10730)) [@bdice](https://github.com/bdice) +- Replace std::make_pair with std::pair (C++17 CTAD) ([#10727](https://github.com/NVIDIA/cudf/pull/10727)) [@karthikeyann](https://github.com/karthikeyann) +- Use structured bindings instead of std::tie ([#10726](https://github.com/NVIDIA/cudf/pull/10726)) [@karthikeyann](https://github.com/karthikeyann) +- Missing `f` prefix on f-strings fix ([#10721](https://github.com/NVIDIA/cudf/pull/10721)) [@code-review-doctor](https://github.com/code-review-doctor) +- Add `max_file_size` parameter to chunked parquet dataset writer ([#10718](https://github.com/NVIDIA/cudf/pull/10718)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `merge_sorted`, change dask cudf usage to internal method ([#10713](https://github.com/NVIDIA/cudf/pull/10713)) [@isVoid](https://github.com/isVoid) +- Prepare dask_cudf test_parquet.py for upcoming API changes ([#10709](https://github.com/NVIDIA/cudf/pull/10709)) [@rjzamora](https://github.com/rjzamora) +- Remove or simplify various utility functions ([#10705](https://github.com/NVIDIA/cudf/pull/10705)) [@vyasr](https://github.com/vyasr) +- Allow building arrow with parquet and not python ([#10702](https://github.com/NVIDIA/cudf/pull/10702)) [@revans2](https://github.com/revans2) +- Partial cuIO GPU decompression refactor ([#10699](https://github.com/NVIDIA/cudf/pull/10699)) [@vuule](https://github.com/vuule) +- Cython API refactor: `merge.pyx` ([#10698](https://github.com/NVIDIA/cudf/pull/10698)) [@isVoid](https://github.com/isVoid) +- Fix random string data length to become variable ([#10697](https://github.com/NVIDIA/cudf/pull/10697)) [@galipremsagar](https://github.com/galipremsagar) +- Add bindings for index_of with column search key ([#10696](https://github.com/NVIDIA/cudf/pull/10696)) [@ChrisJar](https://github.com/ChrisJar) +- Deprecate index merging ([#10689](https://github.com/NVIDIA/cudf/pull/10689)) [@vyasr](https://github.com/vyasr) +- Remove cudf::strings::string namespace ([#10684](https://github.com/NVIDIA/cudf/pull/10684)) [@davidwendt](https://github.com/davidwendt) +- Standardize imports. ([#10680](https://github.com/NVIDIA/cudf/pull/10680)) [@bdice](https://github.com/bdice) +- Standardize usage of collections.abc. ([#10679](https://github.com/NVIDIA/cudf/pull/10679)) [@bdice](https://github.com/bdice) +- Cython API Refactor: `transpose.pyx`, `sort.pyx` ([#10675](https://github.com/NVIDIA/cudf/pull/10675)) [@isVoid](https://github.com/isVoid) +- Add device_memory_resource parameter to create_string_vector_from_column ([#10673](https://github.com/NVIDIA/cudf/pull/10673)) [@davidwendt](https://github.com/davidwendt) +- Split up mixed-join kernels source files ([#10671](https://github.com/NVIDIA/cudf/pull/10671)) [@davidwendt](https://github.com/davidwendt) +- Use `std::filesystem` for temporary directory location and deletion ([#10664](https://github.com/NVIDIA/cudf/pull/10664)) [@vuule](https://github.com/vuule) +- cleanup benchmark includes ([#10661](https://github.com/NVIDIA/cudf/pull/10661)) [@karthikeyann](https://github.com/karthikeyann) +- Use upstream clang-format pre-commit hook. ([#10659](https://github.com/NVIDIA/cudf/pull/10659)) [@bdice](https://github.com/bdice) +- Clean up C++ includes to use <> instead of "". ([#10658](https://github.com/NVIDIA/cudf/pull/10658)) [@bdice](https://github.com/bdice) +- Handle RuntimeError thrown by CUDA Python in `validate_setup` ([#10653](https://github.com/NVIDIA/cudf/pull/10653)) [@shwina](https://github.com/shwina) +- Rework JNI CMake to leverage rapids_find_package ([#10649](https://github.com/NVIDIA/cudf/pull/10649)) [@jlowe](https://github.com/jlowe) +- Use conda to build python packages during GPU tests ([#10648](https://github.com/NVIDIA/cudf/pull/10648)) [@Ethyling](https://github.com/Ethyling) +- Deprecate various functions that don't need to be defined for Index. ([#10647](https://github.com/NVIDIA/cudf/pull/10647)) [@vyasr](https://github.com/vyasr) +- Update pinning to allow newer CMake versions. ([#10646](https://github.com/NVIDIA/cudf/pull/10646)) [@vyasr](https://github.com/vyasr) +- Bump hadoop-common from 3.1.4 to 3.2.3 in /java ([#10645](https://github.com/NVIDIA/cudf/pull/10645)) [@dependabot[bot]](https://github.com/dependabot[bot]) +- Remove `concurrent_unordered_multimap`. ([#10642](https://github.com/NVIDIA/cudf/pull/10642)) [@bdice](https://github.com/bdice) +- Improve parquet dictionary encoding ([#10635](https://github.com/NVIDIA/cudf/pull/10635)) [@PointKernel](https://github.com/PointKernel) +- Improve cudf::cuda_error ([#10630](https://github.com/NVIDIA/cudf/pull/10630)) [@sperlingxx](https://github.com/sperlingxx) +- Add support for null and non-numeric types in Series.diff and DataFrame.diff ([#10625](https://github.com/NVIDIA/cudf/pull/10625)) [@Matt711](https://github.com/Matt711) +- Branch 22.06 merge 22.04 ([#10624](https://github.com/NVIDIA/cudf/pull/10624)) [@vyasr](https://github.com/vyasr) +- Unpin `dask` & `distributed` for development ([#10623](https://github.com/NVIDIA/cudf/pull/10623)) [@galipremsagar](https://github.com/galipremsagar) +- Slightly improve accuracy of stod in to_floats ([#10622](https://github.com/NVIDIA/cudf/pull/10622)) [@davidwendt](https://github.com/davidwendt) +- Allow libcudfjni to be built as a static library ([#10619](https://github.com/NVIDIA/cudf/pull/10619)) [@jlowe](https://github.com/jlowe) +- Change stack-based regex state data to use global memory ([#10600](https://github.com/NVIDIA/cudf/pull/10600)) [@davidwendt](https://github.com/davidwendt) +- Resolve Forward merging of `branch-22.04` into `branch-22.06` ([#10598](https://github.com/NVIDIA/cudf/pull/10598)) [@galipremsagar](https://github.com/galipremsagar) +- KvikIO as an alternative GDS backend ([#10593](https://github.com/NVIDIA/cudf/pull/10593)) [@madsbk](https://github.com/madsbk) +- Rename CUDA_TRY macro to CUDF_CUDA_TRY, rename CHECK_CUDA macro to CUDF_CHECK_CUDA. ([#10589](https://github.com/NVIDIA/cudf/pull/10589)) [@bdice](https://github.com/bdice) +- Upgrade `cudf` to support `pandas` 1.4.x versions ([#10584](https://github.com/NVIDIA/cudf/pull/10584)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor binary ops for timedelta and datetime columns ([#10581](https://github.com/NVIDIA/cudf/pull/10581)) [@vyasr](https://github.com/vyasr) +- Refactor cudf::strings::count_re API to use count_matches utility ([#10580](https://github.com/NVIDIA/cudf/pull/10580)) [@davidwendt](https://github.com/davidwendt) +- Update `Programming Language :: Python` Versions to 3.8 & 3.9 ([#10579](https://github.com/NVIDIA/cudf/pull/10579)) [@madsbk](https://github.com/madsbk) +- Automate Java cudf jar build with statically linked dependencies ([#10578](https://github.com/NVIDIA/cudf/pull/10578)) [@gerashegalov](https://github.com/gerashegalov) +- Add patch for thrust-cub 1.16 to fix sort compile times ([#10577](https://github.com/NVIDIA/cudf/pull/10577)) [@davidwendt](https://github.com/davidwendt) +- Move binop methods from Frame to IndexedFrame and standardize the docstring ([#10576](https://github.com/NVIDIA/cudf/pull/10576)) [@vyasr](https://github.com/vyasr) +- Cleanup libcudf strings regex classes ([#10573](https://github.com/NVIDIA/cudf/pull/10573)) [@davidwendt](https://github.com/davidwendt) +- Simplify preprocessing of arguments for DataFrame binops ([#10563](https://github.com/NVIDIA/cudf/pull/10563)) [@vyasr](https://github.com/vyasr) +- Reduce kernel calls to build strings findall results ([#10559](https://github.com/NVIDIA/cudf/pull/10559)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-22.04 to branch-22.06 ([#10557](https://github.com/NVIDIA/cudf/pull/10557)) [@bdice](https://github.com/bdice) +- Update strings contains benchmark to measure varying match rates ([#10555](https://github.com/NVIDIA/cudf/pull/10555)) [@davidwendt](https://github.com/davidwendt) +- JNI: throw CUDA errors more specifically ([#10551](https://github.com/NVIDIA/cudf/pull/10551)) [@sperlingxx](https://github.com/sperlingxx) +- Enable building static libs ([#10545](https://github.com/NVIDIA/cudf/pull/10545)) [@trxcllnt](https://github.com/trxcllnt) +- Remove pip requirements files. ([#10543](https://github.com/NVIDIA/cudf/pull/10543)) [@bdice](https://github.com/bdice) +- Remove Click pinnings that are unnecessary after upgrading black. ([#10541](https://github.com/NVIDIA/cudf/pull/10541)) [@vyasr](https://github.com/vyasr) +- Refactor `memory_usage` to improve performance ([#10537](https://github.com/NVIDIA/cudf/pull/10537)) [@galipremsagar](https://github.com/galipremsagar) +- Adjust the valid range of group index for replace_with_backrefs ([#10530](https://github.com/NVIDIA/cudf/pull/10530)) [@sperlingxx](https://github.com/sperlingxx) +- add accidentally removed comment. ([#10526](https://github.com/NVIDIA/cudf/pull/10526)) [@vyasr](https://github.com/vyasr) +- Update conda environment. ([#10525](https://github.com/NVIDIA/cudf/pull/10525)) [@vyasr](https://github.com/vyasr) +- Remove ColumnBase.__getitem__ ([#10516](https://github.com/NVIDIA/cudf/pull/10516)) [@vyasr](https://github.com/vyasr) +- Optimize `left_semi_join` by materializing the gather mask ([#10511](https://github.com/NVIDIA/cudf/pull/10511)) [@cheinger](https://github.com/cheinger) +- Define proper binary operation APIs for columns ([#10509](https://github.com/NVIDIA/cudf/pull/10509)) [@vyasr](https://github.com/vyasr) +- Upgrade `arrow-cpp` & `pyarrow` to `7.0.0` ([#10503](https://github.com/NVIDIA/cudf/pull/10503)) [@galipremsagar](https://github.com/galipremsagar) +- Update to Thrust 1.16 ([#10489](https://github.com/NVIDIA/cudf/pull/10489)) [@bdice](https://github.com/bdice) +- Namespace/Docstring Fixes for Reduction ([#10471](https://github.com/NVIDIA/cudf/pull/10471)) [@isVoid](https://github.com/isVoid) +- Update cudfjni 22.06.0-SNAPSHOT ([#10467](https://github.com/NVIDIA/cudf/pull/10467)) [@pxLi](https://github.com/pxLi) +- Use Lists of Columns for Various Files ([#10463](https://github.com/NVIDIA/cudf/pull/10463)) [@isVoid](https://github.com/isVoid) +- Additional refactoring of hash functions ([#10462](https://github.com/NVIDIA/cudf/pull/10462)) [@bdice](https://github.com/bdice) +- Fix Series.str.findall behavior for expand=False. ([#10459](https://github.com/NVIDIA/cudf/pull/10459)) [@bdice](https://github.com/bdice) +- Remove deprecated code. ([#10450](https://github.com/NVIDIA/cudf/pull/10450)) [@vyasr](https://github.com/vyasr) +- Update cmake-format version. ([#10440](https://github.com/NVIDIA/cudf/pull/10440)) [@vyasr](https://github.com/vyasr) +- Consolidate C++ `conda` recipes and add `libcudf-tests` package ([#10326](https://github.com/NVIDIA/cudf/pull/10326)) [@ajschmidt8](https://github.com/ajschmidt8) +- Use conda compilers ([#10275](https://github.com/NVIDIA/cudf/pull/10275)) [@Ethyling](https://github.com/Ethyling) +- Add row bitmask as a `detail::hash_join` member ([#10248](https://github.com/NVIDIA/cudf/pull/10248)) [@PointKernel](https://github.com/PointKernel) # cuDF 22.04.00 (6 Apr 2022) ## 🚨 Breaking Changes -- Drop unsupported method argument from nunique and distinct_count. ([#10411](https://github.com/rapidsai/cudf/pull/10411)) [@bdice](https://github.com/bdice) -- Refactor stream compaction APIs ([#10370](https://github.com/rapidsai/cudf/pull/10370)) [@PointKernel](https://github.com/PointKernel) -- Add scan_aggregation and reduce_aggregation derived types. ([#10357](https://github.com/rapidsai/cudf/pull/10357)) [@nvdbaranec](https://github.com/nvdbaranec) -- Avoid `decimal` type narrowing for decimal binops ([#10299](https://github.com/rapidsai/cudf/pull/10299)) [@galipremsagar](https://github.com/galipremsagar) -- Rewrites `sample` API ([#10262](https://github.com/rapidsai/cudf/pull/10262)) [@isVoid](https://github.com/isVoid) -- Remove probe-time null equality parameters in `cudf::hash_join` ([#10260](https://github.com/rapidsai/cudf/pull/10260)) [@PointKernel](https://github.com/PointKernel) -- Enable proper `Index` round-tripping in `orc` reader and writer ([#10170](https://github.com/rapidsai/cudf/pull/10170)) [@galipremsagar](https://github.com/galipremsagar) -- Add JNI for `strings::split_re` and `strings::split_record_re` ([#10139](https://github.com/rapidsai/cudf/pull/10139)) [@ttnghia](https://github.com/ttnghia) -- Change cudf::strings::find_multiple to return a lists column ([#10134](https://github.com/rapidsai/cudf/pull/10134)) [@davidwendt](https://github.com/davidwendt) -- Remove the option to completely disable decimal128 columns in the ORC reader ([#10127](https://github.com/rapidsai/cudf/pull/10127)) [@vuule](https://github.com/vuule) -- Remove deprecated code ([#10124](https://github.com/rapidsai/cudf/pull/10124)) [@vyasr](https://github.com/vyasr) -- Update gpu_utils.py to reflect current CUDA support. ([#10113](https://github.com/rapidsai/cudf/pull/10113)) [@bdice](https://github.com/bdice) -- Optimize compaction operations ([#10030](https://github.com/rapidsai/cudf/pull/10030)) [@PointKernel](https://github.com/PointKernel) -- Remove deprecated method Series.set_index. ([#9945](https://github.com/rapidsai/cudf/pull/9945)) [@bdice](https://github.com/bdice) -- Add cudf::strings::findall_record API ([#9911](https://github.com/rapidsai/cudf/pull/9911)) [@davidwendt](https://github.com/davidwendt) -- Upgrade `arrow` & `pyarrow` to `6.0.1` ([#9686](https://github.com/rapidsai/cudf/pull/9686)) [@galipremsagar](https://github.com/galipremsagar) +- Drop unsupported method argument from nunique and distinct_count. ([#10411](https://github.com/NVIDIA/cudf/pull/10411)) [@bdice](https://github.com/bdice) +- Refactor stream compaction APIs ([#10370](https://github.com/NVIDIA/cudf/pull/10370)) [@PointKernel](https://github.com/PointKernel) +- Add scan_aggregation and reduce_aggregation derived types. ([#10357](https://github.com/NVIDIA/cudf/pull/10357)) [@nvdbaranec](https://github.com/nvdbaranec) +- Avoid `decimal` type narrowing for decimal binops ([#10299](https://github.com/NVIDIA/cudf/pull/10299)) [@galipremsagar](https://github.com/galipremsagar) +- Rewrites `sample` API ([#10262](https://github.com/NVIDIA/cudf/pull/10262)) [@isVoid](https://github.com/isVoid) +- Remove probe-time null equality parameters in `cudf::hash_join` ([#10260](https://github.com/NVIDIA/cudf/pull/10260)) [@PointKernel](https://github.com/PointKernel) +- Enable proper `Index` round-tripping in `orc` reader and writer ([#10170](https://github.com/NVIDIA/cudf/pull/10170)) [@galipremsagar](https://github.com/galipremsagar) +- Add JNI for `strings::split_re` and `strings::split_record_re` ([#10139](https://github.com/NVIDIA/cudf/pull/10139)) [@ttnghia](https://github.com/ttnghia) +- Change cudf::strings::find_multiple to return a lists column ([#10134](https://github.com/NVIDIA/cudf/pull/10134)) [@davidwendt](https://github.com/davidwendt) +- Remove the option to completely disable decimal128 columns in the ORC reader ([#10127](https://github.com/NVIDIA/cudf/pull/10127)) [@vuule](https://github.com/vuule) +- Remove deprecated code ([#10124](https://github.com/NVIDIA/cudf/pull/10124)) [@vyasr](https://github.com/vyasr) +- Update gpu_utils.py to reflect current CUDA support. ([#10113](https://github.com/NVIDIA/cudf/pull/10113)) [@bdice](https://github.com/bdice) +- Optimize compaction operations ([#10030](https://github.com/NVIDIA/cudf/pull/10030)) [@PointKernel](https://github.com/PointKernel) +- Remove deprecated method Series.set_index. ([#9945](https://github.com/NVIDIA/cudf/pull/9945)) [@bdice](https://github.com/bdice) +- Add cudf::strings::findall_record API ([#9911](https://github.com/NVIDIA/cudf/pull/9911)) [@davidwendt](https://github.com/davidwendt) +- Upgrade `arrow` & `pyarrow` to `6.0.1` ([#9686](https://github.com/NVIDIA/cudf/pull/9686)) [@galipremsagar](https://github.com/galipremsagar) ## 🐛 Bug Fixes -- Fix an issue with tdigest merge aggregations. ([#10506](https://github.com/rapidsai/cudf/pull/10506)) [@nvdbaranec](https://github.com/nvdbaranec) -- Batch of fixes for index overflows in grid stride loops. ([#10448](https://github.com/rapidsai/cudf/pull/10448)) [@nvdbaranec](https://github.com/nvdbaranec) -- Update dask_cudf imports to be compatible with latest dask ([#10442](https://github.com/rapidsai/cudf/pull/10442)) [@rlratzel](https://github.com/rlratzel) -- Fix for integer overflow in contiguous-split ([#10437](https://github.com/rapidsai/cudf/pull/10437)) [@jbrennan333](https://github.com/jbrennan333) -- Fix has_null predicate for drop_list_duplicates on nested structs ([#10436](https://github.com/rapidsai/cudf/pull/10436)) [@sperlingxx](https://github.com/sperlingxx) -- Fix empty reduce with List output and non-List input ([#10435](https://github.com/rapidsai/cudf/pull/10435)) [@sperlingxx](https://github.com/sperlingxx) -- Fix `list` and `struct` meta generation issue in `dask-cudf` ([#10434](https://github.com/rapidsai/cudf/pull/10434)) [@galipremsagar](https://github.com/galipremsagar) -- Fix error in `cudf.to_numeric` when a `bool` input is passed ([#10431](https://github.com/rapidsai/cudf/pull/10431)) [@galipremsagar](https://github.com/galipremsagar) -- Support cupy array in `quantile` input ([#10429](https://github.com/rapidsai/cudf/pull/10429)) [@galipremsagar](https://github.com/galipremsagar) -- Fix benchmarks to work with new aggregation types ([#10428](https://github.com/rapidsai/cudf/pull/10428)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf::shift to handle offset greater than column size ([#10414](https://github.com/rapidsai/cudf/pull/10414)) [@davidwendt](https://github.com/davidwendt) -- Fix lifespan of the temporary directory that holds cuFile configuration file ([#10403](https://github.com/rapidsai/cudf/pull/10403)) [@vuule](https://github.com/vuule) -- Fix error thrown in compiled-binaryop benchmark ([#10398](https://github.com/rapidsai/cudf/pull/10398)) [@davidwendt](https://github.com/davidwendt) -- Limiting async allocator using alignment of 512 ([#10395](https://github.com/rapidsai/cudf/pull/10395)) [@rongou](https://github.com/rongou) -- Include <optional> in multibyte split. ([#10385](https://github.com/rapidsai/cudf/pull/10385)) [@bdice](https://github.com/bdice) -- Fix issue with column and scalar re-assignment ([#10377](https://github.com/rapidsai/cudf/pull/10377)) [@galipremsagar](https://github.com/galipremsagar) -- Fix floating point data generation in benchmarks ([#10372](https://github.com/rapidsai/cudf/pull/10372)) [@vuule](https://github.com/vuule) -- Avoid overflow in fused_concatenate_kernel output_index ([#10344](https://github.com/rapidsai/cudf/pull/10344)) [@abellina](https://github.com/abellina) -- Remove is_relationally_comparable for table device views ([#10342](https://github.com/rapidsai/cudf/pull/10342)) [@davidwendt](https://github.com/davidwendt) -- Fix debug compile error in device_span to column_view conversion ([#10331](https://github.com/rapidsai/cudf/pull/10331)) [@davidwendt](https://github.com/davidwendt) -- Add Pascal support to JCUDF transcode (row_conversion) ([#10329](https://github.com/rapidsai/cudf/pull/10329)) [@mythrocks](https://github.com/mythrocks) -- Fix `std::bad_alloc` exception due to JIT reserving a huge buffer ([#10317](https://github.com/rapidsai/cudf/pull/10317)) [@ttnghia](https://github.com/ttnghia) -- Fixes up the overflowed fixed-point round on nullable column ([#10316](https://github.com/rapidsai/cudf/pull/10316)) [@sperlingxx](https://github.com/sperlingxx) -- Fix DataFrame slicing issues for empty cases ([#10310](https://github.com/rapidsai/cudf/pull/10310)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix documentation issues ([#10307](https://github.com/rapidsai/cudf/pull/10307)) [@ajschmidt8](https://github.com/ajschmidt8) -- Allow Java bindings to use default decimal precisions when writing columns ([#10276](https://github.com/rapidsai/cudf/pull/10276)) [@sperlingxx](https://github.com/sperlingxx) -- Fix incorrect slicing of GDS read/write calls ([#10274](https://github.com/rapidsai/cudf/pull/10274)) [@vuule](https://github.com/vuule) -- Fix out-of-memory error in compiled-binaryop benchmark ([#10269](https://github.com/rapidsai/cudf/pull/10269)) [@davidwendt](https://github.com/davidwendt) -- Add tests of reflected ufuncs and fix behavior of logical reflected ufuncs ([#10261](https://github.com/rapidsai/cudf/pull/10261)) [@vyasr](https://github.com/vyasr) -- Remove probe-time null equality parameters in `cudf::hash_join` ([#10260](https://github.com/rapidsai/cudf/pull/10260)) [@PointKernel](https://github.com/PointKernel) -- Fix out-of-memory error in UrlDecode benchmark ([#10258](https://github.com/rapidsai/cudf/pull/10258)) [@davidwendt](https://github.com/davidwendt) -- Fix groupby reductions that perform operations on source type instead of target type ([#10250](https://github.com/rapidsai/cudf/pull/10250)) [@ttnghia](https://github.com/ttnghia) -- Fix small leak in explode ([#10245](https://github.com/rapidsai/cudf/pull/10245)) [@revans2](https://github.com/revans2) -- Yet another small JNI memory leak ([#10238](https://github.com/rapidsai/cudf/pull/10238)) [@revans2](https://github.com/revans2) -- Fix regex octal parsing to limit to 3 characters ([#10233](https://github.com/rapidsai/cudf/pull/10233)) [@davidwendt](https://github.com/davidwendt) -- Fix string to decimal128 conversion handling large exponents ([#10231](https://github.com/rapidsai/cudf/pull/10231)) [@davidwendt](https://github.com/davidwendt) -- Fix JNI leak on copy to device ([#10229](https://github.com/rapidsai/cudf/pull/10229)) [@revans2](https://github.com/revans2) -- Fix the data generator element size for decimal types ([#10225](https://github.com/rapidsai/cudf/pull/10225)) [@vuule](https://github.com/vuule) -- Fix `decimal` metadata in parquet writer ([#10224](https://github.com/rapidsai/cudf/pull/10224)) [@galipremsagar](https://github.com/galipremsagar) -- Fix strings handling of hex in regex pattern ([#10220](https://github.com/rapidsai/cudf/pull/10220)) [@davidwendt](https://github.com/davidwendt) -- Fix docs builds ([#10216](https://github.com/rapidsai/cudf/pull/10216)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix a leftover _has_nulls change from Nullate ([#10211](https://github.com/rapidsai/cudf/pull/10211)) [@devavret](https://github.com/devavret) -- Fix bitmask of the output for JNI of `lists::drop_list_duplicates` ([#10210](https://github.com/rapidsai/cudf/pull/10210)) [@ttnghia](https://github.com/ttnghia) -- Fix compile error in `binaryop/compiled/util.cpp` ([#10209](https://github.com/rapidsai/cudf/pull/10209)) [@ttnghia](https://github.com/ttnghia) -- Skip ORC and Parquet readers' benchmark cases that are not currently supported ([#10194](https://github.com/rapidsai/cudf/pull/10194)) [@vuule](https://github.com/vuule) -- Fix JNI leak of a cudf::column_view native class. ([#10171](https://github.com/rapidsai/cudf/pull/10171)) [@revans2](https://github.com/revans2) -- Enable proper `Index` round-tripping in `orc` reader and writer ([#10170](https://github.com/rapidsai/cudf/pull/10170)) [@galipremsagar](https://github.com/galipremsagar) -- Convert Column Name to String Before Using Struct Column Factory ([#10156](https://github.com/rapidsai/cudf/pull/10156)) [@isVoid](https://github.com/isVoid) -- Preserve the correct `ListDtype` while creating an identical empty column ([#10151](https://github.com/rapidsai/cudf/pull/10151)) [@galipremsagar](https://github.com/galipremsagar) -- benchmark fixture - static object pointer fix ([#10145](https://github.com/rapidsai/cudf/pull/10145)) [@karthikeyann](https://github.com/karthikeyann) -- Fix UDF Caching ([#10133](https://github.com/rapidsai/cudf/pull/10133)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Raise duplicate column error in `DataFrame.rename` ([#10120](https://github.com/rapidsai/cudf/pull/10120)) [@galipremsagar](https://github.com/galipremsagar) -- Fix flaky memory usage test by guaranteeing array size. ([#10114](https://github.com/rapidsai/cudf/pull/10114)) [@vyasr](https://github.com/vyasr) -- Encode values from python callback for C++ ([#10103](https://github.com/rapidsai/cudf/pull/10103)) [@jdye64](https://github.com/jdye64) -- Add check for regex instructions causing an infinite-loop ([#10095](https://github.com/rapidsai/cudf/pull/10095)) [@davidwendt](https://github.com/davidwendt) -- Remove metadata singleton from nvtext normalizer ([#10090](https://github.com/rapidsai/cudf/pull/10090)) [@davidwendt](https://github.com/davidwendt) -- Column equality testing fixes ([#10011](https://github.com/rapidsai/cudf/pull/10011)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Pin libcudf runtime dependency for cudf / libcudf-kafka nightlies ([#9847](https://github.com/rapidsai/cudf/pull/9847)) [@charlesbluca](https://github.com/charlesbluca) +- Fix an issue with tdigest merge aggregations. ([#10506](https://github.com/NVIDIA/cudf/pull/10506)) [@nvdbaranec](https://github.com/nvdbaranec) +- Batch of fixes for index overflows in grid stride loops. ([#10448](https://github.com/NVIDIA/cudf/pull/10448)) [@nvdbaranec](https://github.com/nvdbaranec) +- Update dask_cudf imports to be compatible with latest dask ([#10442](https://github.com/NVIDIA/cudf/pull/10442)) [@rlratzel](https://github.com/rlratzel) +- Fix for integer overflow in contiguous-split ([#10437](https://github.com/NVIDIA/cudf/pull/10437)) [@jbrennan333](https://github.com/jbrennan333) +- Fix has_null predicate for drop_list_duplicates on nested structs ([#10436](https://github.com/NVIDIA/cudf/pull/10436)) [@sperlingxx](https://github.com/sperlingxx) +- Fix empty reduce with List output and non-List input ([#10435](https://github.com/NVIDIA/cudf/pull/10435)) [@sperlingxx](https://github.com/sperlingxx) +- Fix `list` and `struct` meta generation issue in `dask-cudf` ([#10434](https://github.com/NVIDIA/cudf/pull/10434)) [@galipremsagar](https://github.com/galipremsagar) +- Fix error in `cudf.to_numeric` when a `bool` input is passed ([#10431](https://github.com/NVIDIA/cudf/pull/10431)) [@galipremsagar](https://github.com/galipremsagar) +- Support cupy array in `quantile` input ([#10429](https://github.com/NVIDIA/cudf/pull/10429)) [@galipremsagar](https://github.com/galipremsagar) +- Fix benchmarks to work with new aggregation types ([#10428](https://github.com/NVIDIA/cudf/pull/10428)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf::shift to handle offset greater than column size ([#10414](https://github.com/NVIDIA/cudf/pull/10414)) [@davidwendt](https://github.com/davidwendt) +- Fix lifespan of the temporary directory that holds cuFile configuration file ([#10403](https://github.com/NVIDIA/cudf/pull/10403)) [@vuule](https://github.com/vuule) +- Fix error thrown in compiled-binaryop benchmark ([#10398](https://github.com/NVIDIA/cudf/pull/10398)) [@davidwendt](https://github.com/davidwendt) +- Limiting async allocator using alignment of 512 ([#10395](https://github.com/NVIDIA/cudf/pull/10395)) [@rongou](https://github.com/rongou) +- Include <optional> in multibyte split. ([#10385](https://github.com/NVIDIA/cudf/pull/10385)) [@bdice](https://github.com/bdice) +- Fix issue with column and scalar re-assignment ([#10377](https://github.com/NVIDIA/cudf/pull/10377)) [@galipremsagar](https://github.com/galipremsagar) +- Fix floating point data generation in benchmarks ([#10372](https://github.com/NVIDIA/cudf/pull/10372)) [@vuule](https://github.com/vuule) +- Avoid overflow in fused_concatenate_kernel output_index ([#10344](https://github.com/NVIDIA/cudf/pull/10344)) [@abellina](https://github.com/abellina) +- Remove is_relationally_comparable for table device views ([#10342](https://github.com/NVIDIA/cudf/pull/10342)) [@davidwendt](https://github.com/davidwendt) +- Fix debug compile error in device_span to column_view conversion ([#10331](https://github.com/NVIDIA/cudf/pull/10331)) [@davidwendt](https://github.com/davidwendt) +- Add Pascal support to JCUDF transcode (row_conversion) ([#10329](https://github.com/NVIDIA/cudf/pull/10329)) [@mythrocks](https://github.com/mythrocks) +- Fix `std::bad_alloc` exception due to JIT reserving a huge buffer ([#10317](https://github.com/NVIDIA/cudf/pull/10317)) [@ttnghia](https://github.com/ttnghia) +- Fixes up the overflowed fixed-point round on nullable column ([#10316](https://github.com/NVIDIA/cudf/pull/10316)) [@sperlingxx](https://github.com/sperlingxx) +- Fix DataFrame slicing issues for empty cases ([#10310](https://github.com/NVIDIA/cudf/pull/10310)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix documentation issues ([#10307](https://github.com/NVIDIA/cudf/pull/10307)) [@ajschmidt8](https://github.com/ajschmidt8) +- Allow Java bindings to use default decimal precisions when writing columns ([#10276](https://github.com/NVIDIA/cudf/pull/10276)) [@sperlingxx](https://github.com/sperlingxx) +- Fix incorrect slicing of GDS read/write calls ([#10274](https://github.com/NVIDIA/cudf/pull/10274)) [@vuule](https://github.com/vuule) +- Fix out-of-memory error in compiled-binaryop benchmark ([#10269](https://github.com/NVIDIA/cudf/pull/10269)) [@davidwendt](https://github.com/davidwendt) +- Add tests of reflected ufuncs and fix behavior of logical reflected ufuncs ([#10261](https://github.com/NVIDIA/cudf/pull/10261)) [@vyasr](https://github.com/vyasr) +- Remove probe-time null equality parameters in `cudf::hash_join` ([#10260](https://github.com/NVIDIA/cudf/pull/10260)) [@PointKernel](https://github.com/PointKernel) +- Fix out-of-memory error in UrlDecode benchmark ([#10258](https://github.com/NVIDIA/cudf/pull/10258)) [@davidwendt](https://github.com/davidwendt) +- Fix groupby reductions that perform operations on source type instead of target type ([#10250](https://github.com/NVIDIA/cudf/pull/10250)) [@ttnghia](https://github.com/ttnghia) +- Fix small leak in explode ([#10245](https://github.com/NVIDIA/cudf/pull/10245)) [@revans2](https://github.com/revans2) +- Yet another small JNI memory leak ([#10238](https://github.com/NVIDIA/cudf/pull/10238)) [@revans2](https://github.com/revans2) +- Fix regex octal parsing to limit to 3 characters ([#10233](https://github.com/NVIDIA/cudf/pull/10233)) [@davidwendt](https://github.com/davidwendt) +- Fix string to decimal128 conversion handling large exponents ([#10231](https://github.com/NVIDIA/cudf/pull/10231)) [@davidwendt](https://github.com/davidwendt) +- Fix JNI leak on copy to device ([#10229](https://github.com/NVIDIA/cudf/pull/10229)) [@revans2](https://github.com/revans2) +- Fix the data generator element size for decimal types ([#10225](https://github.com/NVIDIA/cudf/pull/10225)) [@vuule](https://github.com/vuule) +- Fix `decimal` metadata in parquet writer ([#10224](https://github.com/NVIDIA/cudf/pull/10224)) [@galipremsagar](https://github.com/galipremsagar) +- Fix strings handling of hex in regex pattern ([#10220](https://github.com/NVIDIA/cudf/pull/10220)) [@davidwendt](https://github.com/davidwendt) +- Fix docs builds ([#10216](https://github.com/NVIDIA/cudf/pull/10216)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix a leftover _has_nulls change from Nullate ([#10211](https://github.com/NVIDIA/cudf/pull/10211)) [@devavret](https://github.com/devavret) +- Fix bitmask of the output for JNI of `lists::drop_list_duplicates` ([#10210](https://github.com/NVIDIA/cudf/pull/10210)) [@ttnghia](https://github.com/ttnghia) +- Fix compile error in `binaryop/compiled/util.cpp` ([#10209](https://github.com/NVIDIA/cudf/pull/10209)) [@ttnghia](https://github.com/ttnghia) +- Skip ORC and Parquet readers' benchmark cases that are not currently supported ([#10194](https://github.com/NVIDIA/cudf/pull/10194)) [@vuule](https://github.com/vuule) +- Fix JNI leak of a cudf::column_view native class. ([#10171](https://github.com/NVIDIA/cudf/pull/10171)) [@revans2](https://github.com/revans2) +- Enable proper `Index` round-tripping in `orc` reader and writer ([#10170](https://github.com/NVIDIA/cudf/pull/10170)) [@galipremsagar](https://github.com/galipremsagar) +- Convert Column Name to String Before Using Struct Column Factory ([#10156](https://github.com/NVIDIA/cudf/pull/10156)) [@isVoid](https://github.com/isVoid) +- Preserve the correct `ListDtype` while creating an identical empty column ([#10151](https://github.com/NVIDIA/cudf/pull/10151)) [@galipremsagar](https://github.com/galipremsagar) +- benchmark fixture - static object pointer fix ([#10145](https://github.com/NVIDIA/cudf/pull/10145)) [@karthikeyann](https://github.com/karthikeyann) +- Fix UDF Caching ([#10133](https://github.com/NVIDIA/cudf/pull/10133)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Raise duplicate column error in `DataFrame.rename` ([#10120](https://github.com/NVIDIA/cudf/pull/10120)) [@galipremsagar](https://github.com/galipremsagar) +- Fix flaky memory usage test by guaranteeing array size. ([#10114](https://github.com/NVIDIA/cudf/pull/10114)) [@vyasr](https://github.com/vyasr) +- Encode values from python callback for C++ ([#10103](https://github.com/NVIDIA/cudf/pull/10103)) [@jdye64](https://github.com/jdye64) +- Add check for regex instructions causing an infinite-loop ([#10095](https://github.com/NVIDIA/cudf/pull/10095)) [@davidwendt](https://github.com/davidwendt) +- Remove metadata singleton from nvtext normalizer ([#10090](https://github.com/NVIDIA/cudf/pull/10090)) [@davidwendt](https://github.com/davidwendt) +- Column equality testing fixes ([#10011](https://github.com/NVIDIA/cudf/pull/10011)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Pin libcudf runtime dependency for cudf / libcudf-kafka nightlies ([#9847](https://github.com/NVIDIA/cudf/pull/9847)) [@charlesbluca](https://github.com/charlesbluca) ## 📖 Documentation -- Fix documentation for DataFrame.corr and Series.corr. ([#10493](https://github.com/rapidsai/cudf/pull/10493)) [@bdice](https://github.com/bdice) -- Add `cut` to API docs ([#10479](https://github.com/rapidsai/cudf/pull/10479)) [@shwina](https://github.com/shwina) -- Remove documentation for methods removed in #10124. ([#10366](https://github.com/rapidsai/cudf/pull/10366)) [@bdice](https://github.com/bdice) -- Fix documentation issues ([#10306](https://github.com/rapidsai/cudf/pull/10306)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix `fixed_point` binary operation documentation ([#10198](https://github.com/rapidsai/cudf/pull/10198)) [@codereport](https://github.com/codereport) -- Remove cleaned up methods from docs ([#10189](https://github.com/rapidsai/cudf/pull/10189)) [@galipremsagar](https://github.com/galipremsagar) -- Update developer guide to recommend no default stream parameter. ([#10136](https://github.com/rapidsai/cudf/pull/10136)) [@bdice](https://github.com/bdice) -- Update benchmarking guide to use NVBench. ([#10093](https://github.com/rapidsai/cudf/pull/10093)) [@bdice](https://github.com/bdice) +- Fix documentation for DataFrame.corr and Series.corr. ([#10493](https://github.com/NVIDIA/cudf/pull/10493)) [@bdice](https://github.com/bdice) +- Add `cut` to API docs ([#10479](https://github.com/NVIDIA/cudf/pull/10479)) [@shwina](https://github.com/shwina) +- Remove documentation for methods removed in #10124. ([#10366](https://github.com/NVIDIA/cudf/pull/10366)) [@bdice](https://github.com/bdice) +- Fix documentation issues ([#10306](https://github.com/NVIDIA/cudf/pull/10306)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix `fixed_point` binary operation documentation ([#10198](https://github.com/NVIDIA/cudf/pull/10198)) [@codereport](https://github.com/codereport) +- Remove cleaned up methods from docs ([#10189](https://github.com/NVIDIA/cudf/pull/10189)) [@galipremsagar](https://github.com/galipremsagar) +- Update developer guide to recommend no default stream parameter. ([#10136](https://github.com/NVIDIA/cudf/pull/10136)) [@bdice](https://github.com/bdice) +- Update benchmarking guide to use NVBench. ([#10093](https://github.com/NVIDIA/cudf/pull/10093)) [@bdice](https://github.com/bdice) ## 🚀 New Features -- Add StringIO support to read_text ([#10465](https://github.com/rapidsai/cudf/pull/10465)) [@cwharris](https://github.com/cwharris) -- Add support for tdigest and merge_tdigest aggregations through cudf::reduce ([#10433](https://github.com/rapidsai/cudf/pull/10433)) [@nvdbaranec](https://github.com/nvdbaranec) -- JNI support for Collect Ops in Reduction ([#10427](https://github.com/rapidsai/cudf/pull/10427)) [@sperlingxx](https://github.com/sperlingxx) -- Enable read_text with dask_cudf using byte_range ([#10407](https://github.com/rapidsai/cudf/pull/10407)) [@ChrisJar](https://github.com/ChrisJar) -- Add `cudf::stable_sort_by_key` ([#10387](https://github.com/rapidsai/cudf/pull/10387)) [@PointKernel](https://github.com/PointKernel) -- Implement `maps_column_view` abstraction over `LIST<STRUCT<K,V>>` ([#10380](https://github.com/rapidsai/cudf/pull/10380)) [@mythrocks](https://github.com/mythrocks) -- Support Java bindings for Avro reader ([#10373](https://github.com/rapidsai/cudf/pull/10373)) [@HaoYang670](https://github.com/HaoYang670) -- Refactor stream compaction APIs ([#10370](https://github.com/rapidsai/cudf/pull/10370)) [@PointKernel](https://github.com/PointKernel) -- Support collect aggregations in reduction ([#10353](https://github.com/rapidsai/cudf/pull/10353)) [@sperlingxx](https://github.com/sperlingxx) -- Refactor array_ufunc for Index and unify across all classes ([#10346](https://github.com/rapidsai/cudf/pull/10346)) [@vyasr](https://github.com/vyasr) -- Add JNI for extract_list_element with index column ([#10341](https://github.com/rapidsai/cudf/pull/10341)) [@firestarman](https://github.com/firestarman) -- Support `min` and `max` operations for structs in rolling window ([#10332](https://github.com/rapidsai/cudf/pull/10332)) [@ttnghia](https://github.com/ttnghia) -- Add device create_sequence_table for benchmarks ([#10300](https://github.com/rapidsai/cudf/pull/10300)) [@karthikeyann](https://github.com/karthikeyann) -- Enable numpy ufuncs for DataFrame ([#10287](https://github.com/rapidsai/cudf/pull/10287)) [@vyasr](https://github.com/vyasr) -- move input generation for json benchmark to device ([#10281](https://github.com/rapidsai/cudf/pull/10281)) [@karthikeyann](https://github.com/karthikeyann) -- move input generation for type dispatcher benchmark to device ([#10280](https://github.com/rapidsai/cudf/pull/10280)) [@karthikeyann](https://github.com/karthikeyann) -- move input generation for copy benchmark to device ([#10279](https://github.com/rapidsai/cudf/pull/10279)) [@karthikeyann](https://github.com/karthikeyann) -- generate url decode benchmark input in device ([#10278](https://github.com/rapidsai/cudf/pull/10278)) [@karthikeyann](https://github.com/karthikeyann) -- device input generation in join bench ([#10277](https://github.com/rapidsai/cudf/pull/10277)) [@karthikeyann](https://github.com/karthikeyann) -- Add nvtext::byte_pair_encoding API ([#10270](https://github.com/rapidsai/cudf/pull/10270)) [@davidwendt](https://github.com/davidwendt) -- Prevent internal usage of expensive APIs ([#10263](https://github.com/rapidsai/cudf/pull/10263)) [@vyasr](https://github.com/vyasr) -- Column to JCUDF row for tables with strings ([#10235](https://github.com/rapidsai/cudf/pull/10235)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Support `percent_rank()` aggregation ([#10227](https://github.com/rapidsai/cudf/pull/10227)) [@mythrocks](https://github.com/mythrocks) -- Refactor Series.__array_ufunc__ ([#10217](https://github.com/rapidsai/cudf/pull/10217)) [@vyasr](https://github.com/vyasr) -- Reduce pytest runtime ([#10203](https://github.com/rapidsai/cudf/pull/10203)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add regex flags parameter to python cudf strings split ([#10185](https://github.com/rapidsai/cudf/pull/10185)) [@davidwendt](https://github.com/davidwendt) -- Support for `MOD`, `PMOD` and `PYMOD` for `decimal32/64/128` ([#10179](https://github.com/rapidsai/cudf/pull/10179)) [@codereport](https://github.com/codereport) -- Adding string row size iterator for row to column and column to row conversion ([#10157](https://github.com/rapidsai/cudf/pull/10157)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add file size counter to cuIO benchmarks ([#10154](https://github.com/rapidsai/cudf/pull/10154)) [@vuule](https://github.com/vuule) -- byte_range support for multibyte_split/read_text ([#10150](https://github.com/rapidsai/cudf/pull/10150)) [@cwharris](https://github.com/cwharris) -- Add JNI for `strings::split_re` and `strings::split_record_re` ([#10139](https://github.com/rapidsai/cudf/pull/10139)) [@ttnghia](https://github.com/ttnghia) -- Add `maxSplit` parameter to Java binding for `strings:split` ([#10137](https://github.com/rapidsai/cudf/pull/10137)) [@ttnghia](https://github.com/ttnghia) -- Add libcudf strings split API that accepts regex pattern ([#10128](https://github.com/rapidsai/cudf/pull/10128)) [@davidwendt](https://github.com/davidwendt) -- generate benchmark input in device ([#10109](https://github.com/rapidsai/cudf/pull/10109)) [@karthikeyann](https://github.com/karthikeyann) -- Avoid `nan_as_null` op if `nan_count` is 0 ([#10082](https://github.com/rapidsai/cudf/pull/10082)) [@galipremsagar](https://github.com/galipremsagar) -- Add Dataframe and Index nunique ([#10077](https://github.com/rapidsai/cudf/pull/10077)) [@martinfalisse](https://github.com/martinfalisse) -- Support nanosecond timestamps in parquet ([#10063](https://github.com/rapidsai/cudf/pull/10063)) [@PointKernel](https://github.com/PointKernel) -- Java bindings for mixed semi and anti joins ([#10040](https://github.com/rapidsai/cudf/pull/10040)) [@jlowe](https://github.com/jlowe) -- Implement mixed equality/conditional semi/anti joins ([#10037](https://github.com/rapidsai/cudf/pull/10037)) [@vyasr](https://github.com/vyasr) -- Optimize compaction operations ([#10030](https://github.com/rapidsai/cudf/pull/10030)) [@PointKernel](https://github.com/PointKernel) -- Support `args=` in `Series.apply` ([#9982](https://github.com/rapidsai/cudf/pull/9982)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add cudf::strings::findall_record API ([#9911](https://github.com/rapidsai/cudf/pull/9911)) [@davidwendt](https://github.com/davidwendt) -- Add covariance for sort groupby (python) ([#9889](https://github.com/rapidsai/cudf/pull/9889)) [@mayankanand007](https://github.com/mayankanand007) -- Implement DataFrame diff() ([#9817](https://github.com/rapidsai/cudf/pull/9817)) [@skirui-source](https://github.com/skirui-source) -- Implement DataFrame pct_change ([#9805](https://github.com/rapidsai/cudf/pull/9805)) [@skirui-source](https://github.com/skirui-source) -- Support segmented reductions and null mask reductions ([#9621](https://github.com/rapidsai/cudf/pull/9621)) [@isVoid](https://github.com/isVoid) -- Add 'spearman' correlation method for `dataframe.corr` and `series.corr` ([#7141](https://github.com/rapidsai/cudf/pull/7141)) [@dominicshanshan](https://github.com/dominicshanshan) +- Add StringIO support to read_text ([#10465](https://github.com/NVIDIA/cudf/pull/10465)) [@cwharris](https://github.com/cwharris) +- Add support for tdigest and merge_tdigest aggregations through cudf::reduce ([#10433](https://github.com/NVIDIA/cudf/pull/10433)) [@nvdbaranec](https://github.com/nvdbaranec) +- JNI support for Collect Ops in Reduction ([#10427](https://github.com/NVIDIA/cudf/pull/10427)) [@sperlingxx](https://github.com/sperlingxx) +- Enable read_text with dask_cudf using byte_range ([#10407](https://github.com/NVIDIA/cudf/pull/10407)) [@ChrisJar](https://github.com/ChrisJar) +- Add `cudf::stable_sort_by_key` ([#10387](https://github.com/NVIDIA/cudf/pull/10387)) [@PointKernel](https://github.com/PointKernel) +- Implement `maps_column_view` abstraction over `LIST<STRUCT<K,V>>` ([#10380](https://github.com/NVIDIA/cudf/pull/10380)) [@mythrocks](https://github.com/mythrocks) +- Support Java bindings for Avro reader ([#10373](https://github.com/NVIDIA/cudf/pull/10373)) [@HaoYang670](https://github.com/HaoYang670) +- Refactor stream compaction APIs ([#10370](https://github.com/NVIDIA/cudf/pull/10370)) [@PointKernel](https://github.com/PointKernel) +- Support collect aggregations in reduction ([#10353](https://github.com/NVIDIA/cudf/pull/10353)) [@sperlingxx](https://github.com/sperlingxx) +- Refactor array_ufunc for Index and unify across all classes ([#10346](https://github.com/NVIDIA/cudf/pull/10346)) [@vyasr](https://github.com/vyasr) +- Add JNI for extract_list_element with index column ([#10341](https://github.com/NVIDIA/cudf/pull/10341)) [@firestarman](https://github.com/firestarman) +- Support `min` and `max` operations for structs in rolling window ([#10332](https://github.com/NVIDIA/cudf/pull/10332)) [@ttnghia](https://github.com/ttnghia) +- Add device create_sequence_table for benchmarks ([#10300](https://github.com/NVIDIA/cudf/pull/10300)) [@karthikeyann](https://github.com/karthikeyann) +- Enable numpy ufuncs for DataFrame ([#10287](https://github.com/NVIDIA/cudf/pull/10287)) [@vyasr](https://github.com/vyasr) +- move input generation for json benchmark to device ([#10281](https://github.com/NVIDIA/cudf/pull/10281)) [@karthikeyann](https://github.com/karthikeyann) +- move input generation for type dispatcher benchmark to device ([#10280](https://github.com/NVIDIA/cudf/pull/10280)) [@karthikeyann](https://github.com/karthikeyann) +- move input generation for copy benchmark to device ([#10279](https://github.com/NVIDIA/cudf/pull/10279)) [@karthikeyann](https://github.com/karthikeyann) +- generate url decode benchmark input in device ([#10278](https://github.com/NVIDIA/cudf/pull/10278)) [@karthikeyann](https://github.com/karthikeyann) +- device input generation in join bench ([#10277](https://github.com/NVIDIA/cudf/pull/10277)) [@karthikeyann](https://github.com/karthikeyann) +- Add nvtext::byte_pair_encoding API ([#10270](https://github.com/NVIDIA/cudf/pull/10270)) [@davidwendt](https://github.com/davidwendt) +- Prevent internal usage of expensive APIs ([#10263](https://github.com/NVIDIA/cudf/pull/10263)) [@vyasr](https://github.com/vyasr) +- Column to JCUDF row for tables with strings ([#10235](https://github.com/NVIDIA/cudf/pull/10235)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Support `percent_rank()` aggregation ([#10227](https://github.com/NVIDIA/cudf/pull/10227)) [@mythrocks](https://github.com/mythrocks) +- Refactor Series.__array_ufunc__ ([#10217](https://github.com/NVIDIA/cudf/pull/10217)) [@vyasr](https://github.com/vyasr) +- Reduce pytest runtime ([#10203](https://github.com/NVIDIA/cudf/pull/10203)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add regex flags parameter to python cudf strings split ([#10185](https://github.com/NVIDIA/cudf/pull/10185)) [@davidwendt](https://github.com/davidwendt) +- Support for `MOD`, `PMOD` and `PYMOD` for `decimal32/64/128` ([#10179](https://github.com/NVIDIA/cudf/pull/10179)) [@codereport](https://github.com/codereport) +- Adding string row size iterator for row to column and column to row conversion ([#10157](https://github.com/NVIDIA/cudf/pull/10157)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add file size counter to cuIO benchmarks ([#10154](https://github.com/NVIDIA/cudf/pull/10154)) [@vuule](https://github.com/vuule) +- byte_range support for multibyte_split/read_text ([#10150](https://github.com/NVIDIA/cudf/pull/10150)) [@cwharris](https://github.com/cwharris) +- Add JNI for `strings::split_re` and `strings::split_record_re` ([#10139](https://github.com/NVIDIA/cudf/pull/10139)) [@ttnghia](https://github.com/ttnghia) +- Add `maxSplit` parameter to Java binding for `strings:split` ([#10137](https://github.com/NVIDIA/cudf/pull/10137)) [@ttnghia](https://github.com/ttnghia) +- Add libcudf strings split API that accepts regex pattern ([#10128](https://github.com/NVIDIA/cudf/pull/10128)) [@davidwendt](https://github.com/davidwendt) +- generate benchmark input in device ([#10109](https://github.com/NVIDIA/cudf/pull/10109)) [@karthikeyann](https://github.com/karthikeyann) +- Avoid `nan_as_null` op if `nan_count` is 0 ([#10082](https://github.com/NVIDIA/cudf/pull/10082)) [@galipremsagar](https://github.com/galipremsagar) +- Add Dataframe and Index nunique ([#10077](https://github.com/NVIDIA/cudf/pull/10077)) [@martinfalisse](https://github.com/martinfalisse) +- Support nanosecond timestamps in parquet ([#10063](https://github.com/NVIDIA/cudf/pull/10063)) [@PointKernel](https://github.com/PointKernel) +- Java bindings for mixed semi and anti joins ([#10040](https://github.com/NVIDIA/cudf/pull/10040)) [@jlowe](https://github.com/jlowe) +- Implement mixed equality/conditional semi/anti joins ([#10037](https://github.com/NVIDIA/cudf/pull/10037)) [@vyasr](https://github.com/vyasr) +- Optimize compaction operations ([#10030](https://github.com/NVIDIA/cudf/pull/10030)) [@PointKernel](https://github.com/PointKernel) +- Support `args=` in `Series.apply` ([#9982](https://github.com/NVIDIA/cudf/pull/9982)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add cudf::strings::findall_record API ([#9911](https://github.com/NVIDIA/cudf/pull/9911)) [@davidwendt](https://github.com/davidwendt) +- Add covariance for sort groupby (python) ([#9889](https://github.com/NVIDIA/cudf/pull/9889)) [@mayankanand007](https://github.com/mayankanand007) +- Implement DataFrame diff() ([#9817](https://github.com/NVIDIA/cudf/pull/9817)) [@skirui-source](https://github.com/skirui-source) +- Implement DataFrame pct_change ([#9805](https://github.com/NVIDIA/cudf/pull/9805)) [@skirui-source](https://github.com/skirui-source) +- Support segmented reductions and null mask reductions ([#9621](https://github.com/NVIDIA/cudf/pull/9621)) [@isVoid](https://github.com/isVoid) +- Add 'spearman' correlation method for `dataframe.corr` and `series.corr` ([#7141](https://github.com/NVIDIA/cudf/pull/7141)) [@dominicshanshan](https://github.com/dominicshanshan) ## 🛠️ Improvements -- Add `scipy` skip for a test ([#10502](https://github.com/rapidsai/cudf/pull/10502)) [@galipremsagar](https://github.com/galipremsagar) -- Temporarily disable new `ops-bot` functionality ([#10496](https://github.com/rapidsai/cudf/pull/10496)) [@ajschmidt8](https://github.com/ajschmidt8) -- Include <cstddef> to fix compilation of parquet reader on GCC 11. ([#10483](https://github.com/rapidsai/cudf/pull/10483)) [@bdice](https://github.com/bdice) -- Pin `dask` and `distributed` ([#10481](https://github.com/rapidsai/cudf/pull/10481)) [@galipremsagar](https://github.com/galipremsagar) -- MD5 refactoring. ([#10445](https://github.com/rapidsai/cudf/pull/10445)) [@bdice](https://github.com/bdice) -- Remove or split up Frame methods that use the index ([#10439](https://github.com/rapidsai/cudf/pull/10439)) [@vyasr](https://github.com/vyasr) -- Centralization of tdigest aggregation code. ([#10422](https://github.com/rapidsai/cudf/pull/10422)) [@nvdbaranec](https://github.com/nvdbaranec) -- Simplify column binary operations ([#10421](https://github.com/rapidsai/cudf/pull/10421)) [@vyasr](https://github.com/vyasr) -- Add `.github/ops-bot.yaml` config file ([#10420](https://github.com/rapidsai/cudf/pull/10420)) [@ajschmidt8](https://github.com/ajschmidt8) -- Use list of columns for methods in `Groupby.pyx` ([#10419](https://github.com/rapidsai/cudf/pull/10419)) [@isVoid](https://github.com/isVoid) -- Remove warnings in `test_timedelta.py` ([#10418](https://github.com/rapidsai/cudf/pull/10418)) [@galipremsagar](https://github.com/galipremsagar) -- Fix some warnings in `test_parquet.py` ([#10416](https://github.com/rapidsai/cudf/pull/10416)) [@galipremsagar](https://github.com/galipremsagar) -- JNI support for segmented reduce ([#10413](https://github.com/rapidsai/cudf/pull/10413)) [@revans2](https://github.com/revans2) -- Clean up null mask after purging null entries ([#10412](https://github.com/rapidsai/cudf/pull/10412)) [@sperlingxx](https://github.com/sperlingxx) -- Drop unsupported method argument from nunique and distinct_count. ([#10411](https://github.com/rapidsai/cudf/pull/10411)) [@bdice](https://github.com/bdice) -- Use str instead of builtins.str. ([#10410](https://github.com/rapidsai/cudf/pull/10410)) [@bdice](https://github.com/bdice) -- Fix warnings in `test_rolling` ([#10405](https://github.com/rapidsai/cudf/pull/10405)) [@bdice](https://github.com/bdice) -- Enable `codecov` github-check in CI ([#10404](https://github.com/rapidsai/cudf/pull/10404)) [@galipremsagar](https://github.com/galipremsagar) -- Fix warnings in test_cuda_apply, test_numerical, test_pickling, test_unaops. ([#10402](https://github.com/rapidsai/cudf/pull/10402)) [@bdice](https://github.com/bdice) -- Set column names in `_from_columns_like_self` factory ([#10400](https://github.com/rapidsai/cudf/pull/10400)) [@isVoid](https://github.com/isVoid) -- Refactor `nvtx` annotations in `cudf` & `dask-cudf` ([#10396](https://github.com/rapidsai/cudf/pull/10396)) [@galipremsagar](https://github.com/galipremsagar) -- Consolidate .cov and .corr for sort groupby ([#10386](https://github.com/rapidsai/cudf/pull/10386)) [@skirui-source](https://github.com/skirui-source) -- Consolidate some Frame APIs ([#10381](https://github.com/rapidsai/cudf/pull/10381)) [@vyasr](https://github.com/vyasr) -- Refactor hash functions and `hash_combine` ([#10379](https://github.com/rapidsai/cudf/pull/10379)) [@bdice](https://github.com/bdice) -- Add `nvtx` annotations for `Series` and `Index` ([#10374](https://github.com/rapidsai/cudf/pull/10374)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor `filling.repeat` API ([#10371](https://github.com/rapidsai/cudf/pull/10371)) [@isVoid](https://github.com/isVoid) -- Move standalone UTF8 functions from string_view.hpp to utf8.hpp ([#10369](https://github.com/rapidsai/cudf/pull/10369)) [@davidwendt](https://github.com/davidwendt) -- Remove doc for deprecated function `one_hot_encoding` ([#10367](https://github.com/rapidsai/cudf/pull/10367)) [@isVoid](https://github.com/isVoid) -- Refactor array function ([#10364](https://github.com/rapidsai/cudf/pull/10364)) [@vyasr](https://github.com/vyasr) -- Fix warnings in test_csv.py. ([#10362](https://github.com/rapidsai/cudf/pull/10362)) [@bdice](https://github.com/bdice) -- Implement a mixin for binops ([#10360](https://github.com/rapidsai/cudf/pull/10360)) [@vyasr](https://github.com/vyasr) -- Refactor cython interface: `copying.pyx` ([#10359](https://github.com/rapidsai/cudf/pull/10359)) [@isVoid](https://github.com/isVoid) -- Implement a mixin for scans ([#10358](https://github.com/rapidsai/cudf/pull/10358)) [@vyasr](https://github.com/vyasr) -- Add scan_aggregation and reduce_aggregation derived types. ([#10357](https://github.com/rapidsai/cudf/pull/10357)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add cleanup of python artifacts ([#10355](https://github.com/rapidsai/cudf/pull/10355)) [@galipremsagar](https://github.com/galipremsagar) -- Fix warnings in test_categorical.py. ([#10354](https://github.com/rapidsai/cudf/pull/10354)) [@bdice](https://github.com/bdice) -- Create a dispatcher for invoking regex kernel functions ([#10349](https://github.com/rapidsai/cudf/pull/10349)) [@davidwendt](https://github.com/davidwendt) -- Fix `codecov` in CI ([#10347](https://github.com/rapidsai/cudf/pull/10347)) [@galipremsagar](https://github.com/galipremsagar) -- Enable caching for `memory_usage` calculation in `Column` ([#10345](https://github.com/rapidsai/cudf/pull/10345)) [@galipremsagar](https://github.com/galipremsagar) -- C++17 cleanup: traits replace std::enable_if<>::type with std::enable_if_t ([#10343](https://github.com/rapidsai/cudf/pull/10343)) [@karthikeyann](https://github.com/karthikeyann) -- JNI: Support appending DECIMAL128 into ColumnBuilder in terms of byte array ([#10338](https://github.com/rapidsai/cudf/pull/10338)) [@sperlingxx](https://github.com/sperlingxx) -- multibyte_split test improvements ([#10328](https://github.com/rapidsai/cudf/pull/10328)) [@vuule](https://github.com/vuule) -- Fix warnings in test_binops.py. ([#10327](https://github.com/rapidsai/cudf/pull/10327)) [@bdice](https://github.com/bdice) -- Fix warnings from pandas in test_array_ufunc.py. ([#10324](https://github.com/rapidsai/cudf/pull/10324)) [@bdice](https://github.com/bdice) -- Update upload script ([#10321](https://github.com/rapidsai/cudf/pull/10321)) [@ajschmidt8](https://github.com/ajschmidt8) -- Move hash type declarations to hashing.hpp ([#10320](https://github.com/rapidsai/cudf/pull/10320)) [@davidwendt](https://github.com/davidwendt) -- C++17 cleanup: traits replace `::value` with `_v` ([#10319](https://github.com/rapidsai/cudf/pull/10319)) [@karthikeyann](https://github.com/karthikeyann) -- Remove internal columns usage ([#10315](https://github.com/rapidsai/cudf/pull/10315)) [@vyasr](https://github.com/vyasr) -- Remove extraneous `build.sh` parameter ([#10313](https://github.com/rapidsai/cudf/pull/10313)) [@ajschmidt8](https://github.com/ajschmidt8) -- Add const qualifier to MurmurHash3_32::hash_combine ([#10311](https://github.com/rapidsai/cudf/pull/10311)) [@davidwendt](https://github.com/davidwendt) -- Remove `TODO` in `libcudf_kafka` recipe ([#10309](https://github.com/rapidsai/cudf/pull/10309)) [@ajschmidt8](https://github.com/ajschmidt8) -- Add conversions between column_view and device_span<T const>. ([#10302](https://github.com/rapidsai/cudf/pull/10302)) [@bdice](https://github.com/bdice) -- Avoid `decimal` type narrowing for decimal binops ([#10299](https://github.com/rapidsai/cudf/pull/10299)) [@galipremsagar](https://github.com/galipremsagar) -- Deprecate `DataFrame.iteritems` and introduce `.items` ([#10298](https://github.com/rapidsai/cudf/pull/10298)) [@galipremsagar](https://github.com/galipremsagar) -- Explicitly request CMake use `gnu++17` over `c++17` ([#10297](https://github.com/rapidsai/cudf/pull/10297)) [@robertmaynard](https://github.com/robertmaynard) -- Add copyright check as pre-commit hook. ([#10290](https://github.com/rapidsai/cudf/pull/10290)) [@vyasr](https://github.com/vyasr) -- DataFrame `insert` and creation optimizations ([#10285](https://github.com/rapidsai/cudf/pull/10285)) [@galipremsagar](https://github.com/galipremsagar) -- Improve hash join detail functions ([#10273](https://github.com/rapidsai/cudf/pull/10273)) [@PointKernel](https://github.com/PointKernel) -- Replace custom `cached_property` implementation with functools ([#10272](https://github.com/rapidsai/cudf/pull/10272)) [@shwina](https://github.com/shwina) -- Rewrites `sample` API ([#10262](https://github.com/rapidsai/cudf/pull/10262)) [@isVoid](https://github.com/isVoid) -- Bump hadoop-common from 3.1.0 to 3.1.4 in /java ([#10259](https://github.com/rapidsai/cudf/pull/10259)) [@dependabot[bot]](https://github.com/dependabot[bot]) -- Remove making redundant `copy` across code-base ([#10257](https://github.com/rapidsai/cudf/pull/10257)) [@galipremsagar](https://github.com/galipremsagar) -- Add more `nvtx` annotations ([#10256](https://github.com/rapidsai/cudf/pull/10256)) [@galipremsagar](https://github.com/galipremsagar) -- Add `copyright` check in `cudf` ([#10253](https://github.com/rapidsai/cudf/pull/10253)) [@galipremsagar](https://github.com/galipremsagar) -- Remove redundant copies in `fillna` to improve performance ([#10241](https://github.com/rapidsai/cudf/pull/10241)) [@galipremsagar](https://github.com/galipremsagar) -- Remove `std::numeric_limit` specializations for timestamp & durations ([#10239](https://github.com/rapidsai/cudf/pull/10239)) [@codereport](https://github.com/codereport) -- Optimize `DataFrame` creation across code-base ([#10236](https://github.com/rapidsai/cudf/pull/10236)) [@galipremsagar](https://github.com/galipremsagar) -- Change pytest distribution algorithm and increase parallelism in CI ([#10232](https://github.com/rapidsai/cudf/pull/10232)) [@galipremsagar](https://github.com/galipremsagar) -- Add environment variables for I/O thread pool and slice sizes ([#10218](https://github.com/rapidsai/cudf/pull/10218)) [@vuule](https://github.com/vuule) -- Add regex flags to strings findall functions ([#10208](https://github.com/rapidsai/cudf/pull/10208)) [@davidwendt](https://github.com/davidwendt) -- Update dask-cudf parquet tests to reflect upstream bugfixes to `_metadata` ([#10206](https://github.com/rapidsai/cudf/pull/10206)) [@charlesbluca](https://github.com/charlesbluca) -- Remove unnecessary nunique function in Series. ([#10205](https://github.com/rapidsai/cudf/pull/10205)) [@martinfalisse](https://github.com/martinfalisse) -- Refactor DataFrame tests. ([#10204](https://github.com/rapidsai/cudf/pull/10204)) [@bdice](https://github.com/bdice) -- Rewrites `column.__setitem__`, Use `boolean_mask_scatter` ([#10202](https://github.com/rapidsai/cudf/pull/10202)) [@isVoid](https://github.com/isVoid) -- Java utilities to aid in accelerating aggregations on 128-bit types ([#10201](https://github.com/rapidsai/cudf/pull/10201)) [@jlowe](https://github.com/jlowe) -- Fix docstrings alignment in `Frame` methods ([#10199](https://github.com/rapidsai/cudf/pull/10199)) [@galipremsagar](https://github.com/galipremsagar) -- Fix cuco pair issue in hash join ([#10195](https://github.com/rapidsai/cudf/pull/10195)) [@PointKernel](https://github.com/PointKernel) -- Replace `dask` groupby `.index` usages with `.by` ([#10193](https://github.com/rapidsai/cudf/pull/10193)) [@galipremsagar](https://github.com/galipremsagar) -- Add regex flags to strings extract function ([#10192](https://github.com/rapidsai/cudf/pull/10192)) [@davidwendt](https://github.com/davidwendt) -- Forward-merge branch-22.02 to branch-22.04 ([#10191](https://github.com/rapidsai/cudf/pull/10191)) [@bdice](https://github.com/bdice) -- Add CMake `install` rule for tests ([#10190](https://github.com/rapidsai/cudf/pull/10190)) [@ajschmidt8](https://github.com/ajschmidt8) -- Unpin `dask` & `distributed` ([#10182](https://github.com/rapidsai/cudf/pull/10182)) [@galipremsagar](https://github.com/galipremsagar) -- Add comments to explain test validation ([#10176](https://github.com/rapidsai/cudf/pull/10176)) [@galipremsagar](https://github.com/galipremsagar) -- Reduce warnings in pytest output ([#10168](https://github.com/rapidsai/cudf/pull/10168)) [@bdice](https://github.com/bdice) -- Some consolidation of indexed frame methods ([#10167](https://github.com/rapidsai/cudf/pull/10167)) [@vyasr](https://github.com/vyasr) -- Refactor isin implementations ([#10165](https://github.com/rapidsai/cudf/pull/10165)) [@vyasr](https://github.com/vyasr) -- Faster struct row comparator ([#10164](https://github.com/rapidsai/cudf/pull/10164)) [@devavret](https://github.com/devavret) -- Refactor groupby::get_groups. ([#10161](https://github.com/rapidsai/cudf/pull/10161)) [@bdice](https://github.com/bdice) -- Deprecate `decimal_cols_as_float` in ORC reader (C++ layer) ([#10152](https://github.com/rapidsai/cudf/pull/10152)) [@vuule](https://github.com/vuule) -- Replace `ccache` with `sccache` ([#10146](https://github.com/rapidsai/cudf/pull/10146)) [@ajschmidt8](https://github.com/ajschmidt8) -- Murmur3 hash kernel cleanup ([#10143](https://github.com/rapidsai/cudf/pull/10143)) [@rwlee](https://github.com/rwlee) -- Deprecate `decimal_cols_as_float` in ORC reader ([#10142](https://github.com/rapidsai/cudf/pull/10142)) [@galipremsagar](https://github.com/galipremsagar) -- Run pyupgrade 2.31.0. ([#10141](https://github.com/rapidsai/cudf/pull/10141)) [@bdice](https://github.com/bdice) -- Remove `drop_nan` from internal `IndexedFrame._drop_na_rows`. ([#10140](https://github.com/rapidsai/cudf/pull/10140)) [@bdice](https://github.com/bdice) -- Change cudf::strings::find_multiple to return a lists column ([#10134](https://github.com/rapidsai/cudf/pull/10134)) [@davidwendt](https://github.com/davidwendt) -- Update cmake-format script for branch 22.04. ([#10132](https://github.com/rapidsai/cudf/pull/10132)) [@bdice](https://github.com/bdice) -- Accept r-value references in convert_table_for_return(): ([#10131](https://github.com/rapidsai/cudf/pull/10131)) [@mythrocks](https://github.com/mythrocks) -- Remove the option to completely disable decimal128 columns in the ORC reader ([#10127](https://github.com/rapidsai/cudf/pull/10127)) [@vuule](https://github.com/vuule) -- Remove deprecated code ([#10124](https://github.com/rapidsai/cudf/pull/10124)) [@vyasr](https://github.com/vyasr) -- Update gpu_utils.py to reflect current CUDA support. ([#10113](https://github.com/rapidsai/cudf/pull/10113)) [@bdice](https://github.com/bdice) -- Remove benchmarks suffix ([#10112](https://github.com/rapidsai/cudf/pull/10112)) [@bdice](https://github.com/bdice) -- Update cudf java binding version to 22.04.0-SNAPSHOT ([#10084](https://github.com/rapidsai/cudf/pull/10084)) [@pxLi](https://github.com/pxLi) -- Remove unnecessary docker files. ([#10069](https://github.com/rapidsai/cudf/pull/10069)) [@vyasr](https://github.com/vyasr) -- Limit benchmark iterations using environment variable ([#10060](https://github.com/rapidsai/cudf/pull/10060)) [@karthikeyann](https://github.com/karthikeyann) -- Add timing chart for libcudf build metrics report page ([#10038](https://github.com/rapidsai/cudf/pull/10038)) [@davidwendt](https://github.com/davidwendt) -- JNI: Rewrite growBuffersAndRows to accelerate the HostColumnBuilder ([#10025](https://github.com/rapidsai/cudf/pull/10025)) [@sperlingxx](https://github.com/sperlingxx) -- Reduce redundant code in CUDF JNI ([#10019](https://github.com/rapidsai/cudf/pull/10019)) [@mythrocks](https://github.com/mythrocks) -- Make snappy decompress check more efficient ([#9995](https://github.com/rapidsai/cudf/pull/9995)) [@cheinger](https://github.com/cheinger) -- Remove deprecated method Series.set_index. ([#9945](https://github.com/rapidsai/cudf/pull/9945)) [@bdice](https://github.com/bdice) -- Implement a mixin for reductions ([#9925](https://github.com/rapidsai/cudf/pull/9925)) [@vyasr](https://github.com/vyasr) -- JNI: Push back decimal utils from spark-rapids ([#9907](https://github.com/rapidsai/cudf/pull/9907)) [@sperlingxx](https://github.com/sperlingxx) -- Add `assert_column_memory_*` ([#9882](https://github.com/rapidsai/cudf/pull/9882)) [@isVoid](https://github.com/isVoid) -- Add CUDF_UNREACHABLE macro. ([#9727](https://github.com/rapidsai/cudf/pull/9727)) [@bdice](https://github.com/bdice) -- Upgrade `arrow` & `pyarrow` to `6.0.1` ([#9686](https://github.com/rapidsai/cudf/pull/9686)) [@galipremsagar](https://github.com/galipremsagar) +- Add `scipy` skip for a test ([#10502](https://github.com/NVIDIA/cudf/pull/10502)) [@galipremsagar](https://github.com/galipremsagar) +- Temporarily disable new `ops-bot` functionality ([#10496](https://github.com/NVIDIA/cudf/pull/10496)) [@ajschmidt8](https://github.com/ajschmidt8) +- Include <cstddef> to fix compilation of parquet reader on GCC 11. ([#10483](https://github.com/NVIDIA/cudf/pull/10483)) [@bdice](https://github.com/bdice) +- Pin `dask` and `distributed` ([#10481](https://github.com/NVIDIA/cudf/pull/10481)) [@galipremsagar](https://github.com/galipremsagar) +- MD5 refactoring. ([#10445](https://github.com/NVIDIA/cudf/pull/10445)) [@bdice](https://github.com/bdice) +- Remove or split up Frame methods that use the index ([#10439](https://github.com/NVIDIA/cudf/pull/10439)) [@vyasr](https://github.com/vyasr) +- Centralization of tdigest aggregation code. ([#10422](https://github.com/NVIDIA/cudf/pull/10422)) [@nvdbaranec](https://github.com/nvdbaranec) +- Simplify column binary operations ([#10421](https://github.com/NVIDIA/cudf/pull/10421)) [@vyasr](https://github.com/vyasr) +- Add `.github/ops-bot.yaml` config file ([#10420](https://github.com/NVIDIA/cudf/pull/10420)) [@ajschmidt8](https://github.com/ajschmidt8) +- Use list of columns for methods in `Groupby.pyx` ([#10419](https://github.com/NVIDIA/cudf/pull/10419)) [@isVoid](https://github.com/isVoid) +- Remove warnings in `test_timedelta.py` ([#10418](https://github.com/NVIDIA/cudf/pull/10418)) [@galipremsagar](https://github.com/galipremsagar) +- Fix some warnings in `test_parquet.py` ([#10416](https://github.com/NVIDIA/cudf/pull/10416)) [@galipremsagar](https://github.com/galipremsagar) +- JNI support for segmented reduce ([#10413](https://github.com/NVIDIA/cudf/pull/10413)) [@revans2](https://github.com/revans2) +- Clean up null mask after purging null entries ([#10412](https://github.com/NVIDIA/cudf/pull/10412)) [@sperlingxx](https://github.com/sperlingxx) +- Drop unsupported method argument from nunique and distinct_count. ([#10411](https://github.com/NVIDIA/cudf/pull/10411)) [@bdice](https://github.com/bdice) +- Use str instead of builtins.str. ([#10410](https://github.com/NVIDIA/cudf/pull/10410)) [@bdice](https://github.com/bdice) +- Fix warnings in `test_rolling` ([#10405](https://github.com/NVIDIA/cudf/pull/10405)) [@bdice](https://github.com/bdice) +- Enable `codecov` github-check in CI ([#10404](https://github.com/NVIDIA/cudf/pull/10404)) [@galipremsagar](https://github.com/galipremsagar) +- Fix warnings in test_cuda_apply, test_numerical, test_pickling, test_unaops. ([#10402](https://github.com/NVIDIA/cudf/pull/10402)) [@bdice](https://github.com/bdice) +- Set column names in `_from_columns_like_self` factory ([#10400](https://github.com/NVIDIA/cudf/pull/10400)) [@isVoid](https://github.com/isVoid) +- Refactor `nvtx` annotations in `cudf` & `dask-cudf` ([#10396](https://github.com/NVIDIA/cudf/pull/10396)) [@galipremsagar](https://github.com/galipremsagar) +- Consolidate .cov and .corr for sort groupby ([#10386](https://github.com/NVIDIA/cudf/pull/10386)) [@skirui-source](https://github.com/skirui-source) +- Consolidate some Frame APIs ([#10381](https://github.com/NVIDIA/cudf/pull/10381)) [@vyasr](https://github.com/vyasr) +- Refactor hash functions and `hash_combine` ([#10379](https://github.com/NVIDIA/cudf/pull/10379)) [@bdice](https://github.com/bdice) +- Add `nvtx` annotations for `Series` and `Index` ([#10374](https://github.com/NVIDIA/cudf/pull/10374)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor `filling.repeat` API ([#10371](https://github.com/NVIDIA/cudf/pull/10371)) [@isVoid](https://github.com/isVoid) +- Move standalone UTF8 functions from string_view.hpp to utf8.hpp ([#10369](https://github.com/NVIDIA/cudf/pull/10369)) [@davidwendt](https://github.com/davidwendt) +- Remove doc for deprecated function `one_hot_encoding` ([#10367](https://github.com/NVIDIA/cudf/pull/10367)) [@isVoid](https://github.com/isVoid) +- Refactor array function ([#10364](https://github.com/NVIDIA/cudf/pull/10364)) [@vyasr](https://github.com/vyasr) +- Fix warnings in test_csv.py. ([#10362](https://github.com/NVIDIA/cudf/pull/10362)) [@bdice](https://github.com/bdice) +- Implement a mixin for binops ([#10360](https://github.com/NVIDIA/cudf/pull/10360)) [@vyasr](https://github.com/vyasr) +- Refactor cython interface: `copying.pyx` ([#10359](https://github.com/NVIDIA/cudf/pull/10359)) [@isVoid](https://github.com/isVoid) +- Implement a mixin for scans ([#10358](https://github.com/NVIDIA/cudf/pull/10358)) [@vyasr](https://github.com/vyasr) +- Add scan_aggregation and reduce_aggregation derived types. ([#10357](https://github.com/NVIDIA/cudf/pull/10357)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add cleanup of python artifacts ([#10355](https://github.com/NVIDIA/cudf/pull/10355)) [@galipremsagar](https://github.com/galipremsagar) +- Fix warnings in test_categorical.py. ([#10354](https://github.com/NVIDIA/cudf/pull/10354)) [@bdice](https://github.com/bdice) +- Create a dispatcher for invoking regex kernel functions ([#10349](https://github.com/NVIDIA/cudf/pull/10349)) [@davidwendt](https://github.com/davidwendt) +- Fix `codecov` in CI ([#10347](https://github.com/NVIDIA/cudf/pull/10347)) [@galipremsagar](https://github.com/galipremsagar) +- Enable caching for `memory_usage` calculation in `Column` ([#10345](https://github.com/NVIDIA/cudf/pull/10345)) [@galipremsagar](https://github.com/galipremsagar) +- C++17 cleanup: traits replace std::enable_if<>::type with std::enable_if_t ([#10343](https://github.com/NVIDIA/cudf/pull/10343)) [@karthikeyann](https://github.com/karthikeyann) +- JNI: Support appending DECIMAL128 into ColumnBuilder in terms of byte array ([#10338](https://github.com/NVIDIA/cudf/pull/10338)) [@sperlingxx](https://github.com/sperlingxx) +- multibyte_split test improvements ([#10328](https://github.com/NVIDIA/cudf/pull/10328)) [@vuule](https://github.com/vuule) +- Fix warnings in test_binops.py. ([#10327](https://github.com/NVIDIA/cudf/pull/10327)) [@bdice](https://github.com/bdice) +- Fix warnings from pandas in test_array_ufunc.py. ([#10324](https://github.com/NVIDIA/cudf/pull/10324)) [@bdice](https://github.com/bdice) +- Update upload script ([#10321](https://github.com/NVIDIA/cudf/pull/10321)) [@ajschmidt8](https://github.com/ajschmidt8) +- Move hash type declarations to hashing.hpp ([#10320](https://github.com/NVIDIA/cudf/pull/10320)) [@davidwendt](https://github.com/davidwendt) +- C++17 cleanup: traits replace `::value` with `_v` ([#10319](https://github.com/NVIDIA/cudf/pull/10319)) [@karthikeyann](https://github.com/karthikeyann) +- Remove internal columns usage ([#10315](https://github.com/NVIDIA/cudf/pull/10315)) [@vyasr](https://github.com/vyasr) +- Remove extraneous `build.sh` parameter ([#10313](https://github.com/NVIDIA/cudf/pull/10313)) [@ajschmidt8](https://github.com/ajschmidt8) +- Add const qualifier to MurmurHash3_32::hash_combine ([#10311](https://github.com/NVIDIA/cudf/pull/10311)) [@davidwendt](https://github.com/davidwendt) +- Remove `TODO` in `libcudf_kafka` recipe ([#10309](https://github.com/NVIDIA/cudf/pull/10309)) [@ajschmidt8](https://github.com/ajschmidt8) +- Add conversions between column_view and device_span<T const>. ([#10302](https://github.com/NVIDIA/cudf/pull/10302)) [@bdice](https://github.com/bdice) +- Avoid `decimal` type narrowing for decimal binops ([#10299](https://github.com/NVIDIA/cudf/pull/10299)) [@galipremsagar](https://github.com/galipremsagar) +- Deprecate `DataFrame.iteritems` and introduce `.items` ([#10298](https://github.com/NVIDIA/cudf/pull/10298)) [@galipremsagar](https://github.com/galipremsagar) +- Explicitly request CMake use `gnu++17` over `c++17` ([#10297](https://github.com/NVIDIA/cudf/pull/10297)) [@robertmaynard](https://github.com/robertmaynard) +- Add copyright check as pre-commit hook. ([#10290](https://github.com/NVIDIA/cudf/pull/10290)) [@vyasr](https://github.com/vyasr) +- DataFrame `insert` and creation optimizations ([#10285](https://github.com/NVIDIA/cudf/pull/10285)) [@galipremsagar](https://github.com/galipremsagar) +- Improve hash join detail functions ([#10273](https://github.com/NVIDIA/cudf/pull/10273)) [@PointKernel](https://github.com/PointKernel) +- Replace custom `cached_property` implementation with functools ([#10272](https://github.com/NVIDIA/cudf/pull/10272)) [@shwina](https://github.com/shwina) +- Rewrites `sample` API ([#10262](https://github.com/NVIDIA/cudf/pull/10262)) [@isVoid](https://github.com/isVoid) +- Bump hadoop-common from 3.1.0 to 3.1.4 in /java ([#10259](https://github.com/NVIDIA/cudf/pull/10259)) [@dependabot[bot]](https://github.com/dependabot[bot]) +- Remove making redundant `copy` across code-base ([#10257](https://github.com/NVIDIA/cudf/pull/10257)) [@galipremsagar](https://github.com/galipremsagar) +- Add more `nvtx` annotations ([#10256](https://github.com/NVIDIA/cudf/pull/10256)) [@galipremsagar](https://github.com/galipremsagar) +- Add `copyright` check in `cudf` ([#10253](https://github.com/NVIDIA/cudf/pull/10253)) [@galipremsagar](https://github.com/galipremsagar) +- Remove redundant copies in `fillna` to improve performance ([#10241](https://github.com/NVIDIA/cudf/pull/10241)) [@galipremsagar](https://github.com/galipremsagar) +- Remove `std::numeric_limit` specializations for timestamp & durations ([#10239](https://github.com/NVIDIA/cudf/pull/10239)) [@codereport](https://github.com/codereport) +- Optimize `DataFrame` creation across code-base ([#10236](https://github.com/NVIDIA/cudf/pull/10236)) [@galipremsagar](https://github.com/galipremsagar) +- Change pytest distribution algorithm and increase parallelism in CI ([#10232](https://github.com/NVIDIA/cudf/pull/10232)) [@galipremsagar](https://github.com/galipremsagar) +- Add environment variables for I/O thread pool and slice sizes ([#10218](https://github.com/NVIDIA/cudf/pull/10218)) [@vuule](https://github.com/vuule) +- Add regex flags to strings findall functions ([#10208](https://github.com/NVIDIA/cudf/pull/10208)) [@davidwendt](https://github.com/davidwendt) +- Update dask-cudf parquet tests to reflect upstream bugfixes to `_metadata` ([#10206](https://github.com/NVIDIA/cudf/pull/10206)) [@charlesbluca](https://github.com/charlesbluca) +- Remove unnecessary nunique function in Series. ([#10205](https://github.com/NVIDIA/cudf/pull/10205)) [@martinfalisse](https://github.com/martinfalisse) +- Refactor DataFrame tests. ([#10204](https://github.com/NVIDIA/cudf/pull/10204)) [@bdice](https://github.com/bdice) +- Rewrites `column.__setitem__`, Use `boolean_mask_scatter` ([#10202](https://github.com/NVIDIA/cudf/pull/10202)) [@isVoid](https://github.com/isVoid) +- Java utilities to aid in accelerating aggregations on 128-bit types ([#10201](https://github.com/NVIDIA/cudf/pull/10201)) [@jlowe](https://github.com/jlowe) +- Fix docstrings alignment in `Frame` methods ([#10199](https://github.com/NVIDIA/cudf/pull/10199)) [@galipremsagar](https://github.com/galipremsagar) +- Fix cuco pair issue in hash join ([#10195](https://github.com/NVIDIA/cudf/pull/10195)) [@PointKernel](https://github.com/PointKernel) +- Replace `dask` groupby `.index` usages with `.by` ([#10193](https://github.com/NVIDIA/cudf/pull/10193)) [@galipremsagar](https://github.com/galipremsagar) +- Add regex flags to strings extract function ([#10192](https://github.com/NVIDIA/cudf/pull/10192)) [@davidwendt](https://github.com/davidwendt) +- Forward-merge branch-22.02 to branch-22.04 ([#10191](https://github.com/NVIDIA/cudf/pull/10191)) [@bdice](https://github.com/bdice) +- Add CMake `install` rule for tests ([#10190](https://github.com/NVIDIA/cudf/pull/10190)) [@ajschmidt8](https://github.com/ajschmidt8) +- Unpin `dask` & `distributed` ([#10182](https://github.com/NVIDIA/cudf/pull/10182)) [@galipremsagar](https://github.com/galipremsagar) +- Add comments to explain test validation ([#10176](https://github.com/NVIDIA/cudf/pull/10176)) [@galipremsagar](https://github.com/galipremsagar) +- Reduce warnings in pytest output ([#10168](https://github.com/NVIDIA/cudf/pull/10168)) [@bdice](https://github.com/bdice) +- Some consolidation of indexed frame methods ([#10167](https://github.com/NVIDIA/cudf/pull/10167)) [@vyasr](https://github.com/vyasr) +- Refactor isin implementations ([#10165](https://github.com/NVIDIA/cudf/pull/10165)) [@vyasr](https://github.com/vyasr) +- Faster struct row comparator ([#10164](https://github.com/NVIDIA/cudf/pull/10164)) [@devavret](https://github.com/devavret) +- Refactor groupby::get_groups. ([#10161](https://github.com/NVIDIA/cudf/pull/10161)) [@bdice](https://github.com/bdice) +- Deprecate `decimal_cols_as_float` in ORC reader (C++ layer) ([#10152](https://github.com/NVIDIA/cudf/pull/10152)) [@vuule](https://github.com/vuule) +- Replace `ccache` with `sccache` ([#10146](https://github.com/NVIDIA/cudf/pull/10146)) [@ajschmidt8](https://github.com/ajschmidt8) +- Murmur3 hash kernel cleanup ([#10143](https://github.com/NVIDIA/cudf/pull/10143)) [@rwlee](https://github.com/rwlee) +- Deprecate `decimal_cols_as_float` in ORC reader ([#10142](https://github.com/NVIDIA/cudf/pull/10142)) [@galipremsagar](https://github.com/galipremsagar) +- Run pyupgrade 2.31.0. ([#10141](https://github.com/NVIDIA/cudf/pull/10141)) [@bdice](https://github.com/bdice) +- Remove `drop_nan` from internal `IndexedFrame._drop_na_rows`. ([#10140](https://github.com/NVIDIA/cudf/pull/10140)) [@bdice](https://github.com/bdice) +- Change cudf::strings::find_multiple to return a lists column ([#10134](https://github.com/NVIDIA/cudf/pull/10134)) [@davidwendt](https://github.com/davidwendt) +- Update cmake-format script for branch 22.04. ([#10132](https://github.com/NVIDIA/cudf/pull/10132)) [@bdice](https://github.com/bdice) +- Accept r-value references in convert_table_for_return(): ([#10131](https://github.com/NVIDIA/cudf/pull/10131)) [@mythrocks](https://github.com/mythrocks) +- Remove the option to completely disable decimal128 columns in the ORC reader ([#10127](https://github.com/NVIDIA/cudf/pull/10127)) [@vuule](https://github.com/vuule) +- Remove deprecated code ([#10124](https://github.com/NVIDIA/cudf/pull/10124)) [@vyasr](https://github.com/vyasr) +- Update gpu_utils.py to reflect current CUDA support. ([#10113](https://github.com/NVIDIA/cudf/pull/10113)) [@bdice](https://github.com/bdice) +- Remove benchmarks suffix ([#10112](https://github.com/NVIDIA/cudf/pull/10112)) [@bdice](https://github.com/bdice) +- Update cudf java binding version to 22.04.0-SNAPSHOT ([#10084](https://github.com/NVIDIA/cudf/pull/10084)) [@pxLi](https://github.com/pxLi) +- Remove unnecessary docker files. ([#10069](https://github.com/NVIDIA/cudf/pull/10069)) [@vyasr](https://github.com/vyasr) +- Limit benchmark iterations using environment variable ([#10060](https://github.com/NVIDIA/cudf/pull/10060)) [@karthikeyann](https://github.com/karthikeyann) +- Add timing chart for libcudf build metrics report page ([#10038](https://github.com/NVIDIA/cudf/pull/10038)) [@davidwendt](https://github.com/davidwendt) +- JNI: Rewrite growBuffersAndRows to accelerate the HostColumnBuilder ([#10025](https://github.com/NVIDIA/cudf/pull/10025)) [@sperlingxx](https://github.com/sperlingxx) +- Reduce redundant code in CUDF JNI ([#10019](https://github.com/NVIDIA/cudf/pull/10019)) [@mythrocks](https://github.com/mythrocks) +- Make snappy decompress check more efficient ([#9995](https://github.com/NVIDIA/cudf/pull/9995)) [@cheinger](https://github.com/cheinger) +- Remove deprecated method Series.set_index. ([#9945](https://github.com/NVIDIA/cudf/pull/9945)) [@bdice](https://github.com/bdice) +- Implement a mixin for reductions ([#9925](https://github.com/NVIDIA/cudf/pull/9925)) [@vyasr](https://github.com/vyasr) +- JNI: Push back decimal utils from spark-rapids ([#9907](https://github.com/NVIDIA/cudf/pull/9907)) [@sperlingxx](https://github.com/sperlingxx) +- Add `assert_column_memory_*` ([#9882](https://github.com/NVIDIA/cudf/pull/9882)) [@isVoid](https://github.com/isVoid) +- Add CUDF_UNREACHABLE macro. ([#9727](https://github.com/NVIDIA/cudf/pull/9727)) [@bdice](https://github.com/bdice) +- Upgrade `arrow` & `pyarrow` to `6.0.1` ([#9686](https://github.com/NVIDIA/cudf/pull/9686)) [@galipremsagar](https://github.com/galipremsagar) # cuDF 22.02.00 (2 Feb 2022) ## 🚨 Breaking Changes -- ORC writer API changes for granular statistics ([#10058](https://github.com/rapidsai/cudf/pull/10058)) [@mythrocks](https://github.com/mythrocks) -- `decimal128` Support for `to/from_arrow` ([#9986](https://github.com/rapidsai/cudf/pull/9986)) [@codereport](https://github.com/codereport) -- Remove deprecated method `one_hot_encoding` ([#9977](https://github.com/rapidsai/cudf/pull/9977)) [@isVoid](https://github.com/isVoid) -- Remove str.subword_tokenize ([#9968](https://github.com/rapidsai/cudf/pull/9968)) [@VibhuJawa](https://github.com/VibhuJawa) -- Remove deprecated `method` parameter from `merge` and `join`. ([#9944](https://github.com/rapidsai/cudf/pull/9944)) [@bdice](https://github.com/bdice) -- Remove deprecated method DataFrame.hash_columns. ([#9943](https://github.com/rapidsai/cudf/pull/9943)) [@bdice](https://github.com/bdice) -- Remove deprecated method Series.hash_encode. ([#9942](https://github.com/rapidsai/cudf/pull/9942)) [@bdice](https://github.com/bdice) -- Refactoring ceil/round/floor code for datetime64 types ([#9926](https://github.com/rapidsai/cudf/pull/9926)) [@mayankanand007](https://github.com/mayankanand007) -- Introduce `nan_as_null` parameter for `cudf.Index` ([#9893](https://github.com/rapidsai/cudf/pull/9893)) [@galipremsagar](https://github.com/galipremsagar) -- Add regex_flags parameter to strings replace_re functions ([#9878](https://github.com/rapidsai/cudf/pull/9878)) [@davidwendt](https://github.com/davidwendt) -- Break tie for `top` categorical columns in `Series.describe` ([#9867](https://github.com/rapidsai/cudf/pull/9867)) [@isVoid](https://github.com/isVoid) -- Add partitioning support in parquet writer ([#9810](https://github.com/rapidsai/cudf/pull/9810)) [@devavret](https://github.com/devavret) -- Move `drop_duplicates`, `drop_na`, `_gather`, `take` to IndexFrame and create their `_base_index` counterparts ([#9807](https://github.com/rapidsai/cudf/pull/9807)) [@isVoid](https://github.com/isVoid) -- Raise temporary error for `decimal128` types in parquet reader ([#9804](https://github.com/rapidsai/cudf/pull/9804)) [@galipremsagar](https://github.com/galipremsagar) -- Change default `dtype` of all nulls column from `float` to `object` ([#9803](https://github.com/rapidsai/cudf/pull/9803)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unused masked udf cython/c++ code ([#9792](https://github.com/rapidsai/cudf/pull/9792)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Pick smallest decimal type with required precision in ORC reader ([#9775](https://github.com/rapidsai/cudf/pull/9775)) [@vuule](https://github.com/vuule) -- Add decimal128 support to Parquet reader and writer ([#9765](https://github.com/rapidsai/cudf/pull/9765)) [@vuule](https://github.com/vuule) -- Refactor TableTest assertion methods to a separate utility class ([#9762](https://github.com/rapidsai/cudf/pull/9762)) [@jlowe](https://github.com/jlowe) -- Use cuFile direct device reads/writes by default in cuIO ([#9722](https://github.com/rapidsai/cudf/pull/9722)) [@vuule](https://github.com/vuule) -- Match pandas scalar result types in reductions ([#9717](https://github.com/rapidsai/cudf/pull/9717)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add parameters to control row group size in Parquet writer ([#9677](https://github.com/rapidsai/cudf/pull/9677)) [@vuule](https://github.com/vuule) -- Refactor bit counting APIs, introduce valid/null count functions, and split host/device side code for segmented counts. ([#9588](https://github.com/rapidsai/cudf/pull/9588)) [@bdice](https://github.com/bdice) -- Add support for `decimal128` in cudf python ([#9533](https://github.com/rapidsai/cudf/pull/9533)) [@galipremsagar](https://github.com/galipremsagar) -- Implement `lists::index_of()` to find positions in list rows ([#9510](https://github.com/rapidsai/cudf/pull/9510)) [@mythrocks](https://github.com/mythrocks) -- Rewriting row/column conversions for Spark <-> cudf data conversions ([#8444](https://github.com/rapidsai/cudf/pull/8444)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- ORC writer API changes for granular statistics ([#10058](https://github.com/NVIDIA/cudf/pull/10058)) [@mythrocks](https://github.com/mythrocks) +- `decimal128` Support for `to/from_arrow` ([#9986](https://github.com/NVIDIA/cudf/pull/9986)) [@codereport](https://github.com/codereport) +- Remove deprecated method `one_hot_encoding` ([#9977](https://github.com/NVIDIA/cudf/pull/9977)) [@isVoid](https://github.com/isVoid) +- Remove str.subword_tokenize ([#9968](https://github.com/NVIDIA/cudf/pull/9968)) [@VibhuJawa](https://github.com/VibhuJawa) +- Remove deprecated `method` parameter from `merge` and `join`. ([#9944](https://github.com/NVIDIA/cudf/pull/9944)) [@bdice](https://github.com/bdice) +- Remove deprecated method DataFrame.hash_columns. ([#9943](https://github.com/NVIDIA/cudf/pull/9943)) [@bdice](https://github.com/bdice) +- Remove deprecated method Series.hash_encode. ([#9942](https://github.com/NVIDIA/cudf/pull/9942)) [@bdice](https://github.com/bdice) +- Refactoring ceil/round/floor code for datetime64 types ([#9926](https://github.com/NVIDIA/cudf/pull/9926)) [@mayankanand007](https://github.com/mayankanand007) +- Introduce `nan_as_null` parameter for `cudf.Index` ([#9893](https://github.com/NVIDIA/cudf/pull/9893)) [@galipremsagar](https://github.com/galipremsagar) +- Add regex_flags parameter to strings replace_re functions ([#9878](https://github.com/NVIDIA/cudf/pull/9878)) [@davidwendt](https://github.com/davidwendt) +- Break tie for `top` categorical columns in `Series.describe` ([#9867](https://github.com/NVIDIA/cudf/pull/9867)) [@isVoid](https://github.com/isVoid) +- Add partitioning support in parquet writer ([#9810](https://github.com/NVIDIA/cudf/pull/9810)) [@devavret](https://github.com/devavret) +- Move `drop_duplicates`, `drop_na`, `_gather`, `take` to IndexFrame and create their `_base_index` counterparts ([#9807](https://github.com/NVIDIA/cudf/pull/9807)) [@isVoid](https://github.com/isVoid) +- Raise temporary error for `decimal128` types in parquet reader ([#9804](https://github.com/NVIDIA/cudf/pull/9804)) [@galipremsagar](https://github.com/galipremsagar) +- Change default `dtype` of all nulls column from `float` to `object` ([#9803](https://github.com/NVIDIA/cudf/pull/9803)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unused masked udf cython/c++ code ([#9792](https://github.com/NVIDIA/cudf/pull/9792)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Pick smallest decimal type with required precision in ORC reader ([#9775](https://github.com/NVIDIA/cudf/pull/9775)) [@vuule](https://github.com/vuule) +- Add decimal128 support to Parquet reader and writer ([#9765](https://github.com/NVIDIA/cudf/pull/9765)) [@vuule](https://github.com/vuule) +- Refactor TableTest assertion methods to a separate utility class ([#9762](https://github.com/NVIDIA/cudf/pull/9762)) [@jlowe](https://github.com/jlowe) +- Use cuFile direct device reads/writes by default in cuIO ([#9722](https://github.com/NVIDIA/cudf/pull/9722)) [@vuule](https://github.com/vuule) +- Match pandas scalar result types in reductions ([#9717](https://github.com/NVIDIA/cudf/pull/9717)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add parameters to control row group size in Parquet writer ([#9677](https://github.com/NVIDIA/cudf/pull/9677)) [@vuule](https://github.com/vuule) +- Refactor bit counting APIs, introduce valid/null count functions, and split host/device side code for segmented counts. ([#9588](https://github.com/NVIDIA/cudf/pull/9588)) [@bdice](https://github.com/bdice) +- Add support for `decimal128` in cudf python ([#9533](https://github.com/NVIDIA/cudf/pull/9533)) [@galipremsagar](https://github.com/galipremsagar) +- Implement `lists::index_of()` to find positions in list rows ([#9510](https://github.com/NVIDIA/cudf/pull/9510)) [@mythrocks](https://github.com/mythrocks) +- Rewriting row/column conversions for Spark <-> cudf data conversions ([#8444](https://github.com/NVIDIA/cudf/pull/8444)) [@hyperbolic2346](https://github.com/hyperbolic2346) ## 🐛 Bug Fixes -- Add check for negative stripe index in ORC reader ([#10074](https://github.com/rapidsai/cudf/pull/10074)) [@vuule](https://github.com/vuule) -- Update Java tests to expect DECIMAL128 from Arrow ([#10073](https://github.com/rapidsai/cudf/pull/10073)) [@jlowe](https://github.com/jlowe) -- Avoid index materialization when `DataFrame` is created with un-named `Series` objects ([#10071](https://github.com/rapidsai/cudf/pull/10071)) [@galipremsagar](https://github.com/galipremsagar) -- fix gcc 11 compilation errors ([#10067](https://github.com/rapidsai/cudf/pull/10067)) [@rongou](https://github.com/rongou) -- Fix `columns` ordering issue in parquet reader ([#10066](https://github.com/rapidsai/cudf/pull/10066)) [@galipremsagar](https://github.com/galipremsagar) -- Fix dataframe setitem with `ndarray` types ([#10056](https://github.com/rapidsai/cudf/pull/10056)) [@galipremsagar](https://github.com/galipremsagar) -- Remove implicit copy due to conversion from cudf::size_type and size_t ([#10045](https://github.com/rapidsai/cudf/pull/10045)) [@robertmaynard](https://github.com/robertmaynard) -- Include <optional> in headers that use std::optional ([#10044](https://github.com/rapidsai/cudf/pull/10044)) [@robertmaynard](https://github.com/robertmaynard) -- Fix repr and concat of `StructColumn` ([#10042](https://github.com/rapidsai/cudf/pull/10042)) [@galipremsagar](https://github.com/galipremsagar) -- Include row group level stats when writing ORC files ([#10041](https://github.com/rapidsai/cudf/pull/10041)) [@vuule](https://github.com/vuule) -- build.sh respects the `--build_metrics` and `--incl_cache_stats` flags ([#10035](https://github.com/rapidsai/cudf/pull/10035)) [@robertmaynard](https://github.com/robertmaynard) -- Fix memory leaks in JNI native code. ([#10029](https://github.com/rapidsai/cudf/pull/10029)) [@mythrocks](https://github.com/mythrocks) -- Update JNI to use new arena mr constructor ([#10027](https://github.com/rapidsai/cudf/pull/10027)) [@rongou](https://github.com/rongou) -- Fix null check when comparing structs in `arg_min` operation of reduction/groupby ([#10026](https://github.com/rapidsai/cudf/pull/10026)) [@ttnghia](https://github.com/ttnghia) -- Wrap CI script shell variables in quotes to fix local testing. ([#10018](https://github.com/rapidsai/cudf/pull/10018)) [@bdice](https://github.com/bdice) -- cudftestutil no longer propagates compiler flags to external users ([#10017](https://github.com/rapidsai/cudf/pull/10017)) [@robertmaynard](https://github.com/robertmaynard) -- Remove `CUDA_DEVICE_CALLABLE` macro usage ([#10015](https://github.com/rapidsai/cudf/pull/10015)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add missing list filling header in meta.yaml ([#10007](https://github.com/rapidsai/cudf/pull/10007)) [@devavret](https://github.com/devavret) -- Fix `conda` recipes for `custreamz` & `cudf_kafka` ([#10003](https://github.com/rapidsai/cudf/pull/10003)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix matching regex word-boundary () in strings replace ([#9997](https://github.com/rapidsai/cudf/pull/9997)) [@davidwendt](https://github.com/davidwendt) -- Fix null check when comparing structs in `min` and `max` reduction/groupby operations ([#9994](https://github.com/rapidsai/cudf/pull/9994)) [@ttnghia](https://github.com/ttnghia) -- Fix octal pattern matching in regex string ([#9993](https://github.com/rapidsai/cudf/pull/9993)) [@davidwendt](https://github.com/davidwendt) -- `decimal128` Support for `to/from_arrow` ([#9986](https://github.com/rapidsai/cudf/pull/9986)) [@codereport](https://github.com/codereport) -- Fix groupby shift/diff/fill after selecting from a `GroupBy` ([#9984](https://github.com/rapidsai/cudf/pull/9984)) [@shwina](https://github.com/shwina) -- Fix the overflow problem of decimal rescale ([#9966](https://github.com/rapidsai/cudf/pull/9966)) [@sperlingxx](https://github.com/sperlingxx) -- Use default value for decimal precision in parquet writer when not specified ([#9963](https://github.com/rapidsai/cudf/pull/9963)) [@devavret](https://github.com/devavret) -- Fix cudf java build error. ([#9958](https://github.com/rapidsai/cudf/pull/9958)) [@firestarman](https://github.com/firestarman) -- Use gpuci_mamba_retry to install local artifacts. ([#9951](https://github.com/rapidsai/cudf/pull/9951)) [@bdice](https://github.com/bdice) -- Fix regression HostColumnVectorCore requiring native libs ([#9948](https://github.com/rapidsai/cudf/pull/9948)) [@jlowe](https://github.com/jlowe) -- Rename aggregate_metadata in writer to fix name collision ([#9938](https://github.com/rapidsai/cudf/pull/9938)) [@devavret](https://github.com/devavret) -- Fixed issue with percentile_approx where output tdigests could have uninitialized data at the end. ([#9931](https://github.com/rapidsai/cudf/pull/9931)) [@nvdbaranec](https://github.com/nvdbaranec) -- Resolve racecheck errors in ORC kernels ([#9916](https://github.com/rapidsai/cudf/pull/9916)) [@vuule](https://github.com/vuule) -- Fix the java build after parquet partitioning support ([#9908](https://github.com/rapidsai/cudf/pull/9908)) [@revans2](https://github.com/revans2) -- Fix compilation of benchmark for parquet writer. ([#9905](https://github.com/rapidsai/cudf/pull/9905)) [@bdice](https://github.com/bdice) -- Fix a memcheck error in ORC writer ([#9896](https://github.com/rapidsai/cudf/pull/9896)) [@vuule](https://github.com/vuule) -- Introduce `nan_as_null` parameter for `cudf.Index` ([#9893](https://github.com/rapidsai/cudf/pull/9893)) [@galipremsagar](https://github.com/galipremsagar) -- Fix fallback to sort aggregation for grouping only hash aggregate ([#9891](https://github.com/rapidsai/cudf/pull/9891)) [@abellina](https://github.com/abellina) -- Add zlib to cudfjni link when using static libcudf library dependency ([#9890](https://github.com/rapidsai/cudf/pull/9890)) [@jlowe](https://github.com/jlowe) -- TimedeltaIndex constructor raises an AttributeError. ([#9884](https://github.com/rapidsai/cudf/pull/9884)) [@skirui-source](https://github.com/skirui-source) -- Fix cudf.Scalar string datetime construction ([#9875](https://github.com/rapidsai/cudf/pull/9875)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Load libcufile.so with RTLD_NODELETE flag ([#9872](https://github.com/rapidsai/cudf/pull/9872)) [@vuule](https://github.com/vuule) -- Break tie for `top` categorical columns in `Series.describe` ([#9867](https://github.com/rapidsai/cudf/pull/9867)) [@isVoid](https://github.com/isVoid) -- Fix null handling for structs `min` and `arg_min` in groupby, groupby scan, reduction, and inclusive_scan ([#9864](https://github.com/rapidsai/cudf/pull/9864)) [@ttnghia](https://github.com/ttnghia) -- Add one-level list encoding support in parquet reader ([#9848](https://github.com/rapidsai/cudf/pull/9848)) [@PointKernel](https://github.com/PointKernel) -- Fix an out-of-bounds read in validity copying in contiguous_split. ([#9842](https://github.com/rapidsai/cudf/pull/9842)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix join of MultiIndex to Index with one column and overlapping name. ([#9830](https://github.com/rapidsai/cudf/pull/9830)) [@vyasr](https://github.com/vyasr) -- Fix caching in `Series.applymap` ([#9821](https://github.com/rapidsai/cudf/pull/9821)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Enforce boolean `ascending` for dask-cudf `sort_values` ([#9814](https://github.com/rapidsai/cudf/pull/9814)) [@charlesbluca](https://github.com/charlesbluca) -- Fix ORC writer crash with empty input columns ([#9808](https://github.com/rapidsai/cudf/pull/9808)) [@vuule](https://github.com/vuule) -- Change default `dtype` of all nulls column from `float` to `object` ([#9803](https://github.com/rapidsai/cudf/pull/9803)) [@galipremsagar](https://github.com/galipremsagar) -- Load native dependencies when Java ColumnView is loaded ([#9800](https://github.com/rapidsai/cudf/pull/9800)) [@jlowe](https://github.com/jlowe) -- Fix dtype-argument bug in dask_cudf read_csv ([#9796](https://github.com/rapidsai/cudf/pull/9796)) [@rjzamora](https://github.com/rjzamora) -- Fix overflow for min calculation in strings::from_timestamps ([#9793](https://github.com/rapidsai/cudf/pull/9793)) [@revans2](https://github.com/revans2) -- Fix memory error due to lambda return type deduction limitation ([#9778](https://github.com/rapidsai/cudf/pull/9778)) [@karthikeyann](https://github.com/karthikeyann) -- Revert regex $/EOL end-of-string new-line special case handling ([#9774](https://github.com/rapidsai/cudf/pull/9774)) [@davidwendt](https://github.com/davidwendt) -- Fix missing streams ([#9767](https://github.com/rapidsai/cudf/pull/9767)) [@karthikeyann](https://github.com/karthikeyann) -- Fix make_empty_scalar_like on list_type ([#9759](https://github.com/rapidsai/cudf/pull/9759)) [@sperlingxx](https://github.com/sperlingxx) -- Update cmake and conda to 22.02 ([#9746](https://github.com/rapidsai/cudf/pull/9746)) [@devavret](https://github.com/devavret) -- Fix out-of-bounds memory write in decimal128-to-string conversion ([#9740](https://github.com/rapidsai/cudf/pull/9740)) [@davidwendt](https://github.com/davidwendt) -- Match pandas scalar result types in reductions ([#9717](https://github.com/rapidsai/cudf/pull/9717)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix regex non-multiline EOL/$ matching strings ending with a new-line ([#9715](https://github.com/rapidsai/cudf/pull/9715)) [@davidwendt](https://github.com/davidwendt) -- Fixed build by adding more checks for int8, int16 ([#9707](https://github.com/rapidsai/cudf/pull/9707)) [@razajafri](https://github.com/razajafri) -- Fix `null` handling when `boolean` dtype is passed ([#9691](https://github.com/rapidsai/cudf/pull/9691)) [@galipremsagar](https://github.com/galipremsagar) -- Fix stream usage in `segmented_gather()` ([#9679](https://github.com/rapidsai/cudf/pull/9679)) [@mythrocks](https://github.com/mythrocks) +- Add check for negative stripe index in ORC reader ([#10074](https://github.com/NVIDIA/cudf/pull/10074)) [@vuule](https://github.com/vuule) +- Update Java tests to expect DECIMAL128 from Arrow ([#10073](https://github.com/NVIDIA/cudf/pull/10073)) [@jlowe](https://github.com/jlowe) +- Avoid index materialization when `DataFrame` is created with un-named `Series` objects ([#10071](https://github.com/NVIDIA/cudf/pull/10071)) [@galipremsagar](https://github.com/galipremsagar) +- fix gcc 11 compilation errors ([#10067](https://github.com/NVIDIA/cudf/pull/10067)) [@rongou](https://github.com/rongou) +- Fix `columns` ordering issue in parquet reader ([#10066](https://github.com/NVIDIA/cudf/pull/10066)) [@galipremsagar](https://github.com/galipremsagar) +- Fix dataframe setitem with `ndarray` types ([#10056](https://github.com/NVIDIA/cudf/pull/10056)) [@galipremsagar](https://github.com/galipremsagar) +- Remove implicit copy due to conversion from cudf::size_type and size_t ([#10045](https://github.com/NVIDIA/cudf/pull/10045)) [@robertmaynard](https://github.com/robertmaynard) +- Include <optional> in headers that use std::optional ([#10044](https://github.com/NVIDIA/cudf/pull/10044)) [@robertmaynard](https://github.com/robertmaynard) +- Fix repr and concat of `StructColumn` ([#10042](https://github.com/NVIDIA/cudf/pull/10042)) [@galipremsagar](https://github.com/galipremsagar) +- Include row group level stats when writing ORC files ([#10041](https://github.com/NVIDIA/cudf/pull/10041)) [@vuule](https://github.com/vuule) +- build.sh respects the `--build_metrics` and `--incl_cache_stats` flags ([#10035](https://github.com/NVIDIA/cudf/pull/10035)) [@robertmaynard](https://github.com/robertmaynard) +- Fix memory leaks in JNI native code. ([#10029](https://github.com/NVIDIA/cudf/pull/10029)) [@mythrocks](https://github.com/mythrocks) +- Update JNI to use new arena mr constructor ([#10027](https://github.com/NVIDIA/cudf/pull/10027)) [@rongou](https://github.com/rongou) +- Fix null check when comparing structs in `arg_min` operation of reduction/groupby ([#10026](https://github.com/NVIDIA/cudf/pull/10026)) [@ttnghia](https://github.com/ttnghia) +- Wrap CI script shell variables in quotes to fix local testing. ([#10018](https://github.com/NVIDIA/cudf/pull/10018)) [@bdice](https://github.com/bdice) +- cudftestutil no longer propagates compiler flags to external users ([#10017](https://github.com/NVIDIA/cudf/pull/10017)) [@robertmaynard](https://github.com/robertmaynard) +- Remove `CUDA_DEVICE_CALLABLE` macro usage ([#10015](https://github.com/NVIDIA/cudf/pull/10015)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add missing list filling header in meta.yaml ([#10007](https://github.com/NVIDIA/cudf/pull/10007)) [@devavret](https://github.com/devavret) +- Fix `conda` recipes for `custreamz` & `cudf_kafka` ([#10003](https://github.com/NVIDIA/cudf/pull/10003)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix matching regex word-boundary () in strings replace ([#9997](https://github.com/NVIDIA/cudf/pull/9997)) [@davidwendt](https://github.com/davidwendt) +- Fix null check when comparing structs in `min` and `max` reduction/groupby operations ([#9994](https://github.com/NVIDIA/cudf/pull/9994)) [@ttnghia](https://github.com/ttnghia) +- Fix octal pattern matching in regex string ([#9993](https://github.com/NVIDIA/cudf/pull/9993)) [@davidwendt](https://github.com/davidwendt) +- `decimal128` Support for `to/from_arrow` ([#9986](https://github.com/NVIDIA/cudf/pull/9986)) [@codereport](https://github.com/codereport) +- Fix groupby shift/diff/fill after selecting from a `GroupBy` ([#9984](https://github.com/NVIDIA/cudf/pull/9984)) [@shwina](https://github.com/shwina) +- Fix the overflow problem of decimal rescale ([#9966](https://github.com/NVIDIA/cudf/pull/9966)) [@sperlingxx](https://github.com/sperlingxx) +- Use default value for decimal precision in parquet writer when not specified ([#9963](https://github.com/NVIDIA/cudf/pull/9963)) [@devavret](https://github.com/devavret) +- Fix cudf java build error. ([#9958](https://github.com/NVIDIA/cudf/pull/9958)) [@firestarman](https://github.com/firestarman) +- Use gpuci_mamba_retry to install local artifacts. ([#9951](https://github.com/NVIDIA/cudf/pull/9951)) [@bdice](https://github.com/bdice) +- Fix regression HostColumnVectorCore requiring native libs ([#9948](https://github.com/NVIDIA/cudf/pull/9948)) [@jlowe](https://github.com/jlowe) +- Rename aggregate_metadata in writer to fix name collision ([#9938](https://github.com/NVIDIA/cudf/pull/9938)) [@devavret](https://github.com/devavret) +- Fixed issue with percentile_approx where output tdigests could have uninitialized data at the end. ([#9931](https://github.com/NVIDIA/cudf/pull/9931)) [@nvdbaranec](https://github.com/nvdbaranec) +- Resolve racecheck errors in ORC kernels ([#9916](https://github.com/NVIDIA/cudf/pull/9916)) [@vuule](https://github.com/vuule) +- Fix the java build after parquet partitioning support ([#9908](https://github.com/NVIDIA/cudf/pull/9908)) [@revans2](https://github.com/revans2) +- Fix compilation of benchmark for parquet writer. ([#9905](https://github.com/NVIDIA/cudf/pull/9905)) [@bdice](https://github.com/bdice) +- Fix a memcheck error in ORC writer ([#9896](https://github.com/NVIDIA/cudf/pull/9896)) [@vuule](https://github.com/vuule) +- Introduce `nan_as_null` parameter for `cudf.Index` ([#9893](https://github.com/NVIDIA/cudf/pull/9893)) [@galipremsagar](https://github.com/galipremsagar) +- Fix fallback to sort aggregation for grouping only hash aggregate ([#9891](https://github.com/NVIDIA/cudf/pull/9891)) [@abellina](https://github.com/abellina) +- Add zlib to cudfjni link when using static libcudf library dependency ([#9890](https://github.com/NVIDIA/cudf/pull/9890)) [@jlowe](https://github.com/jlowe) +- TimedeltaIndex constructor raises an AttributeError. ([#9884](https://github.com/NVIDIA/cudf/pull/9884)) [@skirui-source](https://github.com/skirui-source) +- Fix cudf.Scalar string datetime construction ([#9875](https://github.com/NVIDIA/cudf/pull/9875)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Load libcufile.so with RTLD_NODELETE flag ([#9872](https://github.com/NVIDIA/cudf/pull/9872)) [@vuule](https://github.com/vuule) +- Break tie for `top` categorical columns in `Series.describe` ([#9867](https://github.com/NVIDIA/cudf/pull/9867)) [@isVoid](https://github.com/isVoid) +- Fix null handling for structs `min` and `arg_min` in groupby, groupby scan, reduction, and inclusive_scan ([#9864](https://github.com/NVIDIA/cudf/pull/9864)) [@ttnghia](https://github.com/ttnghia) +- Add one-level list encoding support in parquet reader ([#9848](https://github.com/NVIDIA/cudf/pull/9848)) [@PointKernel](https://github.com/PointKernel) +- Fix an out-of-bounds read in validity copying in contiguous_split. ([#9842](https://github.com/NVIDIA/cudf/pull/9842)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix join of MultiIndex to Index with one column and overlapping name. ([#9830](https://github.com/NVIDIA/cudf/pull/9830)) [@vyasr](https://github.com/vyasr) +- Fix caching in `Series.applymap` ([#9821](https://github.com/NVIDIA/cudf/pull/9821)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Enforce boolean `ascending` for dask-cudf `sort_values` ([#9814](https://github.com/NVIDIA/cudf/pull/9814)) [@charlesbluca](https://github.com/charlesbluca) +- Fix ORC writer crash with empty input columns ([#9808](https://github.com/NVIDIA/cudf/pull/9808)) [@vuule](https://github.com/vuule) +- Change default `dtype` of all nulls column from `float` to `object` ([#9803](https://github.com/NVIDIA/cudf/pull/9803)) [@galipremsagar](https://github.com/galipremsagar) +- Load native dependencies when Java ColumnView is loaded ([#9800](https://github.com/NVIDIA/cudf/pull/9800)) [@jlowe](https://github.com/jlowe) +- Fix dtype-argument bug in dask_cudf read_csv ([#9796](https://github.com/NVIDIA/cudf/pull/9796)) [@rjzamora](https://github.com/rjzamora) +- Fix overflow for min calculation in strings::from_timestamps ([#9793](https://github.com/NVIDIA/cudf/pull/9793)) [@revans2](https://github.com/revans2) +- Fix memory error due to lambda return type deduction limitation ([#9778](https://github.com/NVIDIA/cudf/pull/9778)) [@karthikeyann](https://github.com/karthikeyann) +- Revert regex $/EOL end-of-string new-line special case handling ([#9774](https://github.com/NVIDIA/cudf/pull/9774)) [@davidwendt](https://github.com/davidwendt) +- Fix missing streams ([#9767](https://github.com/NVIDIA/cudf/pull/9767)) [@karthikeyann](https://github.com/karthikeyann) +- Fix make_empty_scalar_like on list_type ([#9759](https://github.com/NVIDIA/cudf/pull/9759)) [@sperlingxx](https://github.com/sperlingxx) +- Update cmake and conda to 22.02 ([#9746](https://github.com/NVIDIA/cudf/pull/9746)) [@devavret](https://github.com/devavret) +- Fix out-of-bounds memory write in decimal128-to-string conversion ([#9740](https://github.com/NVIDIA/cudf/pull/9740)) [@davidwendt](https://github.com/davidwendt) +- Match pandas scalar result types in reductions ([#9717](https://github.com/NVIDIA/cudf/pull/9717)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix regex non-multiline EOL/$ matching strings ending with a new-line ([#9715](https://github.com/NVIDIA/cudf/pull/9715)) [@davidwendt](https://github.com/davidwendt) +- Fixed build by adding more checks for int8, int16 ([#9707](https://github.com/NVIDIA/cudf/pull/9707)) [@razajafri](https://github.com/razajafri) +- Fix `null` handling when `boolean` dtype is passed ([#9691](https://github.com/NVIDIA/cudf/pull/9691)) [@galipremsagar](https://github.com/galipremsagar) +- Fix stream usage in `segmented_gather()` ([#9679](https://github.com/NVIDIA/cudf/pull/9679)) [@mythrocks](https://github.com/mythrocks) ## 📖 Documentation -- Update `decimal` dtypes related docs entries ([#10072](https://github.com/rapidsai/cudf/pull/10072)) [@galipremsagar](https://github.com/galipremsagar) -- Fix regex doc describing hexadecimal escape characters ([#10009](https://github.com/rapidsai/cudf/pull/10009)) [@davidwendt](https://github.com/davidwendt) -- Fix cudf compilation instructions. ([#9956](https://github.com/rapidsai/cudf/pull/9956)) [@esoha-nvidia](https://github.com/esoha-nvidia) -- Fix see also links for IO APIs ([#9895](https://github.com/rapidsai/cudf/pull/9895)) [@galipremsagar](https://github.com/galipremsagar) -- Fix build instructions for libcudf doxygen ([#9837](https://github.com/rapidsai/cudf/pull/9837)) [@davidwendt](https://github.com/davidwendt) -- Fix some doxygen warnings and add missing documentation ([#9770](https://github.com/rapidsai/cudf/pull/9770)) [@karthikeyann](https://github.com/karthikeyann) -- update cuda version in local build ([#9736](https://github.com/rapidsai/cudf/pull/9736)) [@karthikeyann](https://github.com/karthikeyann) -- Fix doxygen for enum types in libcudf ([#9724](https://github.com/rapidsai/cudf/pull/9724)) [@davidwendt](https://github.com/davidwendt) -- Spell check fixes ([#9682](https://github.com/rapidsai/cudf/pull/9682)) [@karthikeyann](https://github.com/karthikeyann) -- Fix links in C++ Developer Guide. ([#9675](https://github.com/rapidsai/cudf/pull/9675)) [@bdice](https://github.com/bdice) +- Update `decimal` dtypes related docs entries ([#10072](https://github.com/NVIDIA/cudf/pull/10072)) [@galipremsagar](https://github.com/galipremsagar) +- Fix regex doc describing hexadecimal escape characters ([#10009](https://github.com/NVIDIA/cudf/pull/10009)) [@davidwendt](https://github.com/davidwendt) +- Fix cudf compilation instructions. ([#9956](https://github.com/NVIDIA/cudf/pull/9956)) [@esoha-nvidia](https://github.com/esoha-nvidia) +- Fix see also links for IO APIs ([#9895](https://github.com/NVIDIA/cudf/pull/9895)) [@galipremsagar](https://github.com/galipremsagar) +- Fix build instructions for libcudf doxygen ([#9837](https://github.com/NVIDIA/cudf/pull/9837)) [@davidwendt](https://github.com/davidwendt) +- Fix some doxygen warnings and add missing documentation ([#9770](https://github.com/NVIDIA/cudf/pull/9770)) [@karthikeyann](https://github.com/karthikeyann) +- update cuda version in local build ([#9736](https://github.com/NVIDIA/cudf/pull/9736)) [@karthikeyann](https://github.com/karthikeyann) +- Fix doxygen for enum types in libcudf ([#9724](https://github.com/NVIDIA/cudf/pull/9724)) [@davidwendt](https://github.com/davidwendt) +- Spell check fixes ([#9682](https://github.com/NVIDIA/cudf/pull/9682)) [@karthikeyann](https://github.com/karthikeyann) +- Fix links in C++ Developer Guide. ([#9675](https://github.com/NVIDIA/cudf/pull/9675)) [@bdice](https://github.com/bdice) ## 🚀 New Features -- Remove libcudacxx patch needed for nvcc 11.4 ([#10057](https://github.com/rapidsai/cudf/pull/10057)) [@robertmaynard](https://github.com/robertmaynard) -- Allow CuPy 10 ([#10048](https://github.com/rapidsai/cudf/pull/10048)) [@jakirkham](https://github.com/jakirkham) -- Add in support for NULL_LOGICAL_AND and NULL_LOGICAL_OR binops ([#10016](https://github.com/rapidsai/cudf/pull/10016)) [@revans2](https://github.com/revans2) -- Add `groupby.transform` (only support for aggregations) ([#10005](https://github.com/rapidsai/cudf/pull/10005)) [@shwina](https://github.com/shwina) -- Add partitioning support to Parquet chunked writer ([#10000](https://github.com/rapidsai/cudf/pull/10000)) [@devavret](https://github.com/devavret) -- Add jni for sequences ([#9972](https://github.com/rapidsai/cudf/pull/9972)) [@wbo4958](https://github.com/wbo4958) -- Java bindings for mixed left, inner, and full joins ([#9941](https://github.com/rapidsai/cudf/pull/9941)) [@jlowe](https://github.com/jlowe) -- Java bindings for JSON reader support ([#9940](https://github.com/rapidsai/cudf/pull/9940)) [@wbo4958](https://github.com/wbo4958) -- Enable transpose for string columns in cudf python ([#9937](https://github.com/rapidsai/cudf/pull/9937)) [@galipremsagar](https://github.com/galipremsagar) -- Support structs for `cudf::contains` with column/scalar input ([#9929](https://github.com/rapidsai/cudf/pull/9929)) [@ttnghia](https://github.com/ttnghia) -- Implement mixed equality/conditional joins ([#9917](https://github.com/rapidsai/cudf/pull/9917)) [@vyasr](https://github.com/vyasr) -- Add cudf::strings::extract_all API ([#9909](https://github.com/rapidsai/cudf/pull/9909)) [@davidwendt](https://github.com/davidwendt) -- Implement JNI for `cudf::scatter` APIs ([#9903](https://github.com/rapidsai/cudf/pull/9903)) [@ttnghia](https://github.com/ttnghia) -- JNI: Function to copy and set validity from bool column. ([#9901](https://github.com/rapidsai/cudf/pull/9901)) [@mythrocks](https://github.com/mythrocks) -- Add dictionary support to cudf::copy_if_else ([#9887](https://github.com/rapidsai/cudf/pull/9887)) [@davidwendt](https://github.com/davidwendt) -- add run_benchmarks target for running benchmarks with json output ([#9879](https://github.com/rapidsai/cudf/pull/9879)) [@karthikeyann](https://github.com/karthikeyann) -- Add regex_flags parameter to strings replace_re functions ([#9878](https://github.com/rapidsai/cudf/pull/9878)) [@davidwendt](https://github.com/davidwendt) -- Add_suffix and add_prefix for DataFrames and Series ([#9846](https://github.com/rapidsai/cudf/pull/9846)) [@mayankanand007](https://github.com/mayankanand007) -- Add JNI for `cudf::drop_duplicates` ([#9841](https://github.com/rapidsai/cudf/pull/9841)) [@ttnghia](https://github.com/ttnghia) -- Implement per-list sequence ([#9839](https://github.com/rapidsai/cudf/pull/9839)) [@ttnghia](https://github.com/ttnghia) -- adding `series.transpose` ([#9835](https://github.com/rapidsai/cudf/pull/9835)) [@mayankanand007](https://github.com/mayankanand007) -- Adding support for `Series.autocorr` ([#9833](https://github.com/rapidsai/cudf/pull/9833)) [@mayankanand007](https://github.com/mayankanand007) -- Support round operation on datetime64 datatypes ([#9820](https://github.com/rapidsai/cudf/pull/9820)) [@mayankanand007](https://github.com/mayankanand007) -- Add partitioning support in parquet writer ([#9810](https://github.com/rapidsai/cudf/pull/9810)) [@devavret](https://github.com/devavret) -- Raise temporary error for `decimal128` types in parquet reader ([#9804](https://github.com/rapidsai/cudf/pull/9804)) [@galipremsagar](https://github.com/galipremsagar) -- Add decimal128 support to Parquet reader and writer ([#9765](https://github.com/rapidsai/cudf/pull/9765)) [@vuule](https://github.com/vuule) -- Optimize `groupby::scan` ([#9754](https://github.com/rapidsai/cudf/pull/9754)) [@PointKernel](https://github.com/PointKernel) -- Add sample JNI API ([#9728](https://github.com/rapidsai/cudf/pull/9728)) [@res-life](https://github.com/res-life) -- Support `min` and `max` in inclusive scan for structs ([#9725](https://github.com/rapidsai/cudf/pull/9725)) [@ttnghia](https://github.com/ttnghia) -- Add `first` and `last` method to `IndexedFrame` ([#9710](https://github.com/rapidsai/cudf/pull/9710)) [@isVoid](https://github.com/isVoid) -- Support `min` and `max` reduction for structs ([#9697](https://github.com/rapidsai/cudf/pull/9697)) [@ttnghia](https://github.com/ttnghia) -- Add parameters to control row group size in Parquet writer ([#9677](https://github.com/rapidsai/cudf/pull/9677)) [@vuule](https://github.com/vuule) -- Run compute-sanitizer in nightly build ([#9641](https://github.com/rapidsai/cudf/pull/9641)) [@karthikeyann](https://github.com/karthikeyann) -- Implement Series.datetime.floor ([#9571](https://github.com/rapidsai/cudf/pull/9571)) [@skirui-source](https://github.com/skirui-source) -- ceil/floor for `DatetimeIndex` ([#9554](https://github.com/rapidsai/cudf/pull/9554)) [@mayankanand007](https://github.com/mayankanand007) -- Add support for `decimal128` in cudf python ([#9533](https://github.com/rapidsai/cudf/pull/9533)) [@galipremsagar](https://github.com/galipremsagar) -- Implement `lists::index_of()` to find positions in list rows ([#9510](https://github.com/rapidsai/cudf/pull/9510)) [@mythrocks](https://github.com/mythrocks) -- custreamz oauth callback for kafka (librdkafka) ([#9486](https://github.com/rapidsai/cudf/pull/9486)) [@jdye64](https://github.com/jdye64) -- Add Pearson correlation for sort groupby (python) ([#9166](https://github.com/rapidsai/cudf/pull/9166)) [@skirui-source](https://github.com/skirui-source) -- Interchange dataframe protocol ([#9071](https://github.com/rapidsai/cudf/pull/9071)) [@iskode](https://github.com/iskode) -- Rewriting row/column conversions for Spark <-> cudf data conversions ([#8444](https://github.com/rapidsai/cudf/pull/8444)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Remove libcudacxx patch needed for nvcc 11.4 ([#10057](https://github.com/NVIDIA/cudf/pull/10057)) [@robertmaynard](https://github.com/robertmaynard) +- Allow CuPy 10 ([#10048](https://github.com/NVIDIA/cudf/pull/10048)) [@jakirkham](https://github.com/jakirkham) +- Add in support for NULL_LOGICAL_AND and NULL_LOGICAL_OR binops ([#10016](https://github.com/NVIDIA/cudf/pull/10016)) [@revans2](https://github.com/revans2) +- Add `groupby.transform` (only support for aggregations) ([#10005](https://github.com/NVIDIA/cudf/pull/10005)) [@shwina](https://github.com/shwina) +- Add partitioning support to Parquet chunked writer ([#10000](https://github.com/NVIDIA/cudf/pull/10000)) [@devavret](https://github.com/devavret) +- Add jni for sequences ([#9972](https://github.com/NVIDIA/cudf/pull/9972)) [@wbo4958](https://github.com/wbo4958) +- Java bindings for mixed left, inner, and full joins ([#9941](https://github.com/NVIDIA/cudf/pull/9941)) [@jlowe](https://github.com/jlowe) +- Java bindings for JSON reader support ([#9940](https://github.com/NVIDIA/cudf/pull/9940)) [@wbo4958](https://github.com/wbo4958) +- Enable transpose for string columns in cudf python ([#9937](https://github.com/NVIDIA/cudf/pull/9937)) [@galipremsagar](https://github.com/galipremsagar) +- Support structs for `cudf::contains` with column/scalar input ([#9929](https://github.com/NVIDIA/cudf/pull/9929)) [@ttnghia](https://github.com/ttnghia) +- Implement mixed equality/conditional joins ([#9917](https://github.com/NVIDIA/cudf/pull/9917)) [@vyasr](https://github.com/vyasr) +- Add cudf::strings::extract_all API ([#9909](https://github.com/NVIDIA/cudf/pull/9909)) [@davidwendt](https://github.com/davidwendt) +- Implement JNI for `cudf::scatter` APIs ([#9903](https://github.com/NVIDIA/cudf/pull/9903)) [@ttnghia](https://github.com/ttnghia) +- JNI: Function to copy and set validity from bool column. ([#9901](https://github.com/NVIDIA/cudf/pull/9901)) [@mythrocks](https://github.com/mythrocks) +- Add dictionary support to cudf::copy_if_else ([#9887](https://github.com/NVIDIA/cudf/pull/9887)) [@davidwendt](https://github.com/davidwendt) +- add run_benchmarks target for running benchmarks with json output ([#9879](https://github.com/NVIDIA/cudf/pull/9879)) [@karthikeyann](https://github.com/karthikeyann) +- Add regex_flags parameter to strings replace_re functions ([#9878](https://github.com/NVIDIA/cudf/pull/9878)) [@davidwendt](https://github.com/davidwendt) +- Add_suffix and add_prefix for DataFrames and Series ([#9846](https://github.com/NVIDIA/cudf/pull/9846)) [@mayankanand007](https://github.com/mayankanand007) +- Add JNI for `cudf::drop_duplicates` ([#9841](https://github.com/NVIDIA/cudf/pull/9841)) [@ttnghia](https://github.com/ttnghia) +- Implement per-list sequence ([#9839](https://github.com/NVIDIA/cudf/pull/9839)) [@ttnghia](https://github.com/ttnghia) +- adding `series.transpose` ([#9835](https://github.com/NVIDIA/cudf/pull/9835)) [@mayankanand007](https://github.com/mayankanand007) +- Adding support for `Series.autocorr` ([#9833](https://github.com/NVIDIA/cudf/pull/9833)) [@mayankanand007](https://github.com/mayankanand007) +- Support round operation on datetime64 datatypes ([#9820](https://github.com/NVIDIA/cudf/pull/9820)) [@mayankanand007](https://github.com/mayankanand007) +- Add partitioning support in parquet writer ([#9810](https://github.com/NVIDIA/cudf/pull/9810)) [@devavret](https://github.com/devavret) +- Raise temporary error for `decimal128` types in parquet reader ([#9804](https://github.com/NVIDIA/cudf/pull/9804)) [@galipremsagar](https://github.com/galipremsagar) +- Add decimal128 support to Parquet reader and writer ([#9765](https://github.com/NVIDIA/cudf/pull/9765)) [@vuule](https://github.com/vuule) +- Optimize `groupby::scan` ([#9754](https://github.com/NVIDIA/cudf/pull/9754)) [@PointKernel](https://github.com/PointKernel) +- Add sample JNI API ([#9728](https://github.com/NVIDIA/cudf/pull/9728)) [@res-life](https://github.com/res-life) +- Support `min` and `max` in inclusive scan for structs ([#9725](https://github.com/NVIDIA/cudf/pull/9725)) [@ttnghia](https://github.com/ttnghia) +- Add `first` and `last` method to `IndexedFrame` ([#9710](https://github.com/NVIDIA/cudf/pull/9710)) [@isVoid](https://github.com/isVoid) +- Support `min` and `max` reduction for structs ([#9697](https://github.com/NVIDIA/cudf/pull/9697)) [@ttnghia](https://github.com/ttnghia) +- Add parameters to control row group size in Parquet writer ([#9677](https://github.com/NVIDIA/cudf/pull/9677)) [@vuule](https://github.com/vuule) +- Run compute-sanitizer in nightly build ([#9641](https://github.com/NVIDIA/cudf/pull/9641)) [@karthikeyann](https://github.com/karthikeyann) +- Implement Series.datetime.floor ([#9571](https://github.com/NVIDIA/cudf/pull/9571)) [@skirui-source](https://github.com/skirui-source) +- ceil/floor for `DatetimeIndex` ([#9554](https://github.com/NVIDIA/cudf/pull/9554)) [@mayankanand007](https://github.com/mayankanand007) +- Add support for `decimal128` in cudf python ([#9533](https://github.com/NVIDIA/cudf/pull/9533)) [@galipremsagar](https://github.com/galipremsagar) +- Implement `lists::index_of()` to find positions in list rows ([#9510](https://github.com/NVIDIA/cudf/pull/9510)) [@mythrocks](https://github.com/mythrocks) +- custreamz oauth callback for kafka (librdkafka) ([#9486](https://github.com/NVIDIA/cudf/pull/9486)) [@jdye64](https://github.com/jdye64) +- Add Pearson correlation for sort groupby (python) ([#9166](https://github.com/NVIDIA/cudf/pull/9166)) [@skirui-source](https://github.com/skirui-source) +- Interchange dataframe protocol ([#9071](https://github.com/NVIDIA/cudf/pull/9071)) [@iskode](https://github.com/iskode) +- Rewriting row/column conversions for Spark <-> cudf data conversions ([#8444](https://github.com/NVIDIA/cudf/pull/8444)) [@hyperbolic2346](https://github.com/hyperbolic2346) ## 🛠️ Improvements -- Prepare upload scripts for Python 3.7 removal ([#10092](https://github.com/rapidsai/cudf/pull/10092)) [@Ethyling](https://github.com/Ethyling) -- Simplify custreamz and cudf_kafka recipes files ([#10065](https://github.com/rapidsai/cudf/pull/10065)) [@Ethyling](https://github.com/Ethyling) -- ORC writer API changes for granular statistics ([#10058](https://github.com/rapidsai/cudf/pull/10058)) [@mythrocks](https://github.com/mythrocks) -- Remove python constraints in cutreamz and cudf_kafka recipes ([#10052](https://github.com/rapidsai/cudf/pull/10052)) [@Ethyling](https://github.com/Ethyling) -- Unpin `dask` and `distributed` in CI ([#10028](https://github.com/rapidsai/cudf/pull/10028)) [@galipremsagar](https://github.com/galipremsagar) -- Add `_from_column_like_self` factory ([#10022](https://github.com/rapidsai/cudf/pull/10022)) [@isVoid](https://github.com/isVoid) -- Replace custom CUDA bindings previously provided by RMM with official CUDA Python bindings ([#10008](https://github.com/rapidsai/cudf/pull/10008)) [@shwina](https://github.com/shwina) -- Use `cuda::std::is_arithmetic` in `cudf::is_numeric` trait. ([#9996](https://github.com/rapidsai/cudf/pull/9996)) [@bdice](https://github.com/bdice) -- Clean up CUDA stream use in cuIO ([#9991](https://github.com/rapidsai/cudf/pull/9991)) [@vuule](https://github.com/vuule) -- Use addressed-ordered first fit for the pinned memory pool ([#9989](https://github.com/rapidsai/cudf/pull/9989)) [@rongou](https://github.com/rongou) -- Add strings tests to transpose_test.cpp ([#9985](https://github.com/rapidsai/cudf/pull/9985)) [@davidwendt](https://github.com/davidwendt) -- Use gpuci_mamba_retry on Java CI. ([#9983](https://github.com/rapidsai/cudf/pull/9983)) [@bdice](https://github.com/bdice) -- Remove deprecated method `one_hot_encoding` ([#9977](https://github.com/rapidsai/cudf/pull/9977)) [@isVoid](https://github.com/isVoid) -- Minor cleanup of unused Python functions ([#9974](https://github.com/rapidsai/cudf/pull/9974)) [@vyasr](https://github.com/vyasr) -- Use new efficient partitioned parquet writing in cuDF ([#9971](https://github.com/rapidsai/cudf/pull/9971)) [@devavret](https://github.com/devavret) -- Remove str.subword_tokenize ([#9968](https://github.com/rapidsai/cudf/pull/9968)) [@VibhuJawa](https://github.com/VibhuJawa) -- Forward-merge branch-21.12 to branch-22.02 ([#9947](https://github.com/rapidsai/cudf/pull/9947)) [@bdice](https://github.com/bdice) -- Remove deprecated `method` parameter from `merge` and `join`. ([#9944](https://github.com/rapidsai/cudf/pull/9944)) [@bdice](https://github.com/bdice) -- Remove deprecated method DataFrame.hash_columns. ([#9943](https://github.com/rapidsai/cudf/pull/9943)) [@bdice](https://github.com/bdice) -- Remove deprecated method Series.hash_encode. ([#9942](https://github.com/rapidsai/cudf/pull/9942)) [@bdice](https://github.com/bdice) -- use ninja in java ci build ([#9933](https://github.com/rapidsai/cudf/pull/9933)) [@rongou](https://github.com/rongou) -- Add build-time publish step to cpu build script ([#9927](https://github.com/rapidsai/cudf/pull/9927)) [@davidwendt](https://github.com/davidwendt) -- Refactoring ceil/round/floor code for datetime64 types ([#9926](https://github.com/rapidsai/cudf/pull/9926)) [@mayankanand007](https://github.com/mayankanand007) -- Remove various unused functions ([#9922](https://github.com/rapidsai/cudf/pull/9922)) [@vyasr](https://github.com/vyasr) -- Raise in `query` if dtype is not supported ([#9921](https://github.com/rapidsai/cudf/pull/9921)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add missing imports tests ([#9920](https://github.com/rapidsai/cudf/pull/9920)) [@Ethyling](https://github.com/Ethyling) -- Spark Decimal128 hashing ([#9919](https://github.com/rapidsai/cudf/pull/9919)) [@rwlee](https://github.com/rwlee) -- Replace `thrust/std::get` with structured bindings ([#9915](https://github.com/rapidsai/cudf/pull/9915)) [@codereport](https://github.com/codereport) -- Upgrade thrust version to 1.15 ([#9912](https://github.com/rapidsai/cudf/pull/9912)) [@robertmaynard](https://github.com/robertmaynard) -- Remove conda envs for CUDA 11.0 and 11.2. ([#9910](https://github.com/rapidsai/cudf/pull/9910)) [@bdice](https://github.com/bdice) -- Return count of set bits from inplace_bitmask_and. ([#9904](https://github.com/rapidsai/cudf/pull/9904)) [@bdice](https://github.com/bdice) -- Use dynamic nullate for join hasher and equality comparator ([#9902](https://github.com/rapidsai/cudf/pull/9902)) [@davidwendt](https://github.com/davidwendt) -- Update ucx-py version on release using rvc ([#9897](https://github.com/rapidsai/cudf/pull/9897)) [@Ethyling](https://github.com/Ethyling) -- Remove `IncludeCategories` from `.clang-format` ([#9876](https://github.com/rapidsai/cudf/pull/9876)) [@codereport](https://github.com/codereport) -- Support statically linking CUDA runtime for Java bindings ([#9873](https://github.com/rapidsai/cudf/pull/9873)) [@jlowe](https://github.com/jlowe) -- Add `clang-tidy` to libcudf ([#9860](https://github.com/rapidsai/cudf/pull/9860)) [@codereport](https://github.com/codereport) -- Remove deprecated methods from Java Table class ([#9853](https://github.com/rapidsai/cudf/pull/9853)) [@jlowe](https://github.com/jlowe) -- Add test for map column metadata handling in ORC writer ([#9852](https://github.com/rapidsai/cudf/pull/9852)) [@vuule](https://github.com/vuule) -- Use pandas `to_offset` to parse frequency string in `date_range` ([#9843](https://github.com/rapidsai/cudf/pull/9843)) [@isVoid](https://github.com/isVoid) -- add templated benchmark with fixture ([#9838](https://github.com/rapidsai/cudf/pull/9838)) [@karthikeyann](https://github.com/karthikeyann) -- Use list of column inputs for `apply_boolean_mask` ([#9832](https://github.com/rapidsai/cudf/pull/9832)) [@isVoid](https://github.com/isVoid) -- Added a few more tests for Decimal to String cast ([#9818](https://github.com/rapidsai/cudf/pull/9818)) [@razajafri](https://github.com/razajafri) -- Run doctests. ([#9815](https://github.com/rapidsai/cudf/pull/9815)) [@bdice](https://github.com/bdice) -- Avoid overflow for fixed_point round ([#9809](https://github.com/rapidsai/cudf/pull/9809)) [@sperlingxx](https://github.com/sperlingxx) -- Move `drop_duplicates`, `drop_na`, `_gather`, `take` to IndexFrame and create their `_base_index` counterparts ([#9807](https://github.com/rapidsai/cudf/pull/9807)) [@isVoid](https://github.com/isVoid) -- Use vector factories for host-device copies. ([#9806](https://github.com/rapidsai/cudf/pull/9806)) [@bdice](https://github.com/bdice) -- Refactor host device macros ([#9797](https://github.com/rapidsai/cudf/pull/9797)) [@vyasr](https://github.com/vyasr) -- Remove unused masked udf cython/c++ code ([#9792](https://github.com/rapidsai/cudf/pull/9792)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Allow custom sort functions for dask-cudf `sort_values` ([#9789](https://github.com/rapidsai/cudf/pull/9789)) [@charlesbluca](https://github.com/charlesbluca) -- Improve build time of libcudf iterator tests ([#9788](https://github.com/rapidsai/cudf/pull/9788)) [@davidwendt](https://github.com/davidwendt) -- Copy Java native dependencies directly into classpath ([#9787](https://github.com/rapidsai/cudf/pull/9787)) [@jlowe](https://github.com/jlowe) -- Add decimal types to cuIO benchmarks ([#9776](https://github.com/rapidsai/cudf/pull/9776)) [@vuule](https://github.com/vuule) -- Pick smallest decimal type with required precision in ORC reader ([#9775](https://github.com/rapidsai/cudf/pull/9775)) [@vuule](https://github.com/vuule) -- Avoid overflow for `fixed_point` `cudf::cast` and performance optimization ([#9772](https://github.com/rapidsai/cudf/pull/9772)) [@codereport](https://github.com/codereport) -- Use CTAD with Thrust function objects ([#9768](https://github.com/rapidsai/cudf/pull/9768)) [@codereport](https://github.com/codereport) -- Refactor TableTest assertion methods to a separate utility class ([#9762](https://github.com/rapidsai/cudf/pull/9762)) [@jlowe](https://github.com/jlowe) -- Use Java classloader to find test resources ([#9760](https://github.com/rapidsai/cudf/pull/9760)) [@jlowe](https://github.com/jlowe) -- Allow cast decimal128 to string and add tests ([#9756](https://github.com/rapidsai/cudf/pull/9756)) [@razajafri](https://github.com/razajafri) -- Load balance optimization for contiguous_split ([#9755](https://github.com/rapidsai/cudf/pull/9755)) [@nvdbaranec](https://github.com/nvdbaranec) -- Consolidate and improve `reset_index` ([#9750](https://github.com/rapidsai/cudf/pull/9750)) [@isVoid](https://github.com/isVoid) -- Update to UCX-Py 0.24 ([#9748](https://github.com/rapidsai/cudf/pull/9748)) [@pentschev](https://github.com/pentschev) -- Skip cufile tests in JNI build script ([#9744](https://github.com/rapidsai/cudf/pull/9744)) [@pxLi](https://github.com/pxLi) -- Enable string to decimal 128 cast ([#9742](https://github.com/rapidsai/cudf/pull/9742)) [@razajafri](https://github.com/razajafri) -- Use stop instead of stop_. ([#9735](https://github.com/rapidsai/cudf/pull/9735)) [@bdice](https://github.com/bdice) -- Forward-merge branch-21.12 to branch-22.02 ([#9730](https://github.com/rapidsai/cudf/pull/9730)) [@bdice](https://github.com/bdice) -- Improve cmake format script ([#9723](https://github.com/rapidsai/cudf/pull/9723)) [@vyasr](https://github.com/vyasr) -- Use cuFile direct device reads/writes by default in cuIO ([#9722](https://github.com/rapidsai/cudf/pull/9722)) [@vuule](https://github.com/vuule) -- Add directory-partitioned data support to cudf.read_parquet ([#9720](https://github.com/rapidsai/cudf/pull/9720)) [@rjzamora](https://github.com/rjzamora) -- Use stream allocator adaptor for hash join table ([#9704](https://github.com/rapidsai/cudf/pull/9704)) [@PointKernel](https://github.com/PointKernel) -- Update check for inf/nan strings in libcudf float conversion to ignore case ([#9694](https://github.com/rapidsai/cudf/pull/9694)) [@davidwendt](https://github.com/davidwendt) -- Update cudf JNI to 22.02.0-SNAPSHOT ([#9681](https://github.com/rapidsai/cudf/pull/9681)) [@pxLi](https://github.com/pxLi) -- Replace cudf's concurrent_ordered_map with cuco::static_map in semi/anti joins ([#9666](https://github.com/rapidsai/cudf/pull/9666)) [@vyasr](https://github.com/vyasr) -- Some improvements to `parse_decimal` function and bindings for `is_fixed_point` ([#9658](https://github.com/rapidsai/cudf/pull/9658)) [@razajafri](https://github.com/razajafri) -- Add utility to format ninja-log build times ([#9631](https://github.com/rapidsai/cudf/pull/9631)) [@davidwendt](https://github.com/davidwendt) -- Allow runtime has_nulls parameter for row operators ([#9623](https://github.com/rapidsai/cudf/pull/9623)) [@davidwendt](https://github.com/davidwendt) -- Use fsspec.parquet for improved read_parquet performance from remote storage ([#9589](https://github.com/rapidsai/cudf/pull/9589)) [@rjzamora](https://github.com/rjzamora) -- Refactor bit counting APIs, introduce valid/null count functions, and split host/device side code for segmented counts. ([#9588](https://github.com/rapidsai/cudf/pull/9588)) [@bdice](https://github.com/bdice) -- Use List of Columns as Input for `drop_nulls`, `gather` and `drop_duplicates` ([#9558](https://github.com/rapidsai/cudf/pull/9558)) [@isVoid](https://github.com/isVoid) -- Simplify merge internals and reduce overhead ([#9516](https://github.com/rapidsai/cudf/pull/9516)) [@vyasr](https://github.com/vyasr) -- Add `struct` generation support in datagenerator & fuzz tests ([#9180](https://github.com/rapidsai/cudf/pull/9180)) [@galipremsagar](https://github.com/galipremsagar) -- Simplify write_csv by removing unnecessary writer/impl classes ([#9089](https://github.com/rapidsai/cudf/pull/9089)) [@cwharris](https://github.com/cwharris) +- Prepare upload scripts for Python 3.7 removal ([#10092](https://github.com/NVIDIA/cudf/pull/10092)) [@Ethyling](https://github.com/Ethyling) +- Simplify custreamz and cudf_kafka recipes files ([#10065](https://github.com/NVIDIA/cudf/pull/10065)) [@Ethyling](https://github.com/Ethyling) +- ORC writer API changes for granular statistics ([#10058](https://github.com/NVIDIA/cudf/pull/10058)) [@mythrocks](https://github.com/mythrocks) +- Remove python constraints in cutreamz and cudf_kafka recipes ([#10052](https://github.com/NVIDIA/cudf/pull/10052)) [@Ethyling](https://github.com/Ethyling) +- Unpin `dask` and `distributed` in CI ([#10028](https://github.com/NVIDIA/cudf/pull/10028)) [@galipremsagar](https://github.com/galipremsagar) +- Add `_from_column_like_self` factory ([#10022](https://github.com/NVIDIA/cudf/pull/10022)) [@isVoid](https://github.com/isVoid) +- Replace custom CUDA bindings previously provided by RMM with official CUDA Python bindings ([#10008](https://github.com/NVIDIA/cudf/pull/10008)) [@shwina](https://github.com/shwina) +- Use `cuda::std::is_arithmetic` in `cudf::is_numeric` trait. ([#9996](https://github.com/NVIDIA/cudf/pull/9996)) [@bdice](https://github.com/bdice) +- Clean up CUDA stream use in cuIO ([#9991](https://github.com/NVIDIA/cudf/pull/9991)) [@vuule](https://github.com/vuule) +- Use addressed-ordered first fit for the pinned memory pool ([#9989](https://github.com/NVIDIA/cudf/pull/9989)) [@rongou](https://github.com/rongou) +- Add strings tests to transpose_test.cpp ([#9985](https://github.com/NVIDIA/cudf/pull/9985)) [@davidwendt](https://github.com/davidwendt) +- Use gpuci_mamba_retry on Java CI. ([#9983](https://github.com/NVIDIA/cudf/pull/9983)) [@bdice](https://github.com/bdice) +- Remove deprecated method `one_hot_encoding` ([#9977](https://github.com/NVIDIA/cudf/pull/9977)) [@isVoid](https://github.com/isVoid) +- Minor cleanup of unused Python functions ([#9974](https://github.com/NVIDIA/cudf/pull/9974)) [@vyasr](https://github.com/vyasr) +- Use new efficient partitioned parquet writing in cuDF ([#9971](https://github.com/NVIDIA/cudf/pull/9971)) [@devavret](https://github.com/devavret) +- Remove str.subword_tokenize ([#9968](https://github.com/NVIDIA/cudf/pull/9968)) [@VibhuJawa](https://github.com/VibhuJawa) +- Forward-merge branch-21.12 to branch-22.02 ([#9947](https://github.com/NVIDIA/cudf/pull/9947)) [@bdice](https://github.com/bdice) +- Remove deprecated `method` parameter from `merge` and `join`. ([#9944](https://github.com/NVIDIA/cudf/pull/9944)) [@bdice](https://github.com/bdice) +- Remove deprecated method DataFrame.hash_columns. ([#9943](https://github.com/NVIDIA/cudf/pull/9943)) [@bdice](https://github.com/bdice) +- Remove deprecated method Series.hash_encode. ([#9942](https://github.com/NVIDIA/cudf/pull/9942)) [@bdice](https://github.com/bdice) +- use ninja in java ci build ([#9933](https://github.com/NVIDIA/cudf/pull/9933)) [@rongou](https://github.com/rongou) +- Add build-time publish step to cpu build script ([#9927](https://github.com/NVIDIA/cudf/pull/9927)) [@davidwendt](https://github.com/davidwendt) +- Refactoring ceil/round/floor code for datetime64 types ([#9926](https://github.com/NVIDIA/cudf/pull/9926)) [@mayankanand007](https://github.com/mayankanand007) +- Remove various unused functions ([#9922](https://github.com/NVIDIA/cudf/pull/9922)) [@vyasr](https://github.com/vyasr) +- Raise in `query` if dtype is not supported ([#9921](https://github.com/NVIDIA/cudf/pull/9921)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add missing imports tests ([#9920](https://github.com/NVIDIA/cudf/pull/9920)) [@Ethyling](https://github.com/Ethyling) +- Spark Decimal128 hashing ([#9919](https://github.com/NVIDIA/cudf/pull/9919)) [@rwlee](https://github.com/rwlee) +- Replace `thrust/std::get` with structured bindings ([#9915](https://github.com/NVIDIA/cudf/pull/9915)) [@codereport](https://github.com/codereport) +- Upgrade thrust version to 1.15 ([#9912](https://github.com/NVIDIA/cudf/pull/9912)) [@robertmaynard](https://github.com/robertmaynard) +- Remove conda envs for CUDA 11.0 and 11.2. ([#9910](https://github.com/NVIDIA/cudf/pull/9910)) [@bdice](https://github.com/bdice) +- Return count of set bits from inplace_bitmask_and. ([#9904](https://github.com/NVIDIA/cudf/pull/9904)) [@bdice](https://github.com/bdice) +- Use dynamic nullate for join hasher and equality comparator ([#9902](https://github.com/NVIDIA/cudf/pull/9902)) [@davidwendt](https://github.com/davidwendt) +- Update ucx-py version on release using rvc ([#9897](https://github.com/NVIDIA/cudf/pull/9897)) [@Ethyling](https://github.com/Ethyling) +- Remove `IncludeCategories` from `.clang-format` ([#9876](https://github.com/NVIDIA/cudf/pull/9876)) [@codereport](https://github.com/codereport) +- Support statically linking CUDA runtime for Java bindings ([#9873](https://github.com/NVIDIA/cudf/pull/9873)) [@jlowe](https://github.com/jlowe) +- Add `clang-tidy` to libcudf ([#9860](https://github.com/NVIDIA/cudf/pull/9860)) [@codereport](https://github.com/codereport) +- Remove deprecated methods from Java Table class ([#9853](https://github.com/NVIDIA/cudf/pull/9853)) [@jlowe](https://github.com/jlowe) +- Add test for map column metadata handling in ORC writer ([#9852](https://github.com/NVIDIA/cudf/pull/9852)) [@vuule](https://github.com/vuule) +- Use pandas `to_offset` to parse frequency string in `date_range` ([#9843](https://github.com/NVIDIA/cudf/pull/9843)) [@isVoid](https://github.com/isVoid) +- add templated benchmark with fixture ([#9838](https://github.com/NVIDIA/cudf/pull/9838)) [@karthikeyann](https://github.com/karthikeyann) +- Use list of column inputs for `apply_boolean_mask` ([#9832](https://github.com/NVIDIA/cudf/pull/9832)) [@isVoid](https://github.com/isVoid) +- Added a few more tests for Decimal to String cast ([#9818](https://github.com/NVIDIA/cudf/pull/9818)) [@razajafri](https://github.com/razajafri) +- Run doctests. ([#9815](https://github.com/NVIDIA/cudf/pull/9815)) [@bdice](https://github.com/bdice) +- Avoid overflow for fixed_point round ([#9809](https://github.com/NVIDIA/cudf/pull/9809)) [@sperlingxx](https://github.com/sperlingxx) +- Move `drop_duplicates`, `drop_na`, `_gather`, `take` to IndexFrame and create their `_base_index` counterparts ([#9807](https://github.com/NVIDIA/cudf/pull/9807)) [@isVoid](https://github.com/isVoid) +- Use vector factories for host-device copies. ([#9806](https://github.com/NVIDIA/cudf/pull/9806)) [@bdice](https://github.com/bdice) +- Refactor host device macros ([#9797](https://github.com/NVIDIA/cudf/pull/9797)) [@vyasr](https://github.com/vyasr) +- Remove unused masked udf cython/c++ code ([#9792](https://github.com/NVIDIA/cudf/pull/9792)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Allow custom sort functions for dask-cudf `sort_values` ([#9789](https://github.com/NVIDIA/cudf/pull/9789)) [@charlesbluca](https://github.com/charlesbluca) +- Improve build time of libcudf iterator tests ([#9788](https://github.com/NVIDIA/cudf/pull/9788)) [@davidwendt](https://github.com/davidwendt) +- Copy Java native dependencies directly into classpath ([#9787](https://github.com/NVIDIA/cudf/pull/9787)) [@jlowe](https://github.com/jlowe) +- Add decimal types to cuIO benchmarks ([#9776](https://github.com/NVIDIA/cudf/pull/9776)) [@vuule](https://github.com/vuule) +- Pick smallest decimal type with required precision in ORC reader ([#9775](https://github.com/NVIDIA/cudf/pull/9775)) [@vuule](https://github.com/vuule) +- Avoid overflow for `fixed_point` `cudf::cast` and performance optimization ([#9772](https://github.com/NVIDIA/cudf/pull/9772)) [@codereport](https://github.com/codereport) +- Use CTAD with Thrust function objects ([#9768](https://github.com/NVIDIA/cudf/pull/9768)) [@codereport](https://github.com/codereport) +- Refactor TableTest assertion methods to a separate utility class ([#9762](https://github.com/NVIDIA/cudf/pull/9762)) [@jlowe](https://github.com/jlowe) +- Use Java classloader to find test resources ([#9760](https://github.com/NVIDIA/cudf/pull/9760)) [@jlowe](https://github.com/jlowe) +- Allow cast decimal128 to string and add tests ([#9756](https://github.com/NVIDIA/cudf/pull/9756)) [@razajafri](https://github.com/razajafri) +- Load balance optimization for contiguous_split ([#9755](https://github.com/NVIDIA/cudf/pull/9755)) [@nvdbaranec](https://github.com/nvdbaranec) +- Consolidate and improve `reset_index` ([#9750](https://github.com/NVIDIA/cudf/pull/9750)) [@isVoid](https://github.com/isVoid) +- Update to UCX-Py 0.24 ([#9748](https://github.com/NVIDIA/cudf/pull/9748)) [@pentschev](https://github.com/pentschev) +- Skip cufile tests in JNI build script ([#9744](https://github.com/NVIDIA/cudf/pull/9744)) [@pxLi](https://github.com/pxLi) +- Enable string to decimal 128 cast ([#9742](https://github.com/NVIDIA/cudf/pull/9742)) [@razajafri](https://github.com/razajafri) +- Use stop instead of stop_. ([#9735](https://github.com/NVIDIA/cudf/pull/9735)) [@bdice](https://github.com/bdice) +- Forward-merge branch-21.12 to branch-22.02 ([#9730](https://github.com/NVIDIA/cudf/pull/9730)) [@bdice](https://github.com/bdice) +- Improve cmake format script ([#9723](https://github.com/NVIDIA/cudf/pull/9723)) [@vyasr](https://github.com/vyasr) +- Use cuFile direct device reads/writes by default in cuIO ([#9722](https://github.com/NVIDIA/cudf/pull/9722)) [@vuule](https://github.com/vuule) +- Add directory-partitioned data support to cudf.read_parquet ([#9720](https://github.com/NVIDIA/cudf/pull/9720)) [@rjzamora](https://github.com/rjzamora) +- Use stream allocator adaptor for hash join table ([#9704](https://github.com/NVIDIA/cudf/pull/9704)) [@PointKernel](https://github.com/PointKernel) +- Update check for inf/nan strings in libcudf float conversion to ignore case ([#9694](https://github.com/NVIDIA/cudf/pull/9694)) [@davidwendt](https://github.com/davidwendt) +- Update cudf JNI to 22.02.0-SNAPSHOT ([#9681](https://github.com/NVIDIA/cudf/pull/9681)) [@pxLi](https://github.com/pxLi) +- Replace cudf's concurrent_ordered_map with cuco::static_map in semi/anti joins ([#9666](https://github.com/NVIDIA/cudf/pull/9666)) [@vyasr](https://github.com/vyasr) +- Some improvements to `parse_decimal` function and bindings for `is_fixed_point` ([#9658](https://github.com/NVIDIA/cudf/pull/9658)) [@razajafri](https://github.com/razajafri) +- Add utility to format ninja-log build times ([#9631](https://github.com/NVIDIA/cudf/pull/9631)) [@davidwendt](https://github.com/davidwendt) +- Allow runtime has_nulls parameter for row operators ([#9623](https://github.com/NVIDIA/cudf/pull/9623)) [@davidwendt](https://github.com/davidwendt) +- Use fsspec.parquet for improved read_parquet performance from remote storage ([#9589](https://github.com/NVIDIA/cudf/pull/9589)) [@rjzamora](https://github.com/rjzamora) +- Refactor bit counting APIs, introduce valid/null count functions, and split host/device side code for segmented counts. ([#9588](https://github.com/NVIDIA/cudf/pull/9588)) [@bdice](https://github.com/bdice) +- Use List of Columns as Input for `drop_nulls`, `gather` and `drop_duplicates` ([#9558](https://github.com/NVIDIA/cudf/pull/9558)) [@isVoid](https://github.com/isVoid) +- Simplify merge internals and reduce overhead ([#9516](https://github.com/NVIDIA/cudf/pull/9516)) [@vyasr](https://github.com/vyasr) +- Add `struct` generation support in datagenerator & fuzz tests ([#9180](https://github.com/NVIDIA/cudf/pull/9180)) [@galipremsagar](https://github.com/galipremsagar) +- Simplify write_csv by removing unnecessary writer/impl classes ([#9089](https://github.com/NVIDIA/cudf/pull/9089)) [@cwharris](https://github.com/cwharris) # cuDF 21.12.00 (9 Dec 2021) ## 🚨 Breaking Changes -- Update `bitmask_and` and `bitmask_or` to return a pair of resulting mask and count of unset bits ([#9616](https://github.com/rapidsai/cudf/pull/9616)) [@PointKernel](https://github.com/PointKernel) -- Remove sizeof and standardize on memory_usage ([#9544](https://github.com/rapidsai/cudf/pull/9544)) [@vyasr](https://github.com/vyasr) -- Add support for single-line regex anchors ^/$ in contains_re ([#9482](https://github.com/rapidsai/cudf/pull/9482)) [@davidwendt](https://github.com/davidwendt) -- Refactor sorting APIs ([#9464](https://github.com/rapidsai/cudf/pull/9464)) [@vyasr](https://github.com/vyasr) -- Update Java nvcomp JNI bindings to nvcomp 2.x API ([#9384](https://github.com/rapidsai/cudf/pull/9384)) [@jbrennan333](https://github.com/jbrennan333) -- Support Python UDFs written in terms of rows ([#9343](https://github.com/rapidsai/cudf/pull/9343)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- JNI: Support nested types in ORC writer ([#9334](https://github.com/rapidsai/cudf/pull/9334)) [@firestarman](https://github.com/firestarman) -- Optionally nullify out-of-bounds indices in segmented_gather(). ([#9318](https://github.com/rapidsai/cudf/pull/9318)) [@mythrocks](https://github.com/mythrocks) -- Refactor cuIO timestamp processing with `cuda::std::chrono` ([#9278](https://github.com/rapidsai/cudf/pull/9278)) [@PointKernel](https://github.com/PointKernel) -- Various internal MultiIndex improvements ([#9243](https://github.com/rapidsai/cudf/pull/9243)) [@vyasr](https://github.com/vyasr) +- Update `bitmask_and` and `bitmask_or` to return a pair of resulting mask and count of unset bits ([#9616](https://github.com/NVIDIA/cudf/pull/9616)) [@PointKernel](https://github.com/PointKernel) +- Remove sizeof and standardize on memory_usage ([#9544](https://github.com/NVIDIA/cudf/pull/9544)) [@vyasr](https://github.com/vyasr) +- Add support for single-line regex anchors ^/$ in contains_re ([#9482](https://github.com/NVIDIA/cudf/pull/9482)) [@davidwendt](https://github.com/davidwendt) +- Refactor sorting APIs ([#9464](https://github.com/NVIDIA/cudf/pull/9464)) [@vyasr](https://github.com/vyasr) +- Update Java nvcomp JNI bindings to nvcomp 2.x API ([#9384](https://github.com/NVIDIA/cudf/pull/9384)) [@jbrennan333](https://github.com/jbrennan333) +- Support Python UDFs written in terms of rows ([#9343](https://github.com/NVIDIA/cudf/pull/9343)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- JNI: Support nested types in ORC writer ([#9334](https://github.com/NVIDIA/cudf/pull/9334)) [@firestarman](https://github.com/firestarman) +- Optionally nullify out-of-bounds indices in segmented_gather(). ([#9318](https://github.com/NVIDIA/cudf/pull/9318)) [@mythrocks](https://github.com/mythrocks) +- Refactor cuIO timestamp processing with `cuda::std::chrono` ([#9278](https://github.com/NVIDIA/cudf/pull/9278)) [@PointKernel](https://github.com/PointKernel) +- Various internal MultiIndex improvements ([#9243](https://github.com/NVIDIA/cudf/pull/9243)) [@vyasr](https://github.com/vyasr) ## 🐛 Bug Fixes -- Fix read_parquet bug for bytes input ([#9669](https://github.com/rapidsai/cudf/pull/9669)) [@rjzamora](https://github.com/rjzamora) -- Use `_gather` internal for `sort_*` ([#9668](https://github.com/rapidsai/cudf/pull/9668)) [@isVoid](https://github.com/isVoid) -- Fix behavior of equals for non-DataFrame Frames and add tests. ([#9653](https://github.com/rapidsai/cudf/pull/9653)) [@vyasr](https://github.com/vyasr) -- Dont recompute output size if it is already available ([#9649](https://github.com/rapidsai/cudf/pull/9649)) [@abellina](https://github.com/abellina) -- Fix read_parquet bug for extended dtypes from remote storage ([#9638](https://github.com/rapidsai/cudf/pull/9638)) [@rjzamora](https://github.com/rjzamora) -- add const when getting data from a JNI data wrapper ([#9637](https://github.com/rapidsai/cudf/pull/9637)) [@wjxiz1992](https://github.com/wjxiz1992) -- Fix debrotli issue on CUDA 11.5 ([#9632](https://github.com/rapidsai/cudf/pull/9632)) [@vuule](https://github.com/vuule) -- Use std::size_t when computing join output size ([#9626](https://github.com/rapidsai/cudf/pull/9626)) [@jlowe](https://github.com/jlowe) -- Fix `usecols` parameter handling in `dask_cudf.read_csv` ([#9618](https://github.com/rapidsai/cudf/pull/9618)) [@galipremsagar](https://github.com/galipremsagar) -- Add support for string `'nan', 'inf' & '-inf'` values while type-casting to `float` ([#9613](https://github.com/rapidsai/cudf/pull/9613)) [@galipremsagar](https://github.com/galipremsagar) -- Avoid passing NativeFileDatasource to pyarrow in read_parquet ([#9608](https://github.com/rapidsai/cudf/pull/9608)) [@rjzamora](https://github.com/rjzamora) -- Fix test failure with cuda 11.5 in row_bit_count tests. ([#9581](https://github.com/rapidsai/cudf/pull/9581)) [@nvdbaranec](https://github.com/nvdbaranec) -- Correct _LIBCUDACXX_CUDACC_VER value computation ([#9579](https://github.com/rapidsai/cudf/pull/9579)) [@robertmaynard](https://github.com/robertmaynard) -- Increase max RLE stream size estimate to avoid potential overflows ([#9568](https://github.com/rapidsai/cudf/pull/9568)) [@vuule](https://github.com/vuule) -- Fix edge case in tdigest scalar generation for groups containing all nulls. ([#9551](https://github.com/rapidsai/cudf/pull/9551)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix pytests failing in `cuda-11.5` environment ([#9547](https://github.com/rapidsai/cudf/pull/9547)) [@galipremsagar](https://github.com/galipremsagar) -- compile libnvcomp with PTDS if requested ([#9540](https://github.com/rapidsai/cudf/pull/9540)) [@jbrennan333](https://github.com/jbrennan333) -- Fix `segmented_gather()` for null LIST rows ([#9537](https://github.com/rapidsai/cudf/pull/9537)) [@mythrocks](https://github.com/mythrocks) -- Deprecate DataFrame.label_encoding, use private _label_encoding method internally. ([#9535](https://github.com/rapidsai/cudf/pull/9535)) [@bdice](https://github.com/bdice) -- Fix several test and benchmark issues related to bitmask allocations. ([#9521](https://github.com/rapidsai/cudf/pull/9521)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix for inserting duplicates in groupby result cache ([#9508](https://github.com/rapidsai/cudf/pull/9508)) [@karthikeyann](https://github.com/karthikeyann) -- Fix mismatched types error in clip() when using non int64 numeric types ([#9498](https://github.com/rapidsai/cudf/pull/9498)) [@davidwendt](https://github.com/davidwendt) -- Match conda pinnings for style checks (revert part of #9412, #9433). ([#9490](https://github.com/rapidsai/cudf/pull/9490)) [@bdice](https://github.com/bdice) -- Make sure all dask-cudf supported aggs are handled in `_tree_node_agg` ([#9487](https://github.com/rapidsai/cudf/pull/9487)) [@charlesbluca](https://github.com/charlesbluca) -- Resolve `hash_columns` `FutureWarning` in `dask_cudf` ([#9481](https://github.com/rapidsai/cudf/pull/9481)) [@pentschev](https://github.com/pentschev) -- Add fixed point to AllTypes in libcudf unit tests ([#9472](https://github.com/rapidsai/cudf/pull/9472)) [@karthikeyann](https://github.com/karthikeyann) -- Fix regex handling of embedded null characters ([#9470](https://github.com/rapidsai/cudf/pull/9470)) [@davidwendt](https://github.com/davidwendt) -- Fix memcheck error in copy-if-else ([#9467](https://github.com/rapidsai/cudf/pull/9467)) [@davidwendt](https://github.com/davidwendt) -- Fix bug in dask_cudf.read_parquet for index=False ([#9453](https://github.com/rapidsai/cudf/pull/9453)) [@rjzamora](https://github.com/rjzamora) -- Preserve the decimal scale when creating a default scalar ([#9449](https://github.com/rapidsai/cudf/pull/9449)) [@revans2](https://github.com/revans2) -- Push down parent nulls when flattening nested columns. ([#9443](https://github.com/rapidsai/cudf/pull/9443)) [@mythrocks](https://github.com/mythrocks) -- Fix memcheck error in gtest SegmentedGatherTest/GatherSliced ([#9442](https://github.com/rapidsai/cudf/pull/9442)) [@davidwendt](https://github.com/davidwendt) -- Revert "Fix quantile division / partition handling for dask-cudf sort… ([#9438](https://github.com/rapidsai/cudf/pull/9438)) [@charlesbluca](https://github.com/charlesbluca) -- Allow int-like objects for the `decimals` argument in `round` ([#9428](https://github.com/rapidsai/cudf/pull/9428)) [@shwina](https://github.com/shwina) -- Fix stream compaction's `drop_duplicates` API to use stable sort ([#9417](https://github.com/rapidsai/cudf/pull/9417)) [@ttnghia](https://github.com/ttnghia) -- Skip Comparing Uniform Window Results in Var/std Tests ([#9416](https://github.com/rapidsai/cudf/pull/9416)) [@isVoid](https://github.com/isVoid) -- Fix `StructColumn.to_pandas` type handling issues ([#9388](https://github.com/rapidsai/cudf/pull/9388)) [@galipremsagar](https://github.com/galipremsagar) -- Correct issues in the build dir cudf-config.cmake ([#9386](https://github.com/rapidsai/cudf/pull/9386)) [@robertmaynard](https://github.com/robertmaynard) -- Fix Java table partition test to account for non-deterministic ordering ([#9385](https://github.com/rapidsai/cudf/pull/9385)) [@jlowe](https://github.com/jlowe) -- Fix timestamp truncation/overflow bugs in orc/parquet ([#9382](https://github.com/rapidsai/cudf/pull/9382)) [@PointKernel](https://github.com/PointKernel) -- Fix the crash in stats code ([#9368](https://github.com/rapidsai/cudf/pull/9368)) [@devavret](https://github.com/devavret) -- Make Series.hash_encode results reproducible. ([#9366](https://github.com/rapidsai/cudf/pull/9366)) [@bdice](https://github.com/bdice) -- Fix libcudf compile warnings on debug 11.4 build ([#9360](https://github.com/rapidsai/cudf/pull/9360)) [@davidwendt](https://github.com/davidwendt) -- Fail gracefully when compiling python UDFs that attempt to access columns with unsupported dtypes ([#9359](https://github.com/rapidsai/cudf/pull/9359)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Set pass_filenames: false in mypy pre-commit configuration. ([#9349](https://github.com/rapidsai/cudf/pull/9349)) [@bdice](https://github.com/bdice) -- Fix cudf_assert in cudf::io::orc::gpu::gpuDecodeOrcColumnData ([#9348](https://github.com/rapidsai/cudf/pull/9348)) [@davidwendt](https://github.com/davidwendt) -- Fix memcheck error in groupby-tdigest get_scalar_minmax ([#9339](https://github.com/rapidsai/cudf/pull/9339)) [@davidwendt](https://github.com/davidwendt) -- Optimizations for `cudf.concat` when `axis=1` ([#9333](https://github.com/rapidsai/cudf/pull/9333)) [@galipremsagar](https://github.com/galipremsagar) -- Use f-string in join helper warning message. ([#9325](https://github.com/rapidsai/cudf/pull/9325)) [@bdice](https://github.com/bdice) -- Avoid casting to list or struct dtypes in dask_cudf.read_parquet ([#9314](https://github.com/rapidsai/cudf/pull/9314)) [@rjzamora](https://github.com/rjzamora) -- Fix null count in statistics for parquet ([#9303](https://github.com/rapidsai/cudf/pull/9303)) [@devavret](https://github.com/devavret) -- Potential overflow of `decimal32` when casting to `int64_t` ([#9287](https://github.com/rapidsai/cudf/pull/9287)) [@codereport](https://github.com/codereport) -- Fix quantile division / partition handling for dask-cudf sort on null dataframes ([#9259](https://github.com/rapidsai/cudf/pull/9259)) [@charlesbluca](https://github.com/charlesbluca) -- Updating cudf version also updates rapids cmake branch ([#9249](https://github.com/rapidsai/cudf/pull/9249)) [@robertmaynard](https://github.com/robertmaynard) -- Implement `one_hot_encoding` in libcudf and bind to python ([#9229](https://github.com/rapidsai/cudf/pull/9229)) [@isVoid](https://github.com/isVoid) -- BUG FIX: CSV Writer ignores the header parameter when no metadata is provided ([#8740](https://github.com/rapidsai/cudf/pull/8740)) [@skirui-source](https://github.com/skirui-source) +- Fix read_parquet bug for bytes input ([#9669](https://github.com/NVIDIA/cudf/pull/9669)) [@rjzamora](https://github.com/rjzamora) +- Use `_gather` internal for `sort_*` ([#9668](https://github.com/NVIDIA/cudf/pull/9668)) [@isVoid](https://github.com/isVoid) +- Fix behavior of equals for non-DataFrame Frames and add tests. ([#9653](https://github.com/NVIDIA/cudf/pull/9653)) [@vyasr](https://github.com/vyasr) +- Dont recompute output size if it is already available ([#9649](https://github.com/NVIDIA/cudf/pull/9649)) [@abellina](https://github.com/abellina) +- Fix read_parquet bug for extended dtypes from remote storage ([#9638](https://github.com/NVIDIA/cudf/pull/9638)) [@rjzamora](https://github.com/rjzamora) +- add const when getting data from a JNI data wrapper ([#9637](https://github.com/NVIDIA/cudf/pull/9637)) [@wjxiz1992](https://github.com/wjxiz1992) +- Fix debrotli issue on CUDA 11.5 ([#9632](https://github.com/NVIDIA/cudf/pull/9632)) [@vuule](https://github.com/vuule) +- Use std::size_t when computing join output size ([#9626](https://github.com/NVIDIA/cudf/pull/9626)) [@jlowe](https://github.com/jlowe) +- Fix `usecols` parameter handling in `dask_cudf.read_csv` ([#9618](https://github.com/NVIDIA/cudf/pull/9618)) [@galipremsagar](https://github.com/galipremsagar) +- Add support for string `'nan', 'inf' & '-inf'` values while type-casting to `float` ([#9613](https://github.com/NVIDIA/cudf/pull/9613)) [@galipremsagar](https://github.com/galipremsagar) +- Avoid passing NativeFileDatasource to pyarrow in read_parquet ([#9608](https://github.com/NVIDIA/cudf/pull/9608)) [@rjzamora](https://github.com/rjzamora) +- Fix test failure with cuda 11.5 in row_bit_count tests. ([#9581](https://github.com/NVIDIA/cudf/pull/9581)) [@nvdbaranec](https://github.com/nvdbaranec) +- Correct _LIBCUDACXX_CUDACC_VER value computation ([#9579](https://github.com/NVIDIA/cudf/pull/9579)) [@robertmaynard](https://github.com/robertmaynard) +- Increase max RLE stream size estimate to avoid potential overflows ([#9568](https://github.com/NVIDIA/cudf/pull/9568)) [@vuule](https://github.com/vuule) +- Fix edge case in tdigest scalar generation for groups containing all nulls. ([#9551](https://github.com/NVIDIA/cudf/pull/9551)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix pytests failing in `cuda-11.5` environment ([#9547](https://github.com/NVIDIA/cudf/pull/9547)) [@galipremsagar](https://github.com/galipremsagar) +- compile libnvcomp with PTDS if requested ([#9540](https://github.com/NVIDIA/cudf/pull/9540)) [@jbrennan333](https://github.com/jbrennan333) +- Fix `segmented_gather()` for null LIST rows ([#9537](https://github.com/NVIDIA/cudf/pull/9537)) [@mythrocks](https://github.com/mythrocks) +- Deprecate DataFrame.label_encoding, use private _label_encoding method internally. ([#9535](https://github.com/NVIDIA/cudf/pull/9535)) [@bdice](https://github.com/bdice) +- Fix several test and benchmark issues related to bitmask allocations. ([#9521](https://github.com/NVIDIA/cudf/pull/9521)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix for inserting duplicates in groupby result cache ([#9508](https://github.com/NVIDIA/cudf/pull/9508)) [@karthikeyann](https://github.com/karthikeyann) +- Fix mismatched types error in clip() when using non int64 numeric types ([#9498](https://github.com/NVIDIA/cudf/pull/9498)) [@davidwendt](https://github.com/davidwendt) +- Match conda pinnings for style checks (revert part of #9412, #9433). ([#9490](https://github.com/NVIDIA/cudf/pull/9490)) [@bdice](https://github.com/bdice) +- Make sure all dask-cudf supported aggs are handled in `_tree_node_agg` ([#9487](https://github.com/NVIDIA/cudf/pull/9487)) [@charlesbluca](https://github.com/charlesbluca) +- Resolve `hash_columns` `FutureWarning` in `dask_cudf` ([#9481](https://github.com/NVIDIA/cudf/pull/9481)) [@pentschev](https://github.com/pentschev) +- Add fixed point to AllTypes in libcudf unit tests ([#9472](https://github.com/NVIDIA/cudf/pull/9472)) [@karthikeyann](https://github.com/karthikeyann) +- Fix regex handling of embedded null characters ([#9470](https://github.com/NVIDIA/cudf/pull/9470)) [@davidwendt](https://github.com/davidwendt) +- Fix memcheck error in copy-if-else ([#9467](https://github.com/NVIDIA/cudf/pull/9467)) [@davidwendt](https://github.com/davidwendt) +- Fix bug in dask_cudf.read_parquet for index=False ([#9453](https://github.com/NVIDIA/cudf/pull/9453)) [@rjzamora](https://github.com/rjzamora) +- Preserve the decimal scale when creating a default scalar ([#9449](https://github.com/NVIDIA/cudf/pull/9449)) [@revans2](https://github.com/revans2) +- Push down parent nulls when flattening nested columns. ([#9443](https://github.com/NVIDIA/cudf/pull/9443)) [@mythrocks](https://github.com/mythrocks) +- Fix memcheck error in gtest SegmentedGatherTest/GatherSliced ([#9442](https://github.com/NVIDIA/cudf/pull/9442)) [@davidwendt](https://github.com/davidwendt) +- Revert "Fix quantile division / partition handling for dask-cudf sort… ([#9438](https://github.com/NVIDIA/cudf/pull/9438)) [@charlesbluca](https://github.com/charlesbluca) +- Allow int-like objects for the `decimals` argument in `round` ([#9428](https://github.com/NVIDIA/cudf/pull/9428)) [@shwina](https://github.com/shwina) +- Fix stream compaction's `drop_duplicates` API to use stable sort ([#9417](https://github.com/NVIDIA/cudf/pull/9417)) [@ttnghia](https://github.com/ttnghia) +- Skip Comparing Uniform Window Results in Var/std Tests ([#9416](https://github.com/NVIDIA/cudf/pull/9416)) [@isVoid](https://github.com/isVoid) +- Fix `StructColumn.to_pandas` type handling issues ([#9388](https://github.com/NVIDIA/cudf/pull/9388)) [@galipremsagar](https://github.com/galipremsagar) +- Correct issues in the build dir cudf-config.cmake ([#9386](https://github.com/NVIDIA/cudf/pull/9386)) [@robertmaynard](https://github.com/robertmaynard) +- Fix Java table partition test to account for non-deterministic ordering ([#9385](https://github.com/NVIDIA/cudf/pull/9385)) [@jlowe](https://github.com/jlowe) +- Fix timestamp truncation/overflow bugs in orc/parquet ([#9382](https://github.com/NVIDIA/cudf/pull/9382)) [@PointKernel](https://github.com/PointKernel) +- Fix the crash in stats code ([#9368](https://github.com/NVIDIA/cudf/pull/9368)) [@devavret](https://github.com/devavret) +- Make Series.hash_encode results reproducible. ([#9366](https://github.com/NVIDIA/cudf/pull/9366)) [@bdice](https://github.com/bdice) +- Fix libcudf compile warnings on debug 11.4 build ([#9360](https://github.com/NVIDIA/cudf/pull/9360)) [@davidwendt](https://github.com/davidwendt) +- Fail gracefully when compiling python UDFs that attempt to access columns with unsupported dtypes ([#9359](https://github.com/NVIDIA/cudf/pull/9359)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Set pass_filenames: false in mypy pre-commit configuration. ([#9349](https://github.com/NVIDIA/cudf/pull/9349)) [@bdice](https://github.com/bdice) +- Fix cudf_assert in cudf::io::orc::gpu::gpuDecodeOrcColumnData ([#9348](https://github.com/NVIDIA/cudf/pull/9348)) [@davidwendt](https://github.com/davidwendt) +- Fix memcheck error in groupby-tdigest get_scalar_minmax ([#9339](https://github.com/NVIDIA/cudf/pull/9339)) [@davidwendt](https://github.com/davidwendt) +- Optimizations for `cudf.concat` when `axis=1` ([#9333](https://github.com/NVIDIA/cudf/pull/9333)) [@galipremsagar](https://github.com/galipremsagar) +- Use f-string in join helper warning message. ([#9325](https://github.com/NVIDIA/cudf/pull/9325)) [@bdice](https://github.com/bdice) +- Avoid casting to list or struct dtypes in dask_cudf.read_parquet ([#9314](https://github.com/NVIDIA/cudf/pull/9314)) [@rjzamora](https://github.com/rjzamora) +- Fix null count in statistics for parquet ([#9303](https://github.com/NVIDIA/cudf/pull/9303)) [@devavret](https://github.com/devavret) +- Potential overflow of `decimal32` when casting to `int64_t` ([#9287](https://github.com/NVIDIA/cudf/pull/9287)) [@codereport](https://github.com/codereport) +- Fix quantile division / partition handling for dask-cudf sort on null dataframes ([#9259](https://github.com/NVIDIA/cudf/pull/9259)) [@charlesbluca](https://github.com/charlesbluca) +- Updating cudf version also updates rapids cmake branch ([#9249](https://github.com/NVIDIA/cudf/pull/9249)) [@robertmaynard](https://github.com/robertmaynard) +- Implement `one_hot_encoding` in libcudf and bind to python ([#9229](https://github.com/NVIDIA/cudf/pull/9229)) [@isVoid](https://github.com/isVoid) +- BUG FIX: CSV Writer ignores the header parameter when no metadata is provided ([#8740](https://github.com/NVIDIA/cudf/pull/8740)) [@skirui-source](https://github.com/skirui-source) ## 📖 Documentation -- Update Documentation to use `TYPED_TEST_SUITE` ([#9654](https://github.com/rapidsai/cudf/pull/9654)) [@codereport](https://github.com/codereport) -- Add dedicated page for `StringHandling` in python docs ([#9624](https://github.com/rapidsai/cudf/pull/9624)) [@galipremsagar](https://github.com/galipremsagar) -- Update docstring of `DataFrame.merge` ([#9572](https://github.com/rapidsai/cudf/pull/9572)) [@galipremsagar](https://github.com/galipremsagar) -- Use raw strings to avoid SyntaxErrors in parsed docstrings. ([#9526](https://github.com/rapidsai/cudf/pull/9526)) [@bdice](https://github.com/bdice) -- Add example to docstrings in `rolling.apply` ([#9522](https://github.com/rapidsai/cudf/pull/9522)) [@isVoid](https://github.com/isVoid) -- Update help message to escape quotes in ./build.sh --cmake-args. ([#9494](https://github.com/rapidsai/cudf/pull/9494)) [@bdice](https://github.com/bdice) -- Improve Python docstring formatting. ([#9493](https://github.com/rapidsai/cudf/pull/9493)) [@bdice](https://github.com/bdice) -- Update table of I/O supported types ([#9476](https://github.com/rapidsai/cudf/pull/9476)) [@vuule](https://github.com/vuule) -- Document invalid regex patterns as undefined behavior ([#9473](https://github.com/rapidsai/cudf/pull/9473)) [@davidwendt](https://github.com/davidwendt) -- Miscellaneous documentation fixes to `cudf` ([#9471](https://github.com/rapidsai/cudf/pull/9471)) [@galipremsagar](https://github.com/galipremsagar) -- Fix many documentation errors in libcudf. ([#9355](https://github.com/rapidsai/cudf/pull/9355)) [@karthikeyann](https://github.com/karthikeyann) -- Fixing SubwordTokenizer docs issue ([#9354](https://github.com/rapidsai/cudf/pull/9354)) [@mayankanand007](https://github.com/mayankanand007) -- Improved deprecation warnings. ([#9347](https://github.com/rapidsai/cudf/pull/9347)) [@bdice](https://github.com/bdice) -- doc reorder mr, stream to stream, mr ([#9308](https://github.com/rapidsai/cudf/pull/9308)) [@karthikeyann](https://github.com/karthikeyann) -- Deprecate method parameters to DataFrame.join, DataFrame.merge. ([#9291](https://github.com/rapidsai/cudf/pull/9291)) [@bdice](https://github.com/bdice) -- Added deprecation warning for `.label_encoding()` ([#9289](https://github.com/rapidsai/cudf/pull/9289)) [@mayankanand007](https://github.com/mayankanand007) +- Update Documentation to use `TYPED_TEST_SUITE` ([#9654](https://github.com/NVIDIA/cudf/pull/9654)) [@codereport](https://github.com/codereport) +- Add dedicated page for `StringHandling` in python docs ([#9624](https://github.com/NVIDIA/cudf/pull/9624)) [@galipremsagar](https://github.com/galipremsagar) +- Update docstring of `DataFrame.merge` ([#9572](https://github.com/NVIDIA/cudf/pull/9572)) [@galipremsagar](https://github.com/galipremsagar) +- Use raw strings to avoid SyntaxErrors in parsed docstrings. ([#9526](https://github.com/NVIDIA/cudf/pull/9526)) [@bdice](https://github.com/bdice) +- Add example to docstrings in `rolling.apply` ([#9522](https://github.com/NVIDIA/cudf/pull/9522)) [@isVoid](https://github.com/isVoid) +- Update help message to escape quotes in ./build.sh --cmake-args. ([#9494](https://github.com/NVIDIA/cudf/pull/9494)) [@bdice](https://github.com/bdice) +- Improve Python docstring formatting. ([#9493](https://github.com/NVIDIA/cudf/pull/9493)) [@bdice](https://github.com/bdice) +- Update table of I/O supported types ([#9476](https://github.com/NVIDIA/cudf/pull/9476)) [@vuule](https://github.com/vuule) +- Document invalid regex patterns as undefined behavior ([#9473](https://github.com/NVIDIA/cudf/pull/9473)) [@davidwendt](https://github.com/davidwendt) +- Miscellaneous documentation fixes to `cudf` ([#9471](https://github.com/NVIDIA/cudf/pull/9471)) [@galipremsagar](https://github.com/galipremsagar) +- Fix many documentation errors in libcudf. ([#9355](https://github.com/NVIDIA/cudf/pull/9355)) [@karthikeyann](https://github.com/karthikeyann) +- Fixing SubwordTokenizer docs issue ([#9354](https://github.com/NVIDIA/cudf/pull/9354)) [@mayankanand007](https://github.com/mayankanand007) +- Improved deprecation warnings. ([#9347](https://github.com/NVIDIA/cudf/pull/9347)) [@bdice](https://github.com/bdice) +- doc reorder mr, stream to stream, mr ([#9308](https://github.com/NVIDIA/cudf/pull/9308)) [@karthikeyann](https://github.com/karthikeyann) +- Deprecate method parameters to DataFrame.join, DataFrame.merge. ([#9291](https://github.com/NVIDIA/cudf/pull/9291)) [@bdice](https://github.com/bdice) +- Added deprecation warning for `.label_encoding()` ([#9289](https://github.com/NVIDIA/cudf/pull/9289)) [@mayankanand007](https://github.com/mayankanand007) ## 🚀 New Features -- Enable Series.divide and DataFrame.divide ([#9630](https://github.com/rapidsai/cudf/pull/9630)) [@vyasr](https://github.com/vyasr) -- Update `bitmask_and` and `bitmask_or` to return a pair of resulting mask and count of unset bits ([#9616](https://github.com/rapidsai/cudf/pull/9616)) [@PointKernel](https://github.com/PointKernel) -- Add handling of mixed numeric types in `to_dlpack` ([#9585](https://github.com/rapidsai/cudf/pull/9585)) [@galipremsagar](https://github.com/galipremsagar) -- Support re.Pattern object for pat arg in str.replace ([#9573](https://github.com/rapidsai/cudf/pull/9573)) [@davidwendt](https://github.com/davidwendt) -- Add JNI for `lists::drop_list_duplicates` with keys-values input column ([#9553](https://github.com/rapidsai/cudf/pull/9553)) [@ttnghia](https://github.com/ttnghia) -- Support structs column in `min`, `max`, `argmin` and `argmax` groupby aggregate() and scan() ([#9545](https://github.com/rapidsai/cudf/pull/9545)) [@ttnghia](https://github.com/ttnghia) -- Move libcudacxx to use `rapids_cpm` and use newer versions ([#9539](https://github.com/rapidsai/cudf/pull/9539)) [@robertmaynard](https://github.com/robertmaynard) -- Add scan min/max support for chrono types to libcudf reduction-scan (not groupby scan) ([#9518](https://github.com/rapidsai/cudf/pull/9518)) [@davidwendt](https://github.com/davidwendt) -- Support `args=` in `apply` ([#9514](https://github.com/rapidsai/cudf/pull/9514)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add groupby scan min/max support for strings values ([#9502](https://github.com/rapidsai/cudf/pull/9502)) [@davidwendt](https://github.com/davidwendt) -- Add list output option to character_ngrams() function ([#9499](https://github.com/rapidsai/cudf/pull/9499)) [@davidwendt](https://github.com/davidwendt) -- More granular column selection in ORC reader ([#9496](https://github.com/rapidsai/cudf/pull/9496)) [@vuule](https://github.com/vuule) -- add min_periods, ddof to groupby covariance, & correlation aggregation ([#9492](https://github.com/rapidsai/cudf/pull/9492)) [@karthikeyann](https://github.com/karthikeyann) -- Implement Series.datetime.floor ([#9488](https://github.com/rapidsai/cudf/pull/9488)) [@skirui-source](https://github.com/skirui-source) -- Enable linting of CMake files using pre-commit ([#9484](https://github.com/rapidsai/cudf/pull/9484)) [@vyasr](https://github.com/vyasr) -- Add support for single-line regex anchors ^/$ in contains_re ([#9482](https://github.com/rapidsai/cudf/pull/9482)) [@davidwendt](https://github.com/davidwendt) -- Augment `order_by` to Accept a List of `null_precedence` ([#9455](https://github.com/rapidsai/cudf/pull/9455)) [@isVoid](https://github.com/isVoid) -- Add format API for list column of strings ([#9454](https://github.com/rapidsai/cudf/pull/9454)) [@davidwendt](https://github.com/davidwendt) -- Enable Datetime/Timedelta dtypes in Masked UDFs ([#9451](https://github.com/rapidsai/cudf/pull/9451)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add cudf python groupby.diff ([#9446](https://github.com/rapidsai/cudf/pull/9446)) [@karthikeyann](https://github.com/karthikeyann) -- Implement `lists::stable_sort_lists` for stable sorting of elements within each row of lists column ([#9425](https://github.com/rapidsai/cudf/pull/9425)) [@ttnghia](https://github.com/ttnghia) -- add ctest memcheck using cuda-sanitizer ([#9414](https://github.com/rapidsai/cudf/pull/9414)) [@karthikeyann](https://github.com/karthikeyann) -- Support Unary Operations in Masked UDF ([#9409](https://github.com/rapidsai/cudf/pull/9409)) [@isVoid](https://github.com/isVoid) -- Move Several Series Function to Frame ([#9394](https://github.com/rapidsai/cudf/pull/9394)) [@isVoid](https://github.com/isVoid) -- MD5 Python hash API ([#9390](https://github.com/rapidsai/cudf/pull/9390)) [@bdice](https://github.com/bdice) -- Add cudf strings is_title API ([#9380](https://github.com/rapidsai/cudf/pull/9380)) [@davidwendt](https://github.com/davidwendt) -- Enable casting to int64, uint64, and double in AST code. ([#9379](https://github.com/rapidsai/cudf/pull/9379)) [@vyasr](https://github.com/vyasr) -- Add support for writing ORC with map columns ([#9369](https://github.com/rapidsai/cudf/pull/9369)) [@vuule](https://github.com/vuule) -- extract_list_elements() with column_view indices ([#9367](https://github.com/rapidsai/cudf/pull/9367)) [@mythrocks](https://github.com/mythrocks) -- Reimplement `lists::drop_list_duplicates` for keys-values lists columns ([#9345](https://github.com/rapidsai/cudf/pull/9345)) [@ttnghia](https://github.com/ttnghia) -- Support Python UDFs written in terms of rows ([#9343](https://github.com/rapidsai/cudf/pull/9343)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- JNI: Support nested types in ORC writer ([#9334](https://github.com/rapidsai/cudf/pull/9334)) [@firestarman](https://github.com/firestarman) -- Optionally nullify out-of-bounds indices in segmented_gather(). ([#9318](https://github.com/rapidsai/cudf/pull/9318)) [@mythrocks](https://github.com/mythrocks) -- Add shallow hash function and shallow equality comparison for column_view ([#9312](https://github.com/rapidsai/cudf/pull/9312)) [@karthikeyann](https://github.com/karthikeyann) -- Add CudaMemoryBuffer for cudaMalloc memory using RMM cuda_memory_resource ([#9311](https://github.com/rapidsai/cudf/pull/9311)) [@rongou](https://github.com/rongou) -- Add parameters to control row index stride and stripe size in ORC writer ([#9310](https://github.com/rapidsai/cudf/pull/9310)) [@vuule](https://github.com/vuule) -- Add `na_position` param to dask-cudf `sort_values` ([#9264](https://github.com/rapidsai/cudf/pull/9264)) [@charlesbluca](https://github.com/charlesbluca) -- Add `ascending` parameter for dask-cudf `sort_values` ([#9250](https://github.com/rapidsai/cudf/pull/9250)) [@charlesbluca](https://github.com/charlesbluca) -- New array conversion methods ([#9236](https://github.com/rapidsai/cudf/pull/9236)) [@vyasr](https://github.com/vyasr) -- Series `apply` method backed by masked UDFs ([#9217](https://github.com/rapidsai/cudf/pull/9217)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Grouping by frequency and resampling ([#9178](https://github.com/rapidsai/cudf/pull/9178)) [@shwina](https://github.com/shwina) -- Pure-python masked UDFs ([#9174](https://github.com/rapidsai/cudf/pull/9174)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add Covariance, Pearson correlation for sort groupby (libcudf) ([#9154](https://github.com/rapidsai/cudf/pull/9154)) [@karthikeyann](https://github.com/karthikeyann) -- Add `calendrical_month_sequence` in c++ and `date_range` in python ([#8886](https://github.com/rapidsai/cudf/pull/8886)) [@shwina](https://github.com/shwina) +- Enable Series.divide and DataFrame.divide ([#9630](https://github.com/NVIDIA/cudf/pull/9630)) [@vyasr](https://github.com/vyasr) +- Update `bitmask_and` and `bitmask_or` to return a pair of resulting mask and count of unset bits ([#9616](https://github.com/NVIDIA/cudf/pull/9616)) [@PointKernel](https://github.com/PointKernel) +- Add handling of mixed numeric types in `to_dlpack` ([#9585](https://github.com/NVIDIA/cudf/pull/9585)) [@galipremsagar](https://github.com/galipremsagar) +- Support re.Pattern object for pat arg in str.replace ([#9573](https://github.com/NVIDIA/cudf/pull/9573)) [@davidwendt](https://github.com/davidwendt) +- Add JNI for `lists::drop_list_duplicates` with keys-values input column ([#9553](https://github.com/NVIDIA/cudf/pull/9553)) [@ttnghia](https://github.com/ttnghia) +- Support structs column in `min`, `max`, `argmin` and `argmax` groupby aggregate() and scan() ([#9545](https://github.com/NVIDIA/cudf/pull/9545)) [@ttnghia](https://github.com/ttnghia) +- Move libcudacxx to use `rapids_cpm` and use newer versions ([#9539](https://github.com/NVIDIA/cudf/pull/9539)) [@robertmaynard](https://github.com/robertmaynard) +- Add scan min/max support for chrono types to libcudf reduction-scan (not groupby scan) ([#9518](https://github.com/NVIDIA/cudf/pull/9518)) [@davidwendt](https://github.com/davidwendt) +- Support `args=` in `apply` ([#9514](https://github.com/NVIDIA/cudf/pull/9514)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add groupby scan min/max support for strings values ([#9502](https://github.com/NVIDIA/cudf/pull/9502)) [@davidwendt](https://github.com/davidwendt) +- Add list output option to character_ngrams() function ([#9499](https://github.com/NVIDIA/cudf/pull/9499)) [@davidwendt](https://github.com/davidwendt) +- More granular column selection in ORC reader ([#9496](https://github.com/NVIDIA/cudf/pull/9496)) [@vuule](https://github.com/vuule) +- add min_periods, ddof to groupby covariance, & correlation aggregation ([#9492](https://github.com/NVIDIA/cudf/pull/9492)) [@karthikeyann](https://github.com/karthikeyann) +- Implement Series.datetime.floor ([#9488](https://github.com/NVIDIA/cudf/pull/9488)) [@skirui-source](https://github.com/skirui-source) +- Enable linting of CMake files using pre-commit ([#9484](https://github.com/NVIDIA/cudf/pull/9484)) [@vyasr](https://github.com/vyasr) +- Add support for single-line regex anchors ^/$ in contains_re ([#9482](https://github.com/NVIDIA/cudf/pull/9482)) [@davidwendt](https://github.com/davidwendt) +- Augment `order_by` to Accept a List of `null_precedence` ([#9455](https://github.com/NVIDIA/cudf/pull/9455)) [@isVoid](https://github.com/isVoid) +- Add format API for list column of strings ([#9454](https://github.com/NVIDIA/cudf/pull/9454)) [@davidwendt](https://github.com/davidwendt) +- Enable Datetime/Timedelta dtypes in Masked UDFs ([#9451](https://github.com/NVIDIA/cudf/pull/9451)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add cudf python groupby.diff ([#9446](https://github.com/NVIDIA/cudf/pull/9446)) [@karthikeyann](https://github.com/karthikeyann) +- Implement `lists::stable_sort_lists` for stable sorting of elements within each row of lists column ([#9425](https://github.com/NVIDIA/cudf/pull/9425)) [@ttnghia](https://github.com/ttnghia) +- add ctest memcheck using cuda-sanitizer ([#9414](https://github.com/NVIDIA/cudf/pull/9414)) [@karthikeyann](https://github.com/karthikeyann) +- Support Unary Operations in Masked UDF ([#9409](https://github.com/NVIDIA/cudf/pull/9409)) [@isVoid](https://github.com/isVoid) +- Move Several Series Function to Frame ([#9394](https://github.com/NVIDIA/cudf/pull/9394)) [@isVoid](https://github.com/isVoid) +- MD5 Python hash API ([#9390](https://github.com/NVIDIA/cudf/pull/9390)) [@bdice](https://github.com/bdice) +- Add cudf strings is_title API ([#9380](https://github.com/NVIDIA/cudf/pull/9380)) [@davidwendt](https://github.com/davidwendt) +- Enable casting to int64, uint64, and double in AST code. ([#9379](https://github.com/NVIDIA/cudf/pull/9379)) [@vyasr](https://github.com/vyasr) +- Add support for writing ORC with map columns ([#9369](https://github.com/NVIDIA/cudf/pull/9369)) [@vuule](https://github.com/vuule) +- extract_list_elements() with column_view indices ([#9367](https://github.com/NVIDIA/cudf/pull/9367)) [@mythrocks](https://github.com/mythrocks) +- Reimplement `lists::drop_list_duplicates` for keys-values lists columns ([#9345](https://github.com/NVIDIA/cudf/pull/9345)) [@ttnghia](https://github.com/ttnghia) +- Support Python UDFs written in terms of rows ([#9343](https://github.com/NVIDIA/cudf/pull/9343)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- JNI: Support nested types in ORC writer ([#9334](https://github.com/NVIDIA/cudf/pull/9334)) [@firestarman](https://github.com/firestarman) +- Optionally nullify out-of-bounds indices in segmented_gather(). ([#9318](https://github.com/NVIDIA/cudf/pull/9318)) [@mythrocks](https://github.com/mythrocks) +- Add shallow hash function and shallow equality comparison for column_view ([#9312](https://github.com/NVIDIA/cudf/pull/9312)) [@karthikeyann](https://github.com/karthikeyann) +- Add CudaMemoryBuffer for cudaMalloc memory using RMM cuda_memory_resource ([#9311](https://github.com/NVIDIA/cudf/pull/9311)) [@rongou](https://github.com/rongou) +- Add parameters to control row index stride and stripe size in ORC writer ([#9310](https://github.com/NVIDIA/cudf/pull/9310)) [@vuule](https://github.com/vuule) +- Add `na_position` param to dask-cudf `sort_values` ([#9264](https://github.com/NVIDIA/cudf/pull/9264)) [@charlesbluca](https://github.com/charlesbluca) +- Add `ascending` parameter for dask-cudf `sort_values` ([#9250](https://github.com/NVIDIA/cudf/pull/9250)) [@charlesbluca](https://github.com/charlesbluca) +- New array conversion methods ([#9236](https://github.com/NVIDIA/cudf/pull/9236)) [@vyasr](https://github.com/vyasr) +- Series `apply` method backed by masked UDFs ([#9217](https://github.com/NVIDIA/cudf/pull/9217)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Grouping by frequency and resampling ([#9178](https://github.com/NVIDIA/cudf/pull/9178)) [@shwina](https://github.com/shwina) +- Pure-python masked UDFs ([#9174](https://github.com/NVIDIA/cudf/pull/9174)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add Covariance, Pearson correlation for sort groupby (libcudf) ([#9154](https://github.com/NVIDIA/cudf/pull/9154)) [@karthikeyann](https://github.com/karthikeyann) +- Add `calendrical_month_sequence` in c++ and `date_range` in python ([#8886](https://github.com/NVIDIA/cudf/pull/8886)) [@shwina](https://github.com/shwina) ## 🛠️ Improvements -- Followup to PR 9088 comments ([#9659](https://github.com/rapidsai/cudf/pull/9659)) [@cwharris](https://github.com/cwharris) -- Update cuCollections to version that supports installed libcudacxx ([#9633](https://github.com/rapidsai/cudf/pull/9633)) [@robertmaynard](https://github.com/robertmaynard) -- Add `11.5` dev.yml to `cudf` ([#9617](https://github.com/rapidsai/cudf/pull/9617)) [@galipremsagar](https://github.com/galipremsagar) -- Add `xfail` for parquet reader `11.5` issue ([#9612](https://github.com/rapidsai/cudf/pull/9612)) [@galipremsagar](https://github.com/galipremsagar) -- remove deprecated Rmm.initialize method ([#9607](https://github.com/rapidsai/cudf/pull/9607)) [@rongou](https://github.com/rongou) -- Use HostColumnVectorCore for child columns in JCudfSerialization.unpackHostColumnVectors ([#9596](https://github.com/rapidsai/cudf/pull/9596)) [@sperlingxx](https://github.com/sperlingxx) -- Set RMM pool to a fixed size in JNI ([#9583](https://github.com/rapidsai/cudf/pull/9583)) [@rongou](https://github.com/rongou) -- Use nvCOMP for Snappy compression/decompression ([#9582](https://github.com/rapidsai/cudf/pull/9582)) [@vuule](https://github.com/vuule) -- Build CUDA version agnostic packages for dask-cudf ([#9578](https://github.com/rapidsai/cudf/pull/9578)) [@Ethyling](https://github.com/Ethyling) -- Fixed tests warning: "TYPED_TEST_CASE is deprecated, please use TYPED_TEST_SUITE" ([#9574](https://github.com/rapidsai/cudf/pull/9574)) [@ttnghia](https://github.com/ttnghia) -- Enable CMake format in CI and fix style ([#9570](https://github.com/rapidsai/cudf/pull/9570)) [@vyasr](https://github.com/vyasr) -- Add NVTX Start/End Ranges to JNI ([#9563](https://github.com/rapidsai/cudf/pull/9563)) [@abellina](https://github.com/abellina) -- Add librdkafka and python-confluent-kafka to dev conda environments s… ([#9562](https://github.com/rapidsai/cudf/pull/9562)) [@jdye64](https://github.com/jdye64) -- Add offsets_begin/end() to strings_column_view ([#9559](https://github.com/rapidsai/cudf/pull/9559)) [@davidwendt](https://github.com/davidwendt) -- remove alignment options for RMM jni ([#9550](https://github.com/rapidsai/cudf/pull/9550)) [@rongou](https://github.com/rongou) -- Add axis parameter passthrough to `DataFrame` and `Series` take for pandas API compatibility ([#9549](https://github.com/rapidsai/cudf/pull/9549)) [@dantegd](https://github.com/dantegd) -- Remove sizeof and standardize on memory_usage ([#9544](https://github.com/rapidsai/cudf/pull/9544)) [@vyasr](https://github.com/vyasr) -- Adds cudaProfilerStart/cudaProfilerStop in JNI api ([#9543](https://github.com/rapidsai/cudf/pull/9543)) [@abellina](https://github.com/abellina) -- Generalize comparison binary operations ([#9542](https://github.com/rapidsai/cudf/pull/9542)) [@vyasr](https://github.com/vyasr) -- Expose APIs to wrap CUDA or RMM allocations with a Java device buffer instance ([#9538](https://github.com/rapidsai/cudf/pull/9538)) [@jlowe](https://github.com/jlowe) -- Add scan sum support for duration types to libcudf ([#9536](https://github.com/rapidsai/cudf/pull/9536)) [@davidwendt](https://github.com/davidwendt) -- Force inlining to improve AST performance ([#9530](https://github.com/rapidsai/cudf/pull/9530)) [@vyasr](https://github.com/vyasr) -- Generalize some more indexed frame methods ([#9529](https://github.com/rapidsai/cudf/pull/9529)) [@vyasr](https://github.com/vyasr) -- Add Java bindings for rolling window stddev aggregation ([#9527](https://github.com/rapidsai/cudf/pull/9527)) [@razajafri](https://github.com/razajafri) -- catch rmm::out_of_memory exceptions in jni ([#9525](https://github.com/rapidsai/cudf/pull/9525)) [@rongou](https://github.com/rongou) -- Add an overload of `make_empty_column` with `type_id` parameter ([#9524](https://github.com/rapidsai/cudf/pull/9524)) [@ttnghia](https://github.com/ttnghia) -- Accelerate conditional inner joins with larger right tables ([#9523](https://github.com/rapidsai/cudf/pull/9523)) [@vyasr](https://github.com/vyasr) -- Initial pass of generalizing `decimal` support in `cudf` python layer ([#9517](https://github.com/rapidsai/cudf/pull/9517)) [@galipremsagar](https://github.com/galipremsagar) -- Cleanup for flattening nested columns ([#9509](https://github.com/rapidsai/cudf/pull/9509)) [@rwlee](https://github.com/rwlee) -- Enable running tests using RMM arena and async memory resources ([#9506](https://github.com/rapidsai/cudf/pull/9506)) [@rongou](https://github.com/rongou) -- Remove dependency on six. ([#9495](https://github.com/rapidsai/cudf/pull/9495)) [@bdice](https://github.com/bdice) -- Cleanup some libcudf strings gtests ([#9489](https://github.com/rapidsai/cudf/pull/9489)) [@davidwendt](https://github.com/davidwendt) -- Rename strings/array_tests.cu to strings/array_tests.cpp ([#9480](https://github.com/rapidsai/cudf/pull/9480)) [@davidwendt](https://github.com/davidwendt) -- Refactor sorting APIs ([#9464](https://github.com/rapidsai/cudf/pull/9464)) [@vyasr](https://github.com/vyasr) -- Implement DataFrame.hash_values, deprecate DataFrame.hash_columns. ([#9458](https://github.com/rapidsai/cudf/pull/9458)) [@bdice](https://github.com/bdice) -- Deprecate Series.hash_encode. ([#9457](https://github.com/rapidsai/cudf/pull/9457)) [@bdice](https://github.com/bdice) -- Update `conda` recipes for Enhanced Compatibility effort ([#9456](https://github.com/rapidsai/cudf/pull/9456)) [@ajschmidt8](https://github.com/ajschmidt8) -- Small clean up to simplify column selection code in ORC reader ([#9444](https://github.com/rapidsai/cudf/pull/9444)) [@vuule](https://github.com/vuule) -- add missing stream to scalar.is_valid() wherever stream is available ([#9436](https://github.com/rapidsai/cudf/pull/9436)) [@karthikeyann](https://github.com/karthikeyann) -- Adds Deprecation Warnings to `one_hot_encoding` and Implement `get_dummies` with Cython API ([#9435](https://github.com/rapidsai/cudf/pull/9435)) [@isVoid](https://github.com/isVoid) -- Update pre-commit hook URLs. ([#9433](https://github.com/rapidsai/cudf/pull/9433)) [@bdice](https://github.com/bdice) -- Remove pyarrow import in `dask_cudf.io.parquet` ([#9429](https://github.com/rapidsai/cudf/pull/9429)) [@charlesbluca](https://github.com/charlesbluca) -- Miscellaneous improvements for UDFs ([#9422](https://github.com/rapidsai/cudf/pull/9422)) [@isVoid](https://github.com/isVoid) -- Use pre-commit for CI ([#9412](https://github.com/rapidsai/cudf/pull/9412)) [@vyasr](https://github.com/vyasr) -- Update to UCX-Py 0.23 ([#9407](https://github.com/rapidsai/cudf/pull/9407)) [@pentschev](https://github.com/pentschev) -- Expose OutOfBoundsPolicy in JNI for Table.gather ([#9406](https://github.com/rapidsai/cudf/pull/9406)) [@abellina](https://github.com/abellina) -- Improvements to tdigest aggregation code. ([#9403](https://github.com/rapidsai/cudf/pull/9403)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add Java API to deserialize a table to host columns ([#9402](https://github.com/rapidsai/cudf/pull/9402)) [@jlowe](https://github.com/jlowe) -- Frame copy to use __class__ instead of type() ([#9397](https://github.com/rapidsai/cudf/pull/9397)) [@madsbk](https://github.com/madsbk) -- Change all DeprecationWarnings to FutureWarning. ([#9392](https://github.com/rapidsai/cudf/pull/9392)) [@bdice](https://github.com/bdice) -- Update Java nvcomp JNI bindings to nvcomp 2.x API ([#9384](https://github.com/rapidsai/cudf/pull/9384)) [@jbrennan333](https://github.com/jbrennan333) -- Add IndexedFrame class and move SingleColumnFrame to a separate module ([#9378](https://github.com/rapidsai/cudf/pull/9378)) [@vyasr](https://github.com/vyasr) -- Support Arrow NativeFile and PythonFile for remote ORC storage ([#9377](https://github.com/rapidsai/cudf/pull/9377)) [@rjzamora](https://github.com/rjzamora) -- Use Arrow PythonFile for remote CSV storage ([#9376](https://github.com/rapidsai/cudf/pull/9376)) [@rjzamora](https://github.com/rjzamora) -- Add multi-threaded writing to GDS writes ([#9372](https://github.com/rapidsai/cudf/pull/9372)) [@devavret](https://github.com/devavret) -- Miscellaneous column cleanup ([#9370](https://github.com/rapidsai/cudf/pull/9370)) [@vyasr](https://github.com/vyasr) -- Use single kernel to extract all groups in cudf::strings::extract ([#9358](https://github.com/rapidsai/cudf/pull/9358)) [@davidwendt](https://github.com/davidwendt) -- Consolidate binary ops into `Frame` ([#9357](https://github.com/rapidsai/cudf/pull/9357)) [@isVoid](https://github.com/isVoid) -- Move rank scan implementations from scan_inclusive.cu to rank_scan.cu ([#9351](https://github.com/rapidsai/cudf/pull/9351)) [@davidwendt](https://github.com/davidwendt) -- Remove usage of deprecated thrust::host_space_tag. ([#9350](https://github.com/rapidsai/cudf/pull/9350)) [@bdice](https://github.com/bdice) -- Use Default Memory Resource for Temporaries in `reduction.cpp` ([#9344](https://github.com/rapidsai/cudf/pull/9344)) [@isVoid](https://github.com/isVoid) -- Fix Cython compilation warnings. ([#9327](https://github.com/rapidsai/cudf/pull/9327)) [@bdice](https://github.com/bdice) -- Fix some unused variable warnings in libcudf ([#9326](https://github.com/rapidsai/cudf/pull/9326)) [@davidwendt](https://github.com/davidwendt) -- Use optional-iterator for copy-if-else kernel ([#9324](https://github.com/rapidsai/cudf/pull/9324)) [@davidwendt](https://github.com/davidwendt) -- Remove Table class ([#9315](https://github.com/rapidsai/cudf/pull/9315)) [@vyasr](https://github.com/vyasr) -- Unpin `dask` and `distributed` in CI ([#9307](https://github.com/rapidsai/cudf/pull/9307)) [@galipremsagar](https://github.com/galipremsagar) -- Add optional-iterator support to indexalator ([#9306](https://github.com/rapidsai/cudf/pull/9306)) [@davidwendt](https://github.com/davidwendt) -- Consolidate more methods in Frame ([#9305](https://github.com/rapidsai/cudf/pull/9305)) [@vyasr](https://github.com/vyasr) -- Add Arrow-NativeFile and PythonFile support to read_parquet and read_csv in cudf ([#9304](https://github.com/rapidsai/cudf/pull/9304)) [@rjzamora](https://github.com/rjzamora) -- Pin mypy in .pre-commit-config.yaml to match conda environment pinning. ([#9300](https://github.com/rapidsai/cudf/pull/9300)) [@bdice](https://github.com/bdice) -- Use gather.hpp when gather-map exists in device memory ([#9299](https://github.com/rapidsai/cudf/pull/9299)) [@davidwendt](https://github.com/davidwendt) -- Fix Automerger for `Branch-21.12` from `branch-21.10` ([#9285](https://github.com/rapidsai/cudf/pull/9285)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor cuIO timestamp processing with `cuda::std::chrono` ([#9278](https://github.com/rapidsai/cudf/pull/9278)) [@PointKernel](https://github.com/PointKernel) -- Change strings copy_if_else to use optional-iterator instead of pair-iterator ([#9266](https://github.com/rapidsai/cudf/pull/9266)) [@davidwendt](https://github.com/davidwendt) -- Update cudf java bindings to 21.12.0-SNAPSHOT ([#9248](https://github.com/rapidsai/cudf/pull/9248)) [@pxLi](https://github.com/pxLi) -- Various internal MultiIndex improvements ([#9243](https://github.com/rapidsai/cudf/pull/9243)) [@vyasr](https://github.com/vyasr) -- Add detail interface for `split` and `slice(table_view)`, refactors both function with `host_span` ([#9226](https://github.com/rapidsai/cudf/pull/9226)) [@isVoid](https://github.com/isVoid) -- Refactor MD5 implementation. ([#9212](https://github.com/rapidsai/cudf/pull/9212)) [@bdice](https://github.com/bdice) -- Update groupby result_cache to allow sharing intermediate results based on column_view instead of requests. ([#9195](https://github.com/rapidsai/cudf/pull/9195)) [@karthikeyann](https://github.com/karthikeyann) -- Use nvcomp's snappy decompressor in avro reader ([#9181](https://github.com/rapidsai/cudf/pull/9181)) [@devavret](https://github.com/devavret) -- Add `isocalendar` API support ([#9169](https://github.com/rapidsai/cudf/pull/9169)) [@marlenezw](https://github.com/marlenezw) -- Simplify read_json by removing unnecessary reader/impl classes ([#9088](https://github.com/rapidsai/cudf/pull/9088)) [@cwharris](https://github.com/cwharris) -- Simplify read_csv by removing unnecessary reader/impl classes ([#9041](https://github.com/rapidsai/cudf/pull/9041)) [@cwharris](https://github.com/cwharris) -- Refactor hash join with cuCollections multimap ([#8934](https://github.com/rapidsai/cudf/pull/8934)) [@PointKernel](https://github.com/PointKernel) +- Followup to PR 9088 comments ([#9659](https://github.com/NVIDIA/cudf/pull/9659)) [@cwharris](https://github.com/cwharris) +- Update cuCollections to version that supports installed libcudacxx ([#9633](https://github.com/NVIDIA/cudf/pull/9633)) [@robertmaynard](https://github.com/robertmaynard) +- Add `11.5` dev.yml to `cudf` ([#9617](https://github.com/NVIDIA/cudf/pull/9617)) [@galipremsagar](https://github.com/galipremsagar) +- Add `xfail` for parquet reader `11.5` issue ([#9612](https://github.com/NVIDIA/cudf/pull/9612)) [@galipremsagar](https://github.com/galipremsagar) +- remove deprecated Rmm.initialize method ([#9607](https://github.com/NVIDIA/cudf/pull/9607)) [@rongou](https://github.com/rongou) +- Use HostColumnVectorCore for child columns in JCudfSerialization.unpackHostColumnVectors ([#9596](https://github.com/NVIDIA/cudf/pull/9596)) [@sperlingxx](https://github.com/sperlingxx) +- Set RMM pool to a fixed size in JNI ([#9583](https://github.com/NVIDIA/cudf/pull/9583)) [@rongou](https://github.com/rongou) +- Use nvCOMP for Snappy compression/decompression ([#9582](https://github.com/NVIDIA/cudf/pull/9582)) [@vuule](https://github.com/vuule) +- Build CUDA version agnostic packages for dask-cudf ([#9578](https://github.com/NVIDIA/cudf/pull/9578)) [@Ethyling](https://github.com/Ethyling) +- Fixed tests warning: "TYPED_TEST_CASE is deprecated, please use TYPED_TEST_SUITE" ([#9574](https://github.com/NVIDIA/cudf/pull/9574)) [@ttnghia](https://github.com/ttnghia) +- Enable CMake format in CI and fix style ([#9570](https://github.com/NVIDIA/cudf/pull/9570)) [@vyasr](https://github.com/vyasr) +- Add NVTX Start/End Ranges to JNI ([#9563](https://github.com/NVIDIA/cudf/pull/9563)) [@abellina](https://github.com/abellina) +- Add librdkafka and python-confluent-kafka to dev conda environments s… ([#9562](https://github.com/NVIDIA/cudf/pull/9562)) [@jdye64](https://github.com/jdye64) +- Add offsets_begin/end() to strings_column_view ([#9559](https://github.com/NVIDIA/cudf/pull/9559)) [@davidwendt](https://github.com/davidwendt) +- remove alignment options for RMM jni ([#9550](https://github.com/NVIDIA/cudf/pull/9550)) [@rongou](https://github.com/rongou) +- Add axis parameter passthrough to `DataFrame` and `Series` take for pandas API compatibility ([#9549](https://github.com/NVIDIA/cudf/pull/9549)) [@dantegd](https://github.com/dantegd) +- Remove sizeof and standardize on memory_usage ([#9544](https://github.com/NVIDIA/cudf/pull/9544)) [@vyasr](https://github.com/vyasr) +- Adds cudaProfilerStart/cudaProfilerStop in JNI api ([#9543](https://github.com/NVIDIA/cudf/pull/9543)) [@abellina](https://github.com/abellina) +- Generalize comparison binary operations ([#9542](https://github.com/NVIDIA/cudf/pull/9542)) [@vyasr](https://github.com/vyasr) +- Expose APIs to wrap CUDA or RMM allocations with a Java device buffer instance ([#9538](https://github.com/NVIDIA/cudf/pull/9538)) [@jlowe](https://github.com/jlowe) +- Add scan sum support for duration types to libcudf ([#9536](https://github.com/NVIDIA/cudf/pull/9536)) [@davidwendt](https://github.com/davidwendt) +- Force inlining to improve AST performance ([#9530](https://github.com/NVIDIA/cudf/pull/9530)) [@vyasr](https://github.com/vyasr) +- Generalize some more indexed frame methods ([#9529](https://github.com/NVIDIA/cudf/pull/9529)) [@vyasr](https://github.com/vyasr) +- Add Java bindings for rolling window stddev aggregation ([#9527](https://github.com/NVIDIA/cudf/pull/9527)) [@razajafri](https://github.com/razajafri) +- catch rmm::out_of_memory exceptions in jni ([#9525](https://github.com/NVIDIA/cudf/pull/9525)) [@rongou](https://github.com/rongou) +- Add an overload of `make_empty_column` with `type_id` parameter ([#9524](https://github.com/NVIDIA/cudf/pull/9524)) [@ttnghia](https://github.com/ttnghia) +- Accelerate conditional inner joins with larger right tables ([#9523](https://github.com/NVIDIA/cudf/pull/9523)) [@vyasr](https://github.com/vyasr) +- Initial pass of generalizing `decimal` support in `cudf` python layer ([#9517](https://github.com/NVIDIA/cudf/pull/9517)) [@galipremsagar](https://github.com/galipremsagar) +- Cleanup for flattening nested columns ([#9509](https://github.com/NVIDIA/cudf/pull/9509)) [@rwlee](https://github.com/rwlee) +- Enable running tests using RMM arena and async memory resources ([#9506](https://github.com/NVIDIA/cudf/pull/9506)) [@rongou](https://github.com/rongou) +- Remove dependency on six. ([#9495](https://github.com/NVIDIA/cudf/pull/9495)) [@bdice](https://github.com/bdice) +- Cleanup some libcudf strings gtests ([#9489](https://github.com/NVIDIA/cudf/pull/9489)) [@davidwendt](https://github.com/davidwendt) +- Rename strings/array_tests.cu to strings/array_tests.cpp ([#9480](https://github.com/NVIDIA/cudf/pull/9480)) [@davidwendt](https://github.com/davidwendt) +- Refactor sorting APIs ([#9464](https://github.com/NVIDIA/cudf/pull/9464)) [@vyasr](https://github.com/vyasr) +- Implement DataFrame.hash_values, deprecate DataFrame.hash_columns. ([#9458](https://github.com/NVIDIA/cudf/pull/9458)) [@bdice](https://github.com/bdice) +- Deprecate Series.hash_encode. ([#9457](https://github.com/NVIDIA/cudf/pull/9457)) [@bdice](https://github.com/bdice) +- Update `conda` recipes for Enhanced Compatibility effort ([#9456](https://github.com/NVIDIA/cudf/pull/9456)) [@ajschmidt8](https://github.com/ajschmidt8) +- Small clean up to simplify column selection code in ORC reader ([#9444](https://github.com/NVIDIA/cudf/pull/9444)) [@vuule](https://github.com/vuule) +- add missing stream to scalar.is_valid() wherever stream is available ([#9436](https://github.com/NVIDIA/cudf/pull/9436)) [@karthikeyann](https://github.com/karthikeyann) +- Adds Deprecation Warnings to `one_hot_encoding` and Implement `get_dummies` with Cython API ([#9435](https://github.com/NVIDIA/cudf/pull/9435)) [@isVoid](https://github.com/isVoid) +- Update pre-commit hook URLs. ([#9433](https://github.com/NVIDIA/cudf/pull/9433)) [@bdice](https://github.com/bdice) +- Remove pyarrow import in `dask_cudf.io.parquet` ([#9429](https://github.com/NVIDIA/cudf/pull/9429)) [@charlesbluca](https://github.com/charlesbluca) +- Miscellaneous improvements for UDFs ([#9422](https://github.com/NVIDIA/cudf/pull/9422)) [@isVoid](https://github.com/isVoid) +- Use pre-commit for CI ([#9412](https://github.com/NVIDIA/cudf/pull/9412)) [@vyasr](https://github.com/vyasr) +- Update to UCX-Py 0.23 ([#9407](https://github.com/NVIDIA/cudf/pull/9407)) [@pentschev](https://github.com/pentschev) +- Expose OutOfBoundsPolicy in JNI for Table.gather ([#9406](https://github.com/NVIDIA/cudf/pull/9406)) [@abellina](https://github.com/abellina) +- Improvements to tdigest aggregation code. ([#9403](https://github.com/NVIDIA/cudf/pull/9403)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add Java API to deserialize a table to host columns ([#9402](https://github.com/NVIDIA/cudf/pull/9402)) [@jlowe](https://github.com/jlowe) +- Frame copy to use __class__ instead of type() ([#9397](https://github.com/NVIDIA/cudf/pull/9397)) [@madsbk](https://github.com/madsbk) +- Change all DeprecationWarnings to FutureWarning. ([#9392](https://github.com/NVIDIA/cudf/pull/9392)) [@bdice](https://github.com/bdice) +- Update Java nvcomp JNI bindings to nvcomp 2.x API ([#9384](https://github.com/NVIDIA/cudf/pull/9384)) [@jbrennan333](https://github.com/jbrennan333) +- Add IndexedFrame class and move SingleColumnFrame to a separate module ([#9378](https://github.com/NVIDIA/cudf/pull/9378)) [@vyasr](https://github.com/vyasr) +- Support Arrow NativeFile and PythonFile for remote ORC storage ([#9377](https://github.com/NVIDIA/cudf/pull/9377)) [@rjzamora](https://github.com/rjzamora) +- Use Arrow PythonFile for remote CSV storage ([#9376](https://github.com/NVIDIA/cudf/pull/9376)) [@rjzamora](https://github.com/rjzamora) +- Add multi-threaded writing to GDS writes ([#9372](https://github.com/NVIDIA/cudf/pull/9372)) [@devavret](https://github.com/devavret) +- Miscellaneous column cleanup ([#9370](https://github.com/NVIDIA/cudf/pull/9370)) [@vyasr](https://github.com/vyasr) +- Use single kernel to extract all groups in cudf::strings::extract ([#9358](https://github.com/NVIDIA/cudf/pull/9358)) [@davidwendt](https://github.com/davidwendt) +- Consolidate binary ops into `Frame` ([#9357](https://github.com/NVIDIA/cudf/pull/9357)) [@isVoid](https://github.com/isVoid) +- Move rank scan implementations from scan_inclusive.cu to rank_scan.cu ([#9351](https://github.com/NVIDIA/cudf/pull/9351)) [@davidwendt](https://github.com/davidwendt) +- Remove usage of deprecated thrust::host_space_tag. ([#9350](https://github.com/NVIDIA/cudf/pull/9350)) [@bdice](https://github.com/bdice) +- Use Default Memory Resource for Temporaries in `reduction.cpp` ([#9344](https://github.com/NVIDIA/cudf/pull/9344)) [@isVoid](https://github.com/isVoid) +- Fix Cython compilation warnings. ([#9327](https://github.com/NVIDIA/cudf/pull/9327)) [@bdice](https://github.com/bdice) +- Fix some unused variable warnings in libcudf ([#9326](https://github.com/NVIDIA/cudf/pull/9326)) [@davidwendt](https://github.com/davidwendt) +- Use optional-iterator for copy-if-else kernel ([#9324](https://github.com/NVIDIA/cudf/pull/9324)) [@davidwendt](https://github.com/davidwendt) +- Remove Table class ([#9315](https://github.com/NVIDIA/cudf/pull/9315)) [@vyasr](https://github.com/vyasr) +- Unpin `dask` and `distributed` in CI ([#9307](https://github.com/NVIDIA/cudf/pull/9307)) [@galipremsagar](https://github.com/galipremsagar) +- Add optional-iterator support to indexalator ([#9306](https://github.com/NVIDIA/cudf/pull/9306)) [@davidwendt](https://github.com/davidwendt) +- Consolidate more methods in Frame ([#9305](https://github.com/NVIDIA/cudf/pull/9305)) [@vyasr](https://github.com/vyasr) +- Add Arrow-NativeFile and PythonFile support to read_parquet and read_csv in cudf ([#9304](https://github.com/NVIDIA/cudf/pull/9304)) [@rjzamora](https://github.com/rjzamora) +- Pin mypy in .pre-commit-config.yaml to match conda environment pinning. ([#9300](https://github.com/NVIDIA/cudf/pull/9300)) [@bdice](https://github.com/bdice) +- Use gather.hpp when gather-map exists in device memory ([#9299](https://github.com/NVIDIA/cudf/pull/9299)) [@davidwendt](https://github.com/davidwendt) +- Fix Automerger for `Branch-21.12` from `branch-21.10` ([#9285](https://github.com/NVIDIA/cudf/pull/9285)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor cuIO timestamp processing with `cuda::std::chrono` ([#9278](https://github.com/NVIDIA/cudf/pull/9278)) [@PointKernel](https://github.com/PointKernel) +- Change strings copy_if_else to use optional-iterator instead of pair-iterator ([#9266](https://github.com/NVIDIA/cudf/pull/9266)) [@davidwendt](https://github.com/davidwendt) +- Update cudf java bindings to 21.12.0-SNAPSHOT ([#9248](https://github.com/NVIDIA/cudf/pull/9248)) [@pxLi](https://github.com/pxLi) +- Various internal MultiIndex improvements ([#9243](https://github.com/NVIDIA/cudf/pull/9243)) [@vyasr](https://github.com/vyasr) +- Add detail interface for `split` and `slice(table_view)`, refactors both function with `host_span` ([#9226](https://github.com/NVIDIA/cudf/pull/9226)) [@isVoid](https://github.com/isVoid) +- Refactor MD5 implementation. ([#9212](https://github.com/NVIDIA/cudf/pull/9212)) [@bdice](https://github.com/bdice) +- Update groupby result_cache to allow sharing intermediate results based on column_view instead of requests. ([#9195](https://github.com/NVIDIA/cudf/pull/9195)) [@karthikeyann](https://github.com/karthikeyann) +- Use nvcomp's snappy decompressor in avro reader ([#9181](https://github.com/NVIDIA/cudf/pull/9181)) [@devavret](https://github.com/devavret) +- Add `isocalendar` API support ([#9169](https://github.com/NVIDIA/cudf/pull/9169)) [@marlenezw](https://github.com/marlenezw) +- Simplify read_json by removing unnecessary reader/impl classes ([#9088](https://github.com/NVIDIA/cudf/pull/9088)) [@cwharris](https://github.com/cwharris) +- Simplify read_csv by removing unnecessary reader/impl classes ([#9041](https://github.com/NVIDIA/cudf/pull/9041)) [@cwharris](https://github.com/cwharris) +- Refactor hash join with cuCollections multimap ([#8934](https://github.com/NVIDIA/cudf/pull/8934)) [@PointKernel](https://github.com/PointKernel) # cuDF 21.10.00 (7 Oct 2021) ## 🚨 Breaking Changes -- Remove Cython APIs for table view generation ([#9199](https://github.com/rapidsai/cudf/pull/9199)) [@vyasr](https://github.com/vyasr) -- Upgrade `pandas` version in `cudf` ([#9147](https://github.com/rapidsai/cudf/pull/9147)) [@galipremsagar](https://github.com/galipremsagar) -- Make AST operators nullable ([#9096](https://github.com/rapidsai/cudf/pull/9096)) [@vyasr](https://github.com/vyasr) -- Remove the option to pass data types as strings to `read_csv` and `read_json` ([#9079](https://github.com/rapidsai/cudf/pull/9079)) [@vuule](https://github.com/vuule) -- Update JNI java CSV APIs to not use deprecated API ([#9066](https://github.com/rapidsai/cudf/pull/9066)) [@revans2](https://github.com/revans2) -- Support additional format specifiers in from_timestamps ([#9047](https://github.com/rapidsai/cudf/pull/9047)) [@davidwendt](https://github.com/davidwendt) -- Expose expression base class publicly and simplify public AST API ([#9045](https://github.com/rapidsai/cudf/pull/9045)) [@vyasr](https://github.com/vyasr) -- Add support for struct type in ORC writer ([#9025](https://github.com/rapidsai/cudf/pull/9025)) [@vuule](https://github.com/vuule) -- Remove aliases of various api.types APIs from utils.dtypes. ([#9011](https://github.com/rapidsai/cudf/pull/9011)) [@vyasr](https://github.com/vyasr) -- Java bindings for conditional join output sizes ([#9002](https://github.com/rapidsai/cudf/pull/9002)) [@jlowe](https://github.com/jlowe) -- Move compute_column API out of ast namespace ([#8957](https://github.com/rapidsai/cudf/pull/8957)) [@vyasr](https://github.com/vyasr) -- `cudf.dtype` function ([#8949](https://github.com/rapidsai/cudf/pull/8949)) [@shwina](https://github.com/shwina) -- Refactor Frame reductions ([#8944](https://github.com/rapidsai/cudf/pull/8944)) [@vyasr](https://github.com/vyasr) -- Add nested column selection to parquet reader ([#8933](https://github.com/rapidsai/cudf/pull/8933)) [@devavret](https://github.com/devavret) -- JNI Aggregation Type Changes ([#8919](https://github.com/rapidsai/cudf/pull/8919)) [@revans2](https://github.com/revans2) -- Add groupby_aggregation and groupby_scan_aggregation classes and force their usage. ([#8906](https://github.com/rapidsai/cudf/pull/8906)) [@nvdbaranec](https://github.com/nvdbaranec) -- Expand CSV and JSON reader APIs to accept `dtypes` as a vector or map of `data_type` objects ([#8856](https://github.com/rapidsai/cudf/pull/8856)) [@vuule](https://github.com/vuule) -- Change cudf docs theme to pydata theme ([#8746](https://github.com/rapidsai/cudf/pull/8746)) [@galipremsagar](https://github.com/galipremsagar) -- Enable compiled binary ops in libcudf, python and java ([#8741](https://github.com/rapidsai/cudf/pull/8741)) [@karthikeyann](https://github.com/karthikeyann) -- Make groupby transform-like op order match original data order ([#8720](https://github.com/rapidsai/cudf/pull/8720)) [@isVoid](https://github.com/isVoid) +- Remove Cython APIs for table view generation ([#9199](https://github.com/NVIDIA/cudf/pull/9199)) [@vyasr](https://github.com/vyasr) +- Upgrade `pandas` version in `cudf` ([#9147](https://github.com/NVIDIA/cudf/pull/9147)) [@galipremsagar](https://github.com/galipremsagar) +- Make AST operators nullable ([#9096](https://github.com/NVIDIA/cudf/pull/9096)) [@vyasr](https://github.com/vyasr) +- Remove the option to pass data types as strings to `read_csv` and `read_json` ([#9079](https://github.com/NVIDIA/cudf/pull/9079)) [@vuule](https://github.com/vuule) +- Update JNI java CSV APIs to not use deprecated API ([#9066](https://github.com/NVIDIA/cudf/pull/9066)) [@revans2](https://github.com/revans2) +- Support additional format specifiers in from_timestamps ([#9047](https://github.com/NVIDIA/cudf/pull/9047)) [@davidwendt](https://github.com/davidwendt) +- Expose expression base class publicly and simplify public AST API ([#9045](https://github.com/NVIDIA/cudf/pull/9045)) [@vyasr](https://github.com/vyasr) +- Add support for struct type in ORC writer ([#9025](https://github.com/NVIDIA/cudf/pull/9025)) [@vuule](https://github.com/vuule) +- Remove aliases of various api.types APIs from utils.dtypes. ([#9011](https://github.com/NVIDIA/cudf/pull/9011)) [@vyasr](https://github.com/vyasr) +- Java bindings for conditional join output sizes ([#9002](https://github.com/NVIDIA/cudf/pull/9002)) [@jlowe](https://github.com/jlowe) +- Move compute_column API out of ast namespace ([#8957](https://github.com/NVIDIA/cudf/pull/8957)) [@vyasr](https://github.com/vyasr) +- `cudf.dtype` function ([#8949](https://github.com/NVIDIA/cudf/pull/8949)) [@shwina](https://github.com/shwina) +- Refactor Frame reductions ([#8944](https://github.com/NVIDIA/cudf/pull/8944)) [@vyasr](https://github.com/vyasr) +- Add nested column selection to parquet reader ([#8933](https://github.com/NVIDIA/cudf/pull/8933)) [@devavret](https://github.com/devavret) +- JNI Aggregation Type Changes ([#8919](https://github.com/NVIDIA/cudf/pull/8919)) [@revans2](https://github.com/revans2) +- Add groupby_aggregation and groupby_scan_aggregation classes and force their usage. ([#8906](https://github.com/NVIDIA/cudf/pull/8906)) [@nvdbaranec](https://github.com/nvdbaranec) +- Expand CSV and JSON reader APIs to accept `dtypes` as a vector or map of `data_type` objects ([#8856](https://github.com/NVIDIA/cudf/pull/8856)) [@vuule](https://github.com/vuule) +- Change cudf docs theme to pydata theme ([#8746](https://github.com/NVIDIA/cudf/pull/8746)) [@galipremsagar](https://github.com/galipremsagar) +- Enable compiled binary ops in libcudf, python and java ([#8741](https://github.com/NVIDIA/cudf/pull/8741)) [@karthikeyann](https://github.com/karthikeyann) +- Make groupby transform-like op order match original data order ([#8720](https://github.com/NVIDIA/cudf/pull/8720)) [@isVoid](https://github.com/isVoid) ## 🐛 Bug Fixes -- `fixed_point` `cudf::groupby` for `mean` aggregation ([#9296](https://github.com/rapidsai/cudf/pull/9296)) [@codereport](https://github.com/codereport) -- Fix `interleave_columns` when the input string lists column having empty child column ([#9292](https://github.com/rapidsai/cudf/pull/9292)) [@ttnghia](https://github.com/ttnghia) -- Update nvcomp to include fixes for installation of headers ([#9276](https://github.com/rapidsai/cudf/pull/9276)) [@devavret](https://github.com/devavret) -- Fix Java column leak in testParquetWriteMap ([#9271](https://github.com/rapidsai/cudf/pull/9271)) [@jlowe](https://github.com/jlowe) -- Fix call to thrust::reduce_by_key in argmin/argmax libcudf groupby ([#9263](https://github.com/rapidsai/cudf/pull/9263)) [@davidwendt](https://github.com/davidwendt) -- Fixing empty input to getMapValue crashing ([#9262](https://github.com/rapidsai/cudf/pull/9262)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix duplicate names issue in `MultiIndex.deserialize ` ([#9258](https://github.com/rapidsai/cudf/pull/9258)) [@galipremsagar](https://github.com/galipremsagar) -- `Dataframe.sort_index` optimizations ([#9238](https://github.com/rapidsai/cudf/pull/9238)) [@galipremsagar](https://github.com/galipremsagar) -- Temporarily disabling problematic test in parquet writer ([#9230](https://github.com/rapidsai/cudf/pull/9230)) [@devavret](https://github.com/devavret) -- Explicitly disable groupby on unsupported key types. ([#9227](https://github.com/rapidsai/cudf/pull/9227)) [@mythrocks](https://github.com/mythrocks) -- Fix `gather` for sliced input structs column ([#9218](https://github.com/rapidsai/cudf/pull/9218)) [@ttnghia](https://github.com/ttnghia) -- Fix JNI code for left semi and anti joins ([#9207](https://github.com/rapidsai/cudf/pull/9207)) [@jlowe](https://github.com/jlowe) -- Only install thrust when using a non 'system' version ([#9206](https://github.com/rapidsai/cudf/pull/9206)) [@robertmaynard](https://github.com/robertmaynard) -- Remove zlib from libcudf public CMake dependencies ([#9204](https://github.com/rapidsai/cudf/pull/9204)) [@robertmaynard](https://github.com/robertmaynard) -- Fix out-of-bounds memory read in orc gpuEncodeOrcColumnData ([#9196](https://github.com/rapidsai/cudf/pull/9196)) [@davidwendt](https://github.com/davidwendt) -- Fix `gather()` for `STRUCT` inputs with no nulls in members. ([#9194](https://github.com/rapidsai/cudf/pull/9194)) [@mythrocks](https://github.com/mythrocks) -- get_cucollections properly uses rapids_cpm_find ([#9189](https://github.com/rapidsai/cudf/pull/9189)) [@robertmaynard](https://github.com/robertmaynard) -- rapids-export correctly reference build code block and doc strings ([#9186](https://github.com/rapidsai/cudf/pull/9186)) [@robertmaynard](https://github.com/robertmaynard) -- Fix logic while parsing the sum statistic for numerical orc columns ([#9183](https://github.com/rapidsai/cudf/pull/9183)) [@ayushdg](https://github.com/ayushdg) -- Add handling for nulls in `dask_cudf.sorting.quantile_divisions` ([#9171](https://github.com/rapidsai/cudf/pull/9171)) [@charlesbluca](https://github.com/charlesbluca) -- Approximate overflow detection in ORC statistics ([#9163](https://github.com/rapidsai/cudf/pull/9163)) [@vuule](https://github.com/vuule) -- Use decimal precision metadata when reading from parquet files ([#9162](https://github.com/rapidsai/cudf/pull/9162)) [@shwina](https://github.com/shwina) -- Fix variable name in Java build script ([#9161](https://github.com/rapidsai/cudf/pull/9161)) [@jlowe](https://github.com/jlowe) -- Import rapids-cmake modules using the correct cmake variable. ([#9149](https://github.com/rapidsai/cudf/pull/9149)) [@robertmaynard](https://github.com/robertmaynard) -- Fix conditional joins with empty left table ([#9146](https://github.com/rapidsai/cudf/pull/9146)) [@vyasr](https://github.com/vyasr) -- Fix joining on indexes with duplicate level names ([#9137](https://github.com/rapidsai/cudf/pull/9137)) [@shwina](https://github.com/shwina) -- Fixes missing child column name in dtype while reading ORC file. ([#9134](https://github.com/rapidsai/cudf/pull/9134)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Apply type metadata after column is slice-copied ([#9131](https://github.com/rapidsai/cudf/pull/9131)) [@isVoid](https://github.com/isVoid) -- Fix a bug: inner_join_size return zero if build table is empty ([#9128](https://github.com/rapidsai/cudf/pull/9128)) [@PointKernel](https://github.com/PointKernel) -- Fix multi hive-partition parquet reading in dask-cudf ([#9122](https://github.com/rapidsai/cudf/pull/9122)) [@rjzamora](https://github.com/rjzamora) -- Support null literals in expressions ([#9117](https://github.com/rapidsai/cudf/pull/9117)) [@vyasr](https://github.com/vyasr) -- Fix cudf::hash_join output size for struct joins ([#9107](https://github.com/rapidsai/cudf/pull/9107)) [@jlowe](https://github.com/jlowe) -- Import fix ([#9104](https://github.com/rapidsai/cudf/pull/9104)) [@shwina](https://github.com/shwina) -- Fix cudf::strings::is_fixed_point checking of overflow for decimal32 ([#9093](https://github.com/rapidsai/cudf/pull/9093)) [@davidwendt](https://github.com/davidwendt) -- Fix branch_stack calculation in `row_bit_count()` ([#9076](https://github.com/rapidsai/cudf/pull/9076)) [@mythrocks](https://github.com/mythrocks) -- Fetch rapids-cmake to work around cuCollection cmake issue ([#9075](https://github.com/rapidsai/cudf/pull/9075)) [@jlowe](https://github.com/jlowe) -- Fix compilation errors in groupby benchmarks. ([#9072](https://github.com/rapidsai/cudf/pull/9072)) [@nvdbaranec](https://github.com/nvdbaranec) -- Preserve float16 upscaling ([#9069](https://github.com/rapidsai/cudf/pull/9069)) [@galipremsagar](https://github.com/galipremsagar) -- Fix memcheck read error in libcudf contiguous_split ([#9067](https://github.com/rapidsai/cudf/pull/9067)) [@davidwendt](https://github.com/davidwendt) -- Add support for reading ORC file with no row group index ([#9060](https://github.com/rapidsai/cudf/pull/9060)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Various multiindex related fixes ([#9036](https://github.com/rapidsai/cudf/pull/9036)) [@shwina](https://github.com/shwina) -- Avoid rebuilding cython in build.sh ([#9034](https://github.com/rapidsai/cudf/pull/9034)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add support for percentile dispatch in `dask_cudf` ([#9031](https://github.com/rapidsai/cudf/pull/9031)) [@galipremsagar](https://github.com/galipremsagar) -- cudf resolve nvcc 11.0 compiler crashes during codegen ([#9028](https://github.com/rapidsai/cudf/pull/9028)) [@robertmaynard](https://github.com/robertmaynard) -- Fetch correct grouping keys `agg` of dask groupby ([#9022](https://github.com/rapidsai/cudf/pull/9022)) [@galipremsagar](https://github.com/galipremsagar) -- Allow `where()` to work with a Series and `other=cudf.NA` ([#9019](https://github.com/rapidsai/cudf/pull/9019)) [@sarahyurick](https://github.com/sarahyurick) -- Use correct index when returning Series from `GroupBy.apply()` ([#9016](https://github.com/rapidsai/cudf/pull/9016)) [@charlesbluca](https://github.com/charlesbluca) -- Fix `Dataframe` indexer setitem when array is passed ([#9006](https://github.com/rapidsai/cudf/pull/9006)) [@galipremsagar](https://github.com/galipremsagar) -- Fix ORC reading of files with struct columns that have null values ([#9005](https://github.com/rapidsai/cudf/pull/9005)) [@vuule](https://github.com/vuule) -- Ensure JNI native libraries load when CompiledExpression loads ([#8997](https://github.com/rapidsai/cudf/pull/8997)) [@jlowe](https://github.com/jlowe) -- Fix memory read error in get_dremel_data in page_enc.cu ([#8995](https://github.com/rapidsai/cudf/pull/8995)) [@davidwendt](https://github.com/davidwendt) -- Fix memory write error in get_list_child_to_list_row_mapping utility ([#8994](https://github.com/rapidsai/cudf/pull/8994)) [@davidwendt](https://github.com/davidwendt) -- Fix debug compile error for csv_test.cpp ([#8981](https://github.com/rapidsai/cudf/pull/8981)) [@davidwendt](https://github.com/davidwendt) -- Fix memory read/write error in concatenate_lists_ignore_null ([#8978](https://github.com/rapidsai/cudf/pull/8978)) [@davidwendt](https://github.com/davidwendt) -- Fix concatenation of `cudf.RangeIndex` ([#8970](https://github.com/rapidsai/cudf/pull/8970)) [@galipremsagar](https://github.com/galipremsagar) -- Java conditional joins should not require matching column counts ([#8955](https://github.com/rapidsai/cudf/pull/8955)) [@jlowe](https://github.com/jlowe) -- Fix concatenate empty structs ([#8947](https://github.com/rapidsai/cudf/pull/8947)) [@sperlingxx](https://github.com/sperlingxx) -- Fix cuda-memcheck errors for some libcudf functions ([#8941](https://github.com/rapidsai/cudf/pull/8941)) [@davidwendt](https://github.com/davidwendt) -- Apply series name to result of `SeriesGroupby.apply()` ([#8939](https://github.com/rapidsai/cudf/pull/8939)) [@charlesbluca](https://github.com/charlesbluca) -- `cdef packed_columns` as `cppclass` instead of `struct` ([#8936](https://github.com/rapidsai/cudf/pull/8936)) [@charlesbluca](https://github.com/charlesbluca) -- Inserting a `cudf.NA` into a DataFrame ([#8923](https://github.com/rapidsai/cudf/pull/8923)) [@sarahyurick](https://github.com/sarahyurick) -- Support casting with Pandas dtype aliases ([#8920](https://github.com/rapidsai/cudf/pull/8920)) [@sarahyurick](https://github.com/sarahyurick) -- Allow `sort_values` to accept same `kind` values as Pandas ([#8912](https://github.com/rapidsai/cudf/pull/8912)) [@sarahyurick](https://github.com/sarahyurick) -- Enable casting to pandas nullable dtypes ([#8889](https://github.com/rapidsai/cudf/pull/8889)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix libcudf memory errors ([#8884](https://github.com/rapidsai/cudf/pull/8884)) [@karthikeyann](https://github.com/karthikeyann) -- Throw KeyError when accessing field from struct with nonexistent key ([#8880](https://github.com/rapidsai/cudf/pull/8880)) [@NV-jpt](https://github.com/NV-jpt) -- replace auto with auto& ref for cast<&> ([#8866](https://github.com/rapidsai/cudf/pull/8866)) [@karthikeyann](https://github.com/karthikeyann) -- Add missing include<optional> in binops ([#8864](https://github.com/rapidsai/cudf/pull/8864)) [@karthikeyann](https://github.com/karthikeyann) -- Fix `select_dtypes` to work when non-class dtypes present in dataframe ([#8849](https://github.com/rapidsai/cudf/pull/8849)) [@sarahyurick](https://github.com/sarahyurick) -- Re-enable JSON tests ([#8843](https://github.com/rapidsai/cudf/pull/8843)) [@vuule](https://github.com/vuule) -- Support header with embedded delimiter in csv writer ([#8798](https://github.com/rapidsai/cudf/pull/8798)) [@davidwendt](https://github.com/davidwendt) +- `fixed_point` `cudf::groupby` for `mean` aggregation ([#9296](https://github.com/NVIDIA/cudf/pull/9296)) [@codereport](https://github.com/codereport) +- Fix `interleave_columns` when the input string lists column having empty child column ([#9292](https://github.com/NVIDIA/cudf/pull/9292)) [@ttnghia](https://github.com/ttnghia) +- Update nvcomp to include fixes for installation of headers ([#9276](https://github.com/NVIDIA/cudf/pull/9276)) [@devavret](https://github.com/devavret) +- Fix Java column leak in testParquetWriteMap ([#9271](https://github.com/NVIDIA/cudf/pull/9271)) [@jlowe](https://github.com/jlowe) +- Fix call to thrust::reduce_by_key in argmin/argmax libcudf groupby ([#9263](https://github.com/NVIDIA/cudf/pull/9263)) [@davidwendt](https://github.com/davidwendt) +- Fixing empty input to getMapValue crashing ([#9262](https://github.com/NVIDIA/cudf/pull/9262)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix duplicate names issue in `MultiIndex.deserialize ` ([#9258](https://github.com/NVIDIA/cudf/pull/9258)) [@galipremsagar](https://github.com/galipremsagar) +- `Dataframe.sort_index` optimizations ([#9238](https://github.com/NVIDIA/cudf/pull/9238)) [@galipremsagar](https://github.com/galipremsagar) +- Temporarily disabling problematic test in parquet writer ([#9230](https://github.com/NVIDIA/cudf/pull/9230)) [@devavret](https://github.com/devavret) +- Explicitly disable groupby on unsupported key types. ([#9227](https://github.com/NVIDIA/cudf/pull/9227)) [@mythrocks](https://github.com/mythrocks) +- Fix `gather` for sliced input structs column ([#9218](https://github.com/NVIDIA/cudf/pull/9218)) [@ttnghia](https://github.com/ttnghia) +- Fix JNI code for left semi and anti joins ([#9207](https://github.com/NVIDIA/cudf/pull/9207)) [@jlowe](https://github.com/jlowe) +- Only install thrust when using a non 'system' version ([#9206](https://github.com/NVIDIA/cudf/pull/9206)) [@robertmaynard](https://github.com/robertmaynard) +- Remove zlib from libcudf public CMake dependencies ([#9204](https://github.com/NVIDIA/cudf/pull/9204)) [@robertmaynard](https://github.com/robertmaynard) +- Fix out-of-bounds memory read in orc gpuEncodeOrcColumnData ([#9196](https://github.com/NVIDIA/cudf/pull/9196)) [@davidwendt](https://github.com/davidwendt) +- Fix `gather()` for `STRUCT` inputs with no nulls in members. ([#9194](https://github.com/NVIDIA/cudf/pull/9194)) [@mythrocks](https://github.com/mythrocks) +- get_cucollections properly uses rapids_cpm_find ([#9189](https://github.com/NVIDIA/cudf/pull/9189)) [@robertmaynard](https://github.com/robertmaynard) +- rapids-export correctly reference build code block and doc strings ([#9186](https://github.com/NVIDIA/cudf/pull/9186)) [@robertmaynard](https://github.com/robertmaynard) +- Fix logic while parsing the sum statistic for numerical orc columns ([#9183](https://github.com/NVIDIA/cudf/pull/9183)) [@ayushdg](https://github.com/ayushdg) +- Add handling for nulls in `dask_cudf.sorting.quantile_divisions` ([#9171](https://github.com/NVIDIA/cudf/pull/9171)) [@charlesbluca](https://github.com/charlesbluca) +- Approximate overflow detection in ORC statistics ([#9163](https://github.com/NVIDIA/cudf/pull/9163)) [@vuule](https://github.com/vuule) +- Use decimal precision metadata when reading from parquet files ([#9162](https://github.com/NVIDIA/cudf/pull/9162)) [@shwina](https://github.com/shwina) +- Fix variable name in Java build script ([#9161](https://github.com/NVIDIA/cudf/pull/9161)) [@jlowe](https://github.com/jlowe) +- Import rapids-cmake modules using the correct cmake variable. ([#9149](https://github.com/NVIDIA/cudf/pull/9149)) [@robertmaynard](https://github.com/robertmaynard) +- Fix conditional joins with empty left table ([#9146](https://github.com/NVIDIA/cudf/pull/9146)) [@vyasr](https://github.com/vyasr) +- Fix joining on indexes with duplicate level names ([#9137](https://github.com/NVIDIA/cudf/pull/9137)) [@shwina](https://github.com/shwina) +- Fixes missing child column name in dtype while reading ORC file. ([#9134](https://github.com/NVIDIA/cudf/pull/9134)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Apply type metadata after column is slice-copied ([#9131](https://github.com/NVIDIA/cudf/pull/9131)) [@isVoid](https://github.com/isVoid) +- Fix a bug: inner_join_size return zero if build table is empty ([#9128](https://github.com/NVIDIA/cudf/pull/9128)) [@PointKernel](https://github.com/PointKernel) +- Fix multi hive-partition parquet reading in dask-cudf ([#9122](https://github.com/NVIDIA/cudf/pull/9122)) [@rjzamora](https://github.com/rjzamora) +- Support null literals in expressions ([#9117](https://github.com/NVIDIA/cudf/pull/9117)) [@vyasr](https://github.com/vyasr) +- Fix cudf::hash_join output size for struct joins ([#9107](https://github.com/NVIDIA/cudf/pull/9107)) [@jlowe](https://github.com/jlowe) +- Import fix ([#9104](https://github.com/NVIDIA/cudf/pull/9104)) [@shwina](https://github.com/shwina) +- Fix cudf::strings::is_fixed_point checking of overflow for decimal32 ([#9093](https://github.com/NVIDIA/cudf/pull/9093)) [@davidwendt](https://github.com/davidwendt) +- Fix branch_stack calculation in `row_bit_count()` ([#9076](https://github.com/NVIDIA/cudf/pull/9076)) [@mythrocks](https://github.com/mythrocks) +- Fetch rapids-cmake to work around cuCollection cmake issue ([#9075](https://github.com/NVIDIA/cudf/pull/9075)) [@jlowe](https://github.com/jlowe) +- Fix compilation errors in groupby benchmarks. ([#9072](https://github.com/NVIDIA/cudf/pull/9072)) [@nvdbaranec](https://github.com/nvdbaranec) +- Preserve float16 upscaling ([#9069](https://github.com/NVIDIA/cudf/pull/9069)) [@galipremsagar](https://github.com/galipremsagar) +- Fix memcheck read error in libcudf contiguous_split ([#9067](https://github.com/NVIDIA/cudf/pull/9067)) [@davidwendt](https://github.com/davidwendt) +- Add support for reading ORC file with no row group index ([#9060](https://github.com/NVIDIA/cudf/pull/9060)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Various multiindex related fixes ([#9036](https://github.com/NVIDIA/cudf/pull/9036)) [@shwina](https://github.com/shwina) +- Avoid rebuilding cython in build.sh ([#9034](https://github.com/NVIDIA/cudf/pull/9034)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add support for percentile dispatch in `dask_cudf` ([#9031](https://github.com/NVIDIA/cudf/pull/9031)) [@galipremsagar](https://github.com/galipremsagar) +- cudf resolve nvcc 11.0 compiler crashes during codegen ([#9028](https://github.com/NVIDIA/cudf/pull/9028)) [@robertmaynard](https://github.com/robertmaynard) +- Fetch correct grouping keys `agg` of dask groupby ([#9022](https://github.com/NVIDIA/cudf/pull/9022)) [@galipremsagar](https://github.com/galipremsagar) +- Allow `where()` to work with a Series and `other=cudf.NA` ([#9019](https://github.com/NVIDIA/cudf/pull/9019)) [@sarahyurick](https://github.com/sarahyurick) +- Use correct index when returning Series from `GroupBy.apply()` ([#9016](https://github.com/NVIDIA/cudf/pull/9016)) [@charlesbluca](https://github.com/charlesbluca) +- Fix `Dataframe` indexer setitem when array is passed ([#9006](https://github.com/NVIDIA/cudf/pull/9006)) [@galipremsagar](https://github.com/galipremsagar) +- Fix ORC reading of files with struct columns that have null values ([#9005](https://github.com/NVIDIA/cudf/pull/9005)) [@vuule](https://github.com/vuule) +- Ensure JNI native libraries load when CompiledExpression loads ([#8997](https://github.com/NVIDIA/cudf/pull/8997)) [@jlowe](https://github.com/jlowe) +- Fix memory read error in get_dremel_data in page_enc.cu ([#8995](https://github.com/NVIDIA/cudf/pull/8995)) [@davidwendt](https://github.com/davidwendt) +- Fix memory write error in get_list_child_to_list_row_mapping utility ([#8994](https://github.com/NVIDIA/cudf/pull/8994)) [@davidwendt](https://github.com/davidwendt) +- Fix debug compile error for csv_test.cpp ([#8981](https://github.com/NVIDIA/cudf/pull/8981)) [@davidwendt](https://github.com/davidwendt) +- Fix memory read/write error in concatenate_lists_ignore_null ([#8978](https://github.com/NVIDIA/cudf/pull/8978)) [@davidwendt](https://github.com/davidwendt) +- Fix concatenation of `cudf.RangeIndex` ([#8970](https://github.com/NVIDIA/cudf/pull/8970)) [@galipremsagar](https://github.com/galipremsagar) +- Java conditional joins should not require matching column counts ([#8955](https://github.com/NVIDIA/cudf/pull/8955)) [@jlowe](https://github.com/jlowe) +- Fix concatenate empty structs ([#8947](https://github.com/NVIDIA/cudf/pull/8947)) [@sperlingxx](https://github.com/sperlingxx) +- Fix cuda-memcheck errors for some libcudf functions ([#8941](https://github.com/NVIDIA/cudf/pull/8941)) [@davidwendt](https://github.com/davidwendt) +- Apply series name to result of `SeriesGroupby.apply()` ([#8939](https://github.com/NVIDIA/cudf/pull/8939)) [@charlesbluca](https://github.com/charlesbluca) +- `cdef packed_columns` as `cppclass` instead of `struct` ([#8936](https://github.com/NVIDIA/cudf/pull/8936)) [@charlesbluca](https://github.com/charlesbluca) +- Inserting a `cudf.NA` into a DataFrame ([#8923](https://github.com/NVIDIA/cudf/pull/8923)) [@sarahyurick](https://github.com/sarahyurick) +- Support casting with Pandas dtype aliases ([#8920](https://github.com/NVIDIA/cudf/pull/8920)) [@sarahyurick](https://github.com/sarahyurick) +- Allow `sort_values` to accept same `kind` values as Pandas ([#8912](https://github.com/NVIDIA/cudf/pull/8912)) [@sarahyurick](https://github.com/sarahyurick) +- Enable casting to pandas nullable dtypes ([#8889](https://github.com/NVIDIA/cudf/pull/8889)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix libcudf memory errors ([#8884](https://github.com/NVIDIA/cudf/pull/8884)) [@karthikeyann](https://github.com/karthikeyann) +- Throw KeyError when accessing field from struct with nonexistent key ([#8880](https://github.com/NVIDIA/cudf/pull/8880)) [@NV-jpt](https://github.com/NV-jpt) +- replace auto with auto& ref for cast<&> ([#8866](https://github.com/NVIDIA/cudf/pull/8866)) [@karthikeyann](https://github.com/karthikeyann) +- Add missing include<optional> in binops ([#8864](https://github.com/NVIDIA/cudf/pull/8864)) [@karthikeyann](https://github.com/karthikeyann) +- Fix `select_dtypes` to work when non-class dtypes present in dataframe ([#8849](https://github.com/NVIDIA/cudf/pull/8849)) [@sarahyurick](https://github.com/sarahyurick) +- Re-enable JSON tests ([#8843](https://github.com/NVIDIA/cudf/pull/8843)) [@vuule](https://github.com/vuule) +- Support header with embedded delimiter in csv writer ([#8798](https://github.com/NVIDIA/cudf/pull/8798)) [@davidwendt](https://github.com/davidwendt) ## 📖 Documentation -- Add IO docs page in `cudf` documentation ([#9145](https://github.com/rapidsai/cudf/pull/9145)) [@galipremsagar](https://github.com/galipremsagar) -- use correct namespace in cuio code examples ([#9037](https://github.com/rapidsai/cudf/pull/9037)) [@cwharris](https://github.com/cwharris) -- Restructuring `Contributing doc` ([#9026](https://github.com/rapidsai/cudf/pull/9026)) [@iskode](https://github.com/iskode) -- Update stable version in readme ([#9008](https://github.com/rapidsai/cudf/pull/9008)) [@galipremsagar](https://github.com/galipremsagar) -- Add spans and more include guidelines to libcudf developer guide ([#8931](https://github.com/rapidsai/cudf/pull/8931)) [@harrism](https://github.com/harrism) -- Update Java build instructions to mention Arrow S3 and Docker ([#8867](https://github.com/rapidsai/cudf/pull/8867)) [@jlowe](https://github.com/jlowe) -- List GDS-enabled formats in the docs ([#8805](https://github.com/rapidsai/cudf/pull/8805)) [@vuule](https://github.com/vuule) -- Change cudf docs theme to pydata theme ([#8746](https://github.com/rapidsai/cudf/pull/8746)) [@galipremsagar](https://github.com/galipremsagar) +- Add IO docs page in `cudf` documentation ([#9145](https://github.com/NVIDIA/cudf/pull/9145)) [@galipremsagar](https://github.com/galipremsagar) +- use correct namespace in cuio code examples ([#9037](https://github.com/NVIDIA/cudf/pull/9037)) [@cwharris](https://github.com/cwharris) +- Restructuring `Contributing doc` ([#9026](https://github.com/NVIDIA/cudf/pull/9026)) [@iskode](https://github.com/iskode) +- Update stable version in readme ([#9008](https://github.com/NVIDIA/cudf/pull/9008)) [@galipremsagar](https://github.com/galipremsagar) +- Add spans and more include guidelines to libcudf developer guide ([#8931](https://github.com/NVIDIA/cudf/pull/8931)) [@harrism](https://github.com/harrism) +- Update Java build instructions to mention Arrow S3 and Docker ([#8867](https://github.com/NVIDIA/cudf/pull/8867)) [@jlowe](https://github.com/jlowe) +- List GDS-enabled formats in the docs ([#8805](https://github.com/NVIDIA/cudf/pull/8805)) [@vuule](https://github.com/vuule) +- Change cudf docs theme to pydata theme ([#8746](https://github.com/NVIDIA/cudf/pull/8746)) [@galipremsagar](https://github.com/galipremsagar) ## 🚀 New Features -- Revert "Add shallow hash function and shallow equality comparison for column_view ([#9185)" (#9283](https://github.com/rapidsai/cudf/pull/9185)" (#9283)) [@karthikeyann](https://github.com/karthikeyann) -- Align `DataFrame.apply` signature with pandas ([#9275](https://github.com/rapidsai/cudf/pull/9275)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add struct type support for `drop_list_duplicates` ([#9202](https://github.com/rapidsai/cudf/pull/9202)) [@ttnghia](https://github.com/ttnghia) -- support CUDA async memory resource in JNI ([#9201](https://github.com/rapidsai/cudf/pull/9201)) [@rongou](https://github.com/rongou) -- Add shallow hash function and shallow equality comparison for column_view ([#9185](https://github.com/rapidsai/cudf/pull/9185)) [@karthikeyann](https://github.com/karthikeyann) -- Superimpose null masks for STRUCT columns. ([#9144](https://github.com/rapidsai/cudf/pull/9144)) [@mythrocks](https://github.com/mythrocks) -- Implemented bindings for `ceil` timestamp operation ([#9141](https://github.com/rapidsai/cudf/pull/9141)) [@shaneding](https://github.com/shaneding) -- Adding MAP type support for ORC Reader ([#9132](https://github.com/rapidsai/cudf/pull/9132)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Implement `interleave_columns` for lists with arbitrary nested type ([#9130](https://github.com/rapidsai/cudf/pull/9130)) [@ttnghia](https://github.com/ttnghia) -- Add python bindings to fixed-size window and groupby `rolling.var`, `rolling.std` ([#9097](https://github.com/rapidsai/cudf/pull/9097)) [@isVoid](https://github.com/isVoid) -- Make AST operators nullable ([#9096](https://github.com/rapidsai/cudf/pull/9096)) [@vyasr](https://github.com/vyasr) -- Java bindings for approx_percentile ([#9094](https://github.com/rapidsai/cudf/pull/9094)) [@andygrove](https://github.com/andygrove) -- Add `dseries.struct.explode` ([#9086](https://github.com/rapidsai/cudf/pull/9086)) [@isVoid](https://github.com/isVoid) -- Add support for BaseIndexer in Rolling APIs ([#9085](https://github.com/rapidsai/cudf/pull/9085)) [@galipremsagar](https://github.com/galipremsagar) -- Remove the option to pass data types as strings to `read_csv` and `read_json` ([#9079](https://github.com/rapidsai/cudf/pull/9079)) [@vuule](https://github.com/vuule) -- Add handling for nested dicts in dask-cudf groupby ([#9054](https://github.com/rapidsai/cudf/pull/9054)) [@charlesbluca](https://github.com/charlesbluca) -- Added Series.dt.is_quarter_start and Series.dt.is_quarter_end ([#9046](https://github.com/rapidsai/cudf/pull/9046)) [@TravisHester](https://github.com/TravisHester) -- Support nested types for nth_element reduction ([#9043](https://github.com/rapidsai/cudf/pull/9043)) [@sperlingxx](https://github.com/sperlingxx) -- Update sort groupby to use non-atomic operation ([#9035](https://github.com/rapidsai/cudf/pull/9035)) [@karthikeyann](https://github.com/karthikeyann) -- Add support for struct type in ORC writer ([#9025](https://github.com/rapidsai/cudf/pull/9025)) [@vuule](https://github.com/vuule) -- Implement `interleave_columns` for structs columns ([#9012](https://github.com/rapidsai/cudf/pull/9012)) [@ttnghia](https://github.com/ttnghia) -- Add groupby first and last aggregations ([#9004](https://github.com/rapidsai/cudf/pull/9004)) [@shwina](https://github.com/shwina) -- Add `DecimalBaseColumn` and move `as_decimal_column` ([#9001](https://github.com/rapidsai/cudf/pull/9001)) [@isVoid](https://github.com/isVoid) -- Python/Cython bindings for multibyte_split ([#8998](https://github.com/rapidsai/cudf/pull/8998)) [@jdye64](https://github.com/jdye64) -- Support scalar `months` in `add_calendrical_months`, extends API to INT32 support ([#8991](https://github.com/rapidsai/cudf/pull/8991)) [@isVoid](https://github.com/isVoid) -- Added Series.dt.is_month_end ([#8989](https://github.com/rapidsai/cudf/pull/8989)) [@TravisHester](https://github.com/TravisHester) -- Support for using tdigests to compute approximate percentiles. ([#8983](https://github.com/rapidsai/cudf/pull/8983)) [@nvdbaranec](https://github.com/nvdbaranec) -- Support "unflatten" of columns flattened via `flatten_nested_columns()`: ([#8956](https://github.com/rapidsai/cudf/pull/8956)) [@mythrocks](https://github.com/mythrocks) -- Implement timestamp ceil ([#8942](https://github.com/rapidsai/cudf/pull/8942)) [@shaneding](https://github.com/shaneding) -- Add nested column selection to parquet reader ([#8933](https://github.com/rapidsai/cudf/pull/8933)) [@devavret](https://github.com/devavret) -- Expose conditional join size calculation ([#8928](https://github.com/rapidsai/cudf/pull/8928)) [@vyasr](https://github.com/vyasr) -- Support Nulls in Timeseries Generator ([#8925](https://github.com/rapidsai/cudf/pull/8925)) [@isVoid](https://github.com/isVoid) -- Avoid index equality check in `_CPackedColumns.from_py_table()` ([#8917](https://github.com/rapidsai/cudf/pull/8917)) [@charlesbluca](https://github.com/charlesbluca) -- Add dot product binary op ([#8909](https://github.com/rapidsai/cudf/pull/8909)) [@charlesbluca](https://github.com/charlesbluca) -- Expose `days_in_month` function in libcudf and add python bindings ([#8892](https://github.com/rapidsai/cudf/pull/8892)) [@isVoid](https://github.com/isVoid) -- Series string repeat ([#8882](https://github.com/rapidsai/cudf/pull/8882)) [@sarahyurick](https://github.com/sarahyurick) -- Python binding for quarters ([#8862](https://github.com/rapidsai/cudf/pull/8862)) [@shaneding](https://github.com/shaneding) -- Expand CSV and JSON reader APIs to accept `dtypes` as a vector or map of `data_type` objects ([#8856](https://github.com/rapidsai/cudf/pull/8856)) [@vuule](https://github.com/vuule) -- Add Java bindings for AST transform ([#8846](https://github.com/rapidsai/cudf/pull/8846)) [@jlowe](https://github.com/jlowe) -- Series datetime is_month_start ([#8844](https://github.com/rapidsai/cudf/pull/8844)) [@sarahyurick](https://github.com/sarahyurick) -- Support bracket syntax for cudf::strings::replace_with_backrefs group index values ([#8841](https://github.com/rapidsai/cudf/pull/8841)) [@davidwendt](https://github.com/davidwendt) -- Support `VARIANCE` and `STD` aggregation in rolling op ([#8809](https://github.com/rapidsai/cudf/pull/8809)) [@isVoid](https://github.com/isVoid) -- Add quarters to libcudf datetime ([#8779](https://github.com/rapidsai/cudf/pull/8779)) [@shaneding](https://github.com/shaneding) -- Linear Interpolation of `nan`s via `cupy` ([#8767](https://github.com/rapidsai/cudf/pull/8767)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Enable compiled binary ops in libcudf, python and java ([#8741](https://github.com/rapidsai/cudf/pull/8741)) [@karthikeyann](https://github.com/karthikeyann) -- Make groupby transform-like op order match original data order ([#8720](https://github.com/rapidsai/cudf/pull/8720)) [@isVoid](https://github.com/isVoid) -- multibyte_split ([#8702](https://github.com/rapidsai/cudf/pull/8702)) [@cwharris](https://github.com/cwharris) -- Implement JNI for `strings:repeat_strings` that repeats each string separately by different numbers of times ([#8572](https://github.com/rapidsai/cudf/pull/8572)) [@ttnghia](https://github.com/ttnghia) +- Revert "Add shallow hash function and shallow equality comparison for column_view ([#9185)" (#9283](https://github.com/NVIDIA/cudf/pull/9185)" (#9283)) [@karthikeyann](https://github.com/karthikeyann) +- Align `DataFrame.apply` signature with pandas ([#9275](https://github.com/NVIDIA/cudf/pull/9275)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add struct type support for `drop_list_duplicates` ([#9202](https://github.com/NVIDIA/cudf/pull/9202)) [@ttnghia](https://github.com/ttnghia) +- support CUDA async memory resource in JNI ([#9201](https://github.com/NVIDIA/cudf/pull/9201)) [@rongou](https://github.com/rongou) +- Add shallow hash function and shallow equality comparison for column_view ([#9185](https://github.com/NVIDIA/cudf/pull/9185)) [@karthikeyann](https://github.com/karthikeyann) +- Superimpose null masks for STRUCT columns. ([#9144](https://github.com/NVIDIA/cudf/pull/9144)) [@mythrocks](https://github.com/mythrocks) +- Implemented bindings for `ceil` timestamp operation ([#9141](https://github.com/NVIDIA/cudf/pull/9141)) [@shaneding](https://github.com/shaneding) +- Adding MAP type support for ORC Reader ([#9132](https://github.com/NVIDIA/cudf/pull/9132)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Implement `interleave_columns` for lists with arbitrary nested type ([#9130](https://github.com/NVIDIA/cudf/pull/9130)) [@ttnghia](https://github.com/ttnghia) +- Add python bindings to fixed-size window and groupby `rolling.var`, `rolling.std` ([#9097](https://github.com/NVIDIA/cudf/pull/9097)) [@isVoid](https://github.com/isVoid) +- Make AST operators nullable ([#9096](https://github.com/NVIDIA/cudf/pull/9096)) [@vyasr](https://github.com/vyasr) +- Java bindings for approx_percentile ([#9094](https://github.com/NVIDIA/cudf/pull/9094)) [@andygrove](https://github.com/andygrove) +- Add `dseries.struct.explode` ([#9086](https://github.com/NVIDIA/cudf/pull/9086)) [@isVoid](https://github.com/isVoid) +- Add support for BaseIndexer in Rolling APIs ([#9085](https://github.com/NVIDIA/cudf/pull/9085)) [@galipremsagar](https://github.com/galipremsagar) +- Remove the option to pass data types as strings to `read_csv` and `read_json` ([#9079](https://github.com/NVIDIA/cudf/pull/9079)) [@vuule](https://github.com/vuule) +- Add handling for nested dicts in dask-cudf groupby ([#9054](https://github.com/NVIDIA/cudf/pull/9054)) [@charlesbluca](https://github.com/charlesbluca) +- Added Series.dt.is_quarter_start and Series.dt.is_quarter_end ([#9046](https://github.com/NVIDIA/cudf/pull/9046)) [@TravisHester](https://github.com/TravisHester) +- Support nested types for nth_element reduction ([#9043](https://github.com/NVIDIA/cudf/pull/9043)) [@sperlingxx](https://github.com/sperlingxx) +- Update sort groupby to use non-atomic operation ([#9035](https://github.com/NVIDIA/cudf/pull/9035)) [@karthikeyann](https://github.com/karthikeyann) +- Add support for struct type in ORC writer ([#9025](https://github.com/NVIDIA/cudf/pull/9025)) [@vuule](https://github.com/vuule) +- Implement `interleave_columns` for structs columns ([#9012](https://github.com/NVIDIA/cudf/pull/9012)) [@ttnghia](https://github.com/ttnghia) +- Add groupby first and last aggregations ([#9004](https://github.com/NVIDIA/cudf/pull/9004)) [@shwina](https://github.com/shwina) +- Add `DecimalBaseColumn` and move `as_decimal_column` ([#9001](https://github.com/NVIDIA/cudf/pull/9001)) [@isVoid](https://github.com/isVoid) +- Python/Cython bindings for multibyte_split ([#8998](https://github.com/NVIDIA/cudf/pull/8998)) [@jdye64](https://github.com/jdye64) +- Support scalar `months` in `add_calendrical_months`, extends API to INT32 support ([#8991](https://github.com/NVIDIA/cudf/pull/8991)) [@isVoid](https://github.com/isVoid) +- Added Series.dt.is_month_end ([#8989](https://github.com/NVIDIA/cudf/pull/8989)) [@TravisHester](https://github.com/TravisHester) +- Support for using tdigests to compute approximate percentiles. ([#8983](https://github.com/NVIDIA/cudf/pull/8983)) [@nvdbaranec](https://github.com/nvdbaranec) +- Support "unflatten" of columns flattened via `flatten_nested_columns()`: ([#8956](https://github.com/NVIDIA/cudf/pull/8956)) [@mythrocks](https://github.com/mythrocks) +- Implement timestamp ceil ([#8942](https://github.com/NVIDIA/cudf/pull/8942)) [@shaneding](https://github.com/shaneding) +- Add nested column selection to parquet reader ([#8933](https://github.com/NVIDIA/cudf/pull/8933)) [@devavret](https://github.com/devavret) +- Expose conditional join size calculation ([#8928](https://github.com/NVIDIA/cudf/pull/8928)) [@vyasr](https://github.com/vyasr) +- Support Nulls in Timeseries Generator ([#8925](https://github.com/NVIDIA/cudf/pull/8925)) [@isVoid](https://github.com/isVoid) +- Avoid index equality check in `_CPackedColumns.from_py_table()` ([#8917](https://github.com/NVIDIA/cudf/pull/8917)) [@charlesbluca](https://github.com/charlesbluca) +- Add dot product binary op ([#8909](https://github.com/NVIDIA/cudf/pull/8909)) [@charlesbluca](https://github.com/charlesbluca) +- Expose `days_in_month` function in libcudf and add python bindings ([#8892](https://github.com/NVIDIA/cudf/pull/8892)) [@isVoid](https://github.com/isVoid) +- Series string repeat ([#8882](https://github.com/NVIDIA/cudf/pull/8882)) [@sarahyurick](https://github.com/sarahyurick) +- Python binding for quarters ([#8862](https://github.com/NVIDIA/cudf/pull/8862)) [@shaneding](https://github.com/shaneding) +- Expand CSV and JSON reader APIs to accept `dtypes` as a vector or map of `data_type` objects ([#8856](https://github.com/NVIDIA/cudf/pull/8856)) [@vuule](https://github.com/vuule) +- Add Java bindings for AST transform ([#8846](https://github.com/NVIDIA/cudf/pull/8846)) [@jlowe](https://github.com/jlowe) +- Series datetime is_month_start ([#8844](https://github.com/NVIDIA/cudf/pull/8844)) [@sarahyurick](https://github.com/sarahyurick) +- Support bracket syntax for cudf::strings::replace_with_backrefs group index values ([#8841](https://github.com/NVIDIA/cudf/pull/8841)) [@davidwendt](https://github.com/davidwendt) +- Support `VARIANCE` and `STD` aggregation in rolling op ([#8809](https://github.com/NVIDIA/cudf/pull/8809)) [@isVoid](https://github.com/isVoid) +- Add quarters to libcudf datetime ([#8779](https://github.com/NVIDIA/cudf/pull/8779)) [@shaneding](https://github.com/shaneding) +- Linear Interpolation of `nan`s via `cupy` ([#8767](https://github.com/NVIDIA/cudf/pull/8767)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Enable compiled binary ops in libcudf, python and java ([#8741](https://github.com/NVIDIA/cudf/pull/8741)) [@karthikeyann](https://github.com/karthikeyann) +- Make groupby transform-like op order match original data order ([#8720](https://github.com/NVIDIA/cudf/pull/8720)) [@isVoid](https://github.com/isVoid) +- multibyte_split ([#8702](https://github.com/NVIDIA/cudf/pull/8702)) [@cwharris](https://github.com/cwharris) +- Implement JNI for `strings:repeat_strings` that repeats each string separately by different numbers of times ([#8572](https://github.com/NVIDIA/cudf/pull/8572)) [@ttnghia](https://github.com/ttnghia) ## 🛠️ Improvements -- Pin max `dask` and `distributed` versions to `2021.09.1` ([#9286](https://github.com/rapidsai/cudf/pull/9286)) [@galipremsagar](https://github.com/galipremsagar) -- Optimized fsspec data transfer for remote file-systems ([#9265](https://github.com/rapidsai/cudf/pull/9265)) [@rjzamora](https://github.com/rjzamora) -- Skip dask-cudf tests on arm64 ([#9252](https://github.com/rapidsai/cudf/pull/9252)) [@Ethyling](https://github.com/Ethyling) -- Use nvcomp's snappy compressor in ORC writer ([#9242](https://github.com/rapidsai/cudf/pull/9242)) [@devavret](https://github.com/devavret) -- Only run imports tests on x86_64 ([#9241](https://github.com/rapidsai/cudf/pull/9241)) [@Ethyling](https://github.com/Ethyling) -- Remove unnecessary call to device_uvector::release() ([#9237](https://github.com/rapidsai/cudf/pull/9237)) [@harrism](https://github.com/harrism) -- Use nvcomp's snappy decompression in ORC reader ([#9235](https://github.com/rapidsai/cudf/pull/9235)) [@devavret](https://github.com/devavret) -- Add grouped_rolling test with STRUCT groupby keys. ([#9228](https://github.com/rapidsai/cudf/pull/9228)) [@mythrocks](https://github.com/mythrocks) -- Optimize `cudf.concat` for `axis=0` ([#9222](https://github.com/rapidsai/cudf/pull/9222)) [@galipremsagar](https://github.com/galipremsagar) -- Fix some libcudf calls not passing the stream parameter ([#9220](https://github.com/rapidsai/cudf/pull/9220)) [@davidwendt](https://github.com/davidwendt) -- Add min and max bounds for random dataframe generator numeric types ([#9211](https://github.com/rapidsai/cudf/pull/9211)) [@galipremsagar](https://github.com/galipremsagar) -- Improve performance of expression evaluation ([#9210](https://github.com/rapidsai/cudf/pull/9210)) [@vyasr](https://github.com/vyasr) -- Misc optimizations in `cudf` ([#9203](https://github.com/rapidsai/cudf/pull/9203)) [@galipremsagar](https://github.com/galipremsagar) -- Remove Cython APIs for table view generation ([#9199](https://github.com/rapidsai/cudf/pull/9199)) [@vyasr](https://github.com/vyasr) -- Add JNI support for drop_list_duplicates ([#9198](https://github.com/rapidsai/cudf/pull/9198)) [@revans2](https://github.com/revans2) -- Update pandas versions in conda recipes and requirements.txt files ([#9197](https://github.com/rapidsai/cudf/pull/9197)) [@galipremsagar](https://github.com/galipremsagar) -- Minor C++17 cleanup of `groupby.cu`: structured bindings, more concise lambda, etc ([#9193](https://github.com/rapidsai/cudf/pull/9193)) [@codereport](https://github.com/codereport) -- Explicit about bitwidth difference between cudf boolean and arrow boolean ([#9192](https://github.com/rapidsai/cudf/pull/9192)) [@isVoid](https://github.com/isVoid) -- Remove _source_index from MultiIndex ([#9191](https://github.com/rapidsai/cudf/pull/9191)) [@vyasr](https://github.com/vyasr) -- Fix typo in the name of `cudf-testing-targets.cmake` ([#9190](https://github.com/rapidsai/cudf/pull/9190)) [@trxcllnt](https://github.com/trxcllnt) -- Add support for single-digits in cudf::to_timestamps ([#9173](https://github.com/rapidsai/cudf/pull/9173)) [@davidwendt](https://github.com/davidwendt) -- Fix cufilejni build include path ([#9168](https://github.com/rapidsai/cudf/pull/9168)) [@pxLi](https://github.com/pxLi) -- `dask_cudf` dispatch registering cleanup ([#9160](https://github.com/rapidsai/cudf/pull/9160)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unneeded stream/mr from a cudf::make_strings_column ([#9148](https://github.com/rapidsai/cudf/pull/9148)) [@davidwendt](https://github.com/davidwendt) -- Upgrade `pandas` version in `cudf` ([#9147](https://github.com/rapidsai/cudf/pull/9147)) [@galipremsagar](https://github.com/galipremsagar) -- make data chunk reader return unique_ptr ([#9129](https://github.com/rapidsai/cudf/pull/9129)) [@cwharris](https://github.com/cwharris) -- Add backend for `percentile_lookup` dispatch ([#9118](https://github.com/rapidsai/cudf/pull/9118)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor implementation of column setitem ([#9110](https://github.com/rapidsai/cudf/pull/9110)) [@vyasr](https://github.com/vyasr) -- Fix compile warnings found using nvcc 11.4 ([#9101](https://github.com/rapidsai/cudf/pull/9101)) [@davidwendt](https://github.com/davidwendt) -- Update to UCX-Py 0.22 ([#9099](https://github.com/rapidsai/cudf/pull/9099)) [@pentschev](https://github.com/pentschev) -- Simplify read_avro by removing unnecessary writer/impl classes ([#9090](https://github.com/rapidsai/cudf/pull/9090)) [@cwharris](https://github.com/cwharris) -- Allowing %f in format to return nanoseconds ([#9081](https://github.com/rapidsai/cudf/pull/9081)) [@marlenezw](https://github.com/marlenezw) -- Java bindings for cudf::hash_join ([#9080](https://github.com/rapidsai/cudf/pull/9080)) [@jlowe](https://github.com/jlowe) -- Remove stale code in `ColumnBase._fill` ([#9078](https://github.com/rapidsai/cudf/pull/9078)) [@isVoid](https://github.com/isVoid) -- Add support for `get_group` in GroupBy ([#9070](https://github.com/rapidsai/cudf/pull/9070)) [@galipremsagar](https://github.com/galipremsagar) -- Remove remaining "support" methods from DataFrame ([#9068](https://github.com/rapidsai/cudf/pull/9068)) [@vyasr](https://github.com/vyasr) -- Update JNI java CSV APIs to not use deprecated API ([#9066](https://github.com/rapidsai/cudf/pull/9066)) [@revans2](https://github.com/revans2) -- Added method to remove null_masks if the column has no nulls ([#9061](https://github.com/rapidsai/cudf/pull/9061)) [@razajafri](https://github.com/razajafri) -- Consolidate Several Series and Dataframe Methods ([#9059](https://github.com/rapidsai/cudf/pull/9059)) [@isVoid](https://github.com/isVoid) -- Remove usage of string based `set_dtypes` for `csv` & `json` readers ([#9049](https://github.com/rapidsai/cudf/pull/9049)) [@galipremsagar](https://github.com/galipremsagar) -- Remove some debug print statements from gtests ([#9048](https://github.com/rapidsai/cudf/pull/9048)) [@davidwendt](https://github.com/davidwendt) -- Support additional format specifiers in from_timestamps ([#9047](https://github.com/rapidsai/cudf/pull/9047)) [@davidwendt](https://github.com/davidwendt) -- Expose expression base class publicly and simplify public AST API ([#9045](https://github.com/rapidsai/cudf/pull/9045)) [@vyasr](https://github.com/vyasr) -- move filepath and mmap logic out of json/csv up to functions.cpp ([#9040](https://github.com/rapidsai/cudf/pull/9040)) [@cwharris](https://github.com/cwharris) -- Refactor Index hierarchy ([#9039](https://github.com/rapidsai/cudf/pull/9039)) [@vyasr](https://github.com/vyasr) -- cudf now leverages rapids-cmake to reduce CMake boilerplate ([#9030](https://github.com/rapidsai/cudf/pull/9030)) [@robertmaynard](https://github.com/robertmaynard) -- Add support for `STRUCT` input to `groupby` ([#9024](https://github.com/rapidsai/cudf/pull/9024)) [@mythrocks](https://github.com/mythrocks) -- Refactor Frame scans ([#9021](https://github.com/rapidsai/cudf/pull/9021)) [@vyasr](https://github.com/vyasr) -- Remove duplicate `set_categories` code ([#9018](https://github.com/rapidsai/cudf/pull/9018)) [@isVoid](https://github.com/isVoid) -- Map support for ParquetWriter ([#9013](https://github.com/rapidsai/cudf/pull/9013)) [@razajafri](https://github.com/razajafri) -- Remove aliases of various api.types APIs from utils.dtypes. ([#9011](https://github.com/rapidsai/cudf/pull/9011)) [@vyasr](https://github.com/vyasr) -- Java bindings for conditional join output sizes ([#9002](https://github.com/rapidsai/cudf/pull/9002)) [@jlowe](https://github.com/jlowe) -- Remove _copy_construct factory ([#8999](https://github.com/rapidsai/cudf/pull/8999)) [@vyasr](https://github.com/vyasr) -- ENH Allow arbitrary CMake config options in build.sh ([#8996](https://github.com/rapidsai/cudf/pull/8996)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- A small optimization for JNI copy column view to column vector ([#8985](https://github.com/rapidsai/cudf/pull/8985)) [@revans2](https://github.com/revans2) -- Fix nvcc warnings in ORC writer ([#8975](https://github.com/rapidsai/cudf/pull/8975)) [@devavret](https://github.com/devavret) -- Support nested structs in rank and dense rank ([#8962](https://github.com/rapidsai/cudf/pull/8962)) [@rwlee](https://github.com/rwlee) -- Move compute_column API out of ast namespace ([#8957](https://github.com/rapidsai/cudf/pull/8957)) [@vyasr](https://github.com/vyasr) -- Series datetime is_year_end and is_year_start ([#8954](https://github.com/rapidsai/cudf/pull/8954)) [@marlenezw](https://github.com/marlenezw) -- Make Java AstNode public ([#8953](https://github.com/rapidsai/cudf/pull/8953)) [@jlowe](https://github.com/jlowe) -- Replace allocate with device_uvector for subword_tokenize internal tables ([#8952](https://github.com/rapidsai/cudf/pull/8952)) [@davidwendt](https://github.com/davidwendt) -- `cudf.dtype` function ([#8949](https://github.com/rapidsai/cudf/pull/8949)) [@shwina](https://github.com/shwina) -- Refactor Frame reductions ([#8944](https://github.com/rapidsai/cudf/pull/8944)) [@vyasr](https://github.com/vyasr) -- Add deprecation warning for `Series.set_mask` API ([#8943](https://github.com/rapidsai/cudf/pull/8943)) [@galipremsagar](https://github.com/galipremsagar) -- Move AST evaluator into a separate header ([#8930](https://github.com/rapidsai/cudf/pull/8930)) [@vyasr](https://github.com/vyasr) -- JNI Aggregation Type Changes ([#8919](https://github.com/rapidsai/cudf/pull/8919)) [@revans2](https://github.com/revans2) -- Move template parameter to function parameter in cudf::detail::left_semi_anti_join ([#8914](https://github.com/rapidsai/cudf/pull/8914)) [@davidwendt](https://github.com/davidwendt) -- Upgrade `arrow` & `pyarrow` to `5.0.0` ([#8908](https://github.com/rapidsai/cudf/pull/8908)) [@galipremsagar](https://github.com/galipremsagar) -- Add groupby_aggregation and groupby_scan_aggregation classes and force their usage. ([#8906](https://github.com/rapidsai/cudf/pull/8906)) [@nvdbaranec](https://github.com/nvdbaranec) -- Move `structs_column_tests.cu` to `.cpp`. ([#8902](https://github.com/rapidsai/cudf/pull/8902)) [@mythrocks](https://github.com/mythrocks) -- Add stream and memory-resource parameters to struct-scalar copy ctor ([#8901](https://github.com/rapidsai/cudf/pull/8901)) [@davidwendt](https://github.com/davidwendt) -- Combine linearizer and ast_plan ([#8900](https://github.com/rapidsai/cudf/pull/8900)) [@vyasr](https://github.com/vyasr) -- Add Java bindings for conditional join gather maps ([#8888](https://github.com/rapidsai/cudf/pull/8888)) [@jlowe](https://github.com/jlowe) -- Remove max version pin for `dask` & `distributed` on development branch ([#8881](https://github.com/rapidsai/cudf/pull/8881)) [@galipremsagar](https://github.com/galipremsagar) -- fix cufilejni build w/ c++17 ([#8877](https://github.com/rapidsai/cudf/pull/8877)) [@pxLi](https://github.com/pxLi) -- Add struct accessor to dask-cudf ([#8874](https://github.com/rapidsai/cudf/pull/8874)) [@NV-jpt](https://github.com/NV-jpt) -- Migrate dask-cudf CudfEngine to leverage ArrowDatasetEngine ([#8871](https://github.com/rapidsai/cudf/pull/8871)) [@rjzamora](https://github.com/rjzamora) -- Add JNI for extract_quarter, add_calendrical_months, and is_leap_year ([#8863](https://github.com/rapidsai/cudf/pull/8863)) [@revans2](https://github.com/revans2) -- Change cudf::scalar copy and move constructors to protected ([#8857](https://github.com/rapidsai/cudf/pull/8857)) [@davidwendt](https://github.com/davidwendt) -- Replace `is_same<>::value` with `is_same_v<>` ([#8852](https://github.com/rapidsai/cudf/pull/8852)) [@codereport](https://github.com/codereport) -- Add min `pytorch` version to `importorskip` in pytest ([#8851](https://github.com/rapidsai/cudf/pull/8851)) [@galipremsagar](https://github.com/galipremsagar) -- Java bindings for regex replace ([#8847](https://github.com/rapidsai/cudf/pull/8847)) [@jlowe](https://github.com/jlowe) -- Remove make strings children with null mask ([#8830](https://github.com/rapidsai/cudf/pull/8830)) [@davidwendt](https://github.com/davidwendt) -- Refactor conditional joins ([#8815](https://github.com/rapidsai/cudf/pull/8815)) [@vyasr](https://github.com/vyasr) -- Small cleanup (unused headers / commented code removals) ([#8799](https://github.com/rapidsai/cudf/pull/8799)) [@codereport](https://github.com/codereport) -- ENH Replace gpuci_conda_retry with gpuci_mamba_retry ([#8770](https://github.com/rapidsai/cudf/pull/8770)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- Update cudf java bindings to 21.10.0-SNAPSHOT ([#8765](https://github.com/rapidsai/cudf/pull/8765)) [@pxLi](https://github.com/pxLi) -- Refactor and improve join benchmarks with nvbench ([#8734](https://github.com/rapidsai/cudf/pull/8734)) [@PointKernel](https://github.com/PointKernel) -- Refactor Python factories and remove usage of Table for libcudf output handling ([#8687](https://github.com/rapidsai/cudf/pull/8687)) [@vyasr](https://github.com/vyasr) -- Optimize URL Decoding ([#8622](https://github.com/rapidsai/cudf/pull/8622)) [@gaohao95](https://github.com/gaohao95) -- Parquet writer dictionary encoding refactor ([#8476](https://github.com/rapidsai/cudf/pull/8476)) [@devavret](https://github.com/devavret) -- Use nvcomp's snappy decompression in parquet reader ([#8252](https://github.com/rapidsai/cudf/pull/8252)) [@devavret](https://github.com/devavret) -- Use nvcomp's snappy compressor in parquet writer ([#8229](https://github.com/rapidsai/cudf/pull/8229)) [@devavret](https://github.com/devavret) +- Pin max `dask` and `distributed` versions to `2021.09.1` ([#9286](https://github.com/NVIDIA/cudf/pull/9286)) [@galipremsagar](https://github.com/galipremsagar) +- Optimized fsspec data transfer for remote file-systems ([#9265](https://github.com/NVIDIA/cudf/pull/9265)) [@rjzamora](https://github.com/rjzamora) +- Skip dask-cudf tests on arm64 ([#9252](https://github.com/NVIDIA/cudf/pull/9252)) [@Ethyling](https://github.com/Ethyling) +- Use nvcomp's snappy compressor in ORC writer ([#9242](https://github.com/NVIDIA/cudf/pull/9242)) [@devavret](https://github.com/devavret) +- Only run imports tests on x86_64 ([#9241](https://github.com/NVIDIA/cudf/pull/9241)) [@Ethyling](https://github.com/Ethyling) +- Remove unnecessary call to device_uvector::release() ([#9237](https://github.com/NVIDIA/cudf/pull/9237)) [@harrism](https://github.com/harrism) +- Use nvcomp's snappy decompression in ORC reader ([#9235](https://github.com/NVIDIA/cudf/pull/9235)) [@devavret](https://github.com/devavret) +- Add grouped_rolling test with STRUCT groupby keys. ([#9228](https://github.com/NVIDIA/cudf/pull/9228)) [@mythrocks](https://github.com/mythrocks) +- Optimize `cudf.concat` for `axis=0` ([#9222](https://github.com/NVIDIA/cudf/pull/9222)) [@galipremsagar](https://github.com/galipremsagar) +- Fix some libcudf calls not passing the stream parameter ([#9220](https://github.com/NVIDIA/cudf/pull/9220)) [@davidwendt](https://github.com/davidwendt) +- Add min and max bounds for random dataframe generator numeric types ([#9211](https://github.com/NVIDIA/cudf/pull/9211)) [@galipremsagar](https://github.com/galipremsagar) +- Improve performance of expression evaluation ([#9210](https://github.com/NVIDIA/cudf/pull/9210)) [@vyasr](https://github.com/vyasr) +- Misc optimizations in `cudf` ([#9203](https://github.com/NVIDIA/cudf/pull/9203)) [@galipremsagar](https://github.com/galipremsagar) +- Remove Cython APIs for table view generation ([#9199](https://github.com/NVIDIA/cudf/pull/9199)) [@vyasr](https://github.com/vyasr) +- Add JNI support for drop_list_duplicates ([#9198](https://github.com/NVIDIA/cudf/pull/9198)) [@revans2](https://github.com/revans2) +- Update pandas versions in conda recipes and requirements.txt files ([#9197](https://github.com/NVIDIA/cudf/pull/9197)) [@galipremsagar](https://github.com/galipremsagar) +- Minor C++17 cleanup of `groupby.cu`: structured bindings, more concise lambda, etc ([#9193](https://github.com/NVIDIA/cudf/pull/9193)) [@codereport](https://github.com/codereport) +- Explicit about bitwidth difference between cudf boolean and arrow boolean ([#9192](https://github.com/NVIDIA/cudf/pull/9192)) [@isVoid](https://github.com/isVoid) +- Remove _source_index from MultiIndex ([#9191](https://github.com/NVIDIA/cudf/pull/9191)) [@vyasr](https://github.com/vyasr) +- Fix typo in the name of `cudf-testing-targets.cmake` ([#9190](https://github.com/NVIDIA/cudf/pull/9190)) [@trxcllnt](https://github.com/trxcllnt) +- Add support for single-digits in cudf::to_timestamps ([#9173](https://github.com/NVIDIA/cudf/pull/9173)) [@davidwendt](https://github.com/davidwendt) +- Fix cufilejni build include path ([#9168](https://github.com/NVIDIA/cudf/pull/9168)) [@pxLi](https://github.com/pxLi) +- `dask_cudf` dispatch registering cleanup ([#9160](https://github.com/NVIDIA/cudf/pull/9160)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unneeded stream/mr from a cudf::make_strings_column ([#9148](https://github.com/NVIDIA/cudf/pull/9148)) [@davidwendt](https://github.com/davidwendt) +- Upgrade `pandas` version in `cudf` ([#9147](https://github.com/NVIDIA/cudf/pull/9147)) [@galipremsagar](https://github.com/galipremsagar) +- make data chunk reader return unique_ptr ([#9129](https://github.com/NVIDIA/cudf/pull/9129)) [@cwharris](https://github.com/cwharris) +- Add backend for `percentile_lookup` dispatch ([#9118](https://github.com/NVIDIA/cudf/pull/9118)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor implementation of column setitem ([#9110](https://github.com/NVIDIA/cudf/pull/9110)) [@vyasr](https://github.com/vyasr) +- Fix compile warnings found using nvcc 11.4 ([#9101](https://github.com/NVIDIA/cudf/pull/9101)) [@davidwendt](https://github.com/davidwendt) +- Update to UCX-Py 0.22 ([#9099](https://github.com/NVIDIA/cudf/pull/9099)) [@pentschev](https://github.com/pentschev) +- Simplify read_avro by removing unnecessary writer/impl classes ([#9090](https://github.com/NVIDIA/cudf/pull/9090)) [@cwharris](https://github.com/cwharris) +- Allowing %f in format to return nanoseconds ([#9081](https://github.com/NVIDIA/cudf/pull/9081)) [@marlenezw](https://github.com/marlenezw) +- Java bindings for cudf::hash_join ([#9080](https://github.com/NVIDIA/cudf/pull/9080)) [@jlowe](https://github.com/jlowe) +- Remove stale code in `ColumnBase._fill` ([#9078](https://github.com/NVIDIA/cudf/pull/9078)) [@isVoid](https://github.com/isVoid) +- Add support for `get_group` in GroupBy ([#9070](https://github.com/NVIDIA/cudf/pull/9070)) [@galipremsagar](https://github.com/galipremsagar) +- Remove remaining "support" methods from DataFrame ([#9068](https://github.com/NVIDIA/cudf/pull/9068)) [@vyasr](https://github.com/vyasr) +- Update JNI java CSV APIs to not use deprecated API ([#9066](https://github.com/NVIDIA/cudf/pull/9066)) [@revans2](https://github.com/revans2) +- Added method to remove null_masks if the column has no nulls ([#9061](https://github.com/NVIDIA/cudf/pull/9061)) [@razajafri](https://github.com/razajafri) +- Consolidate Several Series and Dataframe Methods ([#9059](https://github.com/NVIDIA/cudf/pull/9059)) [@isVoid](https://github.com/isVoid) +- Remove usage of string based `set_dtypes` for `csv` & `json` readers ([#9049](https://github.com/NVIDIA/cudf/pull/9049)) [@galipremsagar](https://github.com/galipremsagar) +- Remove some debug print statements from gtests ([#9048](https://github.com/NVIDIA/cudf/pull/9048)) [@davidwendt](https://github.com/davidwendt) +- Support additional format specifiers in from_timestamps ([#9047](https://github.com/NVIDIA/cudf/pull/9047)) [@davidwendt](https://github.com/davidwendt) +- Expose expression base class publicly and simplify public AST API ([#9045](https://github.com/NVIDIA/cudf/pull/9045)) [@vyasr](https://github.com/vyasr) +- move filepath and mmap logic out of json/csv up to functions.cpp ([#9040](https://github.com/NVIDIA/cudf/pull/9040)) [@cwharris](https://github.com/cwharris) +- Refactor Index hierarchy ([#9039](https://github.com/NVIDIA/cudf/pull/9039)) [@vyasr](https://github.com/vyasr) +- cudf now leverages rapids-cmake to reduce CMake boilerplate ([#9030](https://github.com/NVIDIA/cudf/pull/9030)) [@robertmaynard](https://github.com/robertmaynard) +- Add support for `STRUCT` input to `groupby` ([#9024](https://github.com/NVIDIA/cudf/pull/9024)) [@mythrocks](https://github.com/mythrocks) +- Refactor Frame scans ([#9021](https://github.com/NVIDIA/cudf/pull/9021)) [@vyasr](https://github.com/vyasr) +- Remove duplicate `set_categories` code ([#9018](https://github.com/NVIDIA/cudf/pull/9018)) [@isVoid](https://github.com/isVoid) +- Map support for ParquetWriter ([#9013](https://github.com/NVIDIA/cudf/pull/9013)) [@razajafri](https://github.com/razajafri) +- Remove aliases of various api.types APIs from utils.dtypes. ([#9011](https://github.com/NVIDIA/cudf/pull/9011)) [@vyasr](https://github.com/vyasr) +- Java bindings for conditional join output sizes ([#9002](https://github.com/NVIDIA/cudf/pull/9002)) [@jlowe](https://github.com/jlowe) +- Remove _copy_construct factory ([#8999](https://github.com/NVIDIA/cudf/pull/8999)) [@vyasr](https://github.com/vyasr) +- ENH Allow arbitrary CMake config options in build.sh ([#8996](https://github.com/NVIDIA/cudf/pull/8996)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- A small optimization for JNI copy column view to column vector ([#8985](https://github.com/NVIDIA/cudf/pull/8985)) [@revans2](https://github.com/revans2) +- Fix nvcc warnings in ORC writer ([#8975](https://github.com/NVIDIA/cudf/pull/8975)) [@devavret](https://github.com/devavret) +- Support nested structs in rank and dense rank ([#8962](https://github.com/NVIDIA/cudf/pull/8962)) [@rwlee](https://github.com/rwlee) +- Move compute_column API out of ast namespace ([#8957](https://github.com/NVIDIA/cudf/pull/8957)) [@vyasr](https://github.com/vyasr) +- Series datetime is_year_end and is_year_start ([#8954](https://github.com/NVIDIA/cudf/pull/8954)) [@marlenezw](https://github.com/marlenezw) +- Make Java AstNode public ([#8953](https://github.com/NVIDIA/cudf/pull/8953)) [@jlowe](https://github.com/jlowe) +- Replace allocate with device_uvector for subword_tokenize internal tables ([#8952](https://github.com/NVIDIA/cudf/pull/8952)) [@davidwendt](https://github.com/davidwendt) +- `cudf.dtype` function ([#8949](https://github.com/NVIDIA/cudf/pull/8949)) [@shwina](https://github.com/shwina) +- Refactor Frame reductions ([#8944](https://github.com/NVIDIA/cudf/pull/8944)) [@vyasr](https://github.com/vyasr) +- Add deprecation warning for `Series.set_mask` API ([#8943](https://github.com/NVIDIA/cudf/pull/8943)) [@galipremsagar](https://github.com/galipremsagar) +- Move AST evaluator into a separate header ([#8930](https://github.com/NVIDIA/cudf/pull/8930)) [@vyasr](https://github.com/vyasr) +- JNI Aggregation Type Changes ([#8919](https://github.com/NVIDIA/cudf/pull/8919)) [@revans2](https://github.com/revans2) +- Move template parameter to function parameter in cudf::detail::left_semi_anti_join ([#8914](https://github.com/NVIDIA/cudf/pull/8914)) [@davidwendt](https://github.com/davidwendt) +- Upgrade `arrow` & `pyarrow` to `5.0.0` ([#8908](https://github.com/NVIDIA/cudf/pull/8908)) [@galipremsagar](https://github.com/galipremsagar) +- Add groupby_aggregation and groupby_scan_aggregation classes and force their usage. ([#8906](https://github.com/NVIDIA/cudf/pull/8906)) [@nvdbaranec](https://github.com/nvdbaranec) +- Move `structs_column_tests.cu` to `.cpp`. ([#8902](https://github.com/NVIDIA/cudf/pull/8902)) [@mythrocks](https://github.com/mythrocks) +- Add stream and memory-resource parameters to struct-scalar copy ctor ([#8901](https://github.com/NVIDIA/cudf/pull/8901)) [@davidwendt](https://github.com/davidwendt) +- Combine linearizer and ast_plan ([#8900](https://github.com/NVIDIA/cudf/pull/8900)) [@vyasr](https://github.com/vyasr) +- Add Java bindings for conditional join gather maps ([#8888](https://github.com/NVIDIA/cudf/pull/8888)) [@jlowe](https://github.com/jlowe) +- Remove max version pin for `dask` & `distributed` on development branch ([#8881](https://github.com/NVIDIA/cudf/pull/8881)) [@galipremsagar](https://github.com/galipremsagar) +- fix cufilejni build w/ c++17 ([#8877](https://github.com/NVIDIA/cudf/pull/8877)) [@pxLi](https://github.com/pxLi) +- Add struct accessor to dask-cudf ([#8874](https://github.com/NVIDIA/cudf/pull/8874)) [@NV-jpt](https://github.com/NV-jpt) +- Migrate dask-cudf CudfEngine to leverage ArrowDatasetEngine ([#8871](https://github.com/NVIDIA/cudf/pull/8871)) [@rjzamora](https://github.com/rjzamora) +- Add JNI for extract_quarter, add_calendrical_months, and is_leap_year ([#8863](https://github.com/NVIDIA/cudf/pull/8863)) [@revans2](https://github.com/revans2) +- Change cudf::scalar copy and move constructors to protected ([#8857](https://github.com/NVIDIA/cudf/pull/8857)) [@davidwendt](https://github.com/davidwendt) +- Replace `is_same<>::value` with `is_same_v<>` ([#8852](https://github.com/NVIDIA/cudf/pull/8852)) [@codereport](https://github.com/codereport) +- Add min `pytorch` version to `importorskip` in pytest ([#8851](https://github.com/NVIDIA/cudf/pull/8851)) [@galipremsagar](https://github.com/galipremsagar) +- Java bindings for regex replace ([#8847](https://github.com/NVIDIA/cudf/pull/8847)) [@jlowe](https://github.com/jlowe) +- Remove make strings children with null mask ([#8830](https://github.com/NVIDIA/cudf/pull/8830)) [@davidwendt](https://github.com/davidwendt) +- Refactor conditional joins ([#8815](https://github.com/NVIDIA/cudf/pull/8815)) [@vyasr](https://github.com/vyasr) +- Small cleanup (unused headers / commented code removals) ([#8799](https://github.com/NVIDIA/cudf/pull/8799)) [@codereport](https://github.com/codereport) +- ENH Replace gpuci_conda_retry with gpuci_mamba_retry ([#8770](https://github.com/NVIDIA/cudf/pull/8770)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- Update cudf java bindings to 21.10.0-SNAPSHOT ([#8765](https://github.com/NVIDIA/cudf/pull/8765)) [@pxLi](https://github.com/pxLi) +- Refactor and improve join benchmarks with nvbench ([#8734](https://github.com/NVIDIA/cudf/pull/8734)) [@PointKernel](https://github.com/PointKernel) +- Refactor Python factories and remove usage of Table for libcudf output handling ([#8687](https://github.com/NVIDIA/cudf/pull/8687)) [@vyasr](https://github.com/vyasr) +- Optimize URL Decoding ([#8622](https://github.com/NVIDIA/cudf/pull/8622)) [@gaohao95](https://github.com/gaohao95) +- Parquet writer dictionary encoding refactor ([#8476](https://github.com/NVIDIA/cudf/pull/8476)) [@devavret](https://github.com/devavret) +- Use nvcomp's snappy decompression in parquet reader ([#8252](https://github.com/NVIDIA/cudf/pull/8252)) [@devavret](https://github.com/devavret) +- Use nvcomp's snappy compressor in parquet writer ([#8229](https://github.com/NVIDIA/cudf/pull/8229)) [@devavret](https://github.com/devavret) # cuDF 21.08.00 (4 Aug 2021) ## 🚨 Breaking Changes -- Fix a crash in pack() when being handed tables with no columns. ([#8697](https://github.com/rapidsai/cudf/pull/8697)) [@nvdbaranec](https://github.com/nvdbaranec) -- Remove unused cudf::strings::create_offsets ([#8663](https://github.com/rapidsai/cudf/pull/8663)) [@davidwendt](https://github.com/davidwendt) -- Add delimiter parameter to cudf::strings::capitalize() ([#8620](https://github.com/rapidsai/cudf/pull/8620)) [@davidwendt](https://github.com/davidwendt) -- Change default datetime index resolution to ns to match pandas ([#8611](https://github.com/rapidsai/cudf/pull/8611)) [@vyasr](https://github.com/vyasr) -- Add sequence_type parameter to cudf::strings::title function ([#8602](https://github.com/rapidsai/cudf/pull/8602)) [@davidwendt](https://github.com/davidwendt) -- Add `strings::repeat_strings` API that can repeat each string a different number of times ([#8561](https://github.com/rapidsai/cudf/pull/8561)) [@ttnghia](https://github.com/ttnghia) -- String-to-boolean conversion is different from Pandas ([#8549](https://github.com/rapidsai/cudf/pull/8549)) [@skirui-source](https://github.com/skirui-source) -- Add accurate hash join size functions ([#8453](https://github.com/rapidsai/cudf/pull/8453)) [@PointKernel](https://github.com/PointKernel) -- Expose a Decimal32Dtype in cuDF Python ([#8438](https://github.com/rapidsai/cudf/pull/8438)) [@skirui-source](https://github.com/skirui-source) -- Update dask make_meta changes to be compatible with dask upstream ([#8426](https://github.com/rapidsai/cudf/pull/8426)) [@galipremsagar](https://github.com/galipremsagar) -- Adapt `cudf::scalar` classes to changes in `rmm::device_scalar` ([#8411](https://github.com/rapidsai/cudf/pull/8411)) [@harrism](https://github.com/harrism) -- Remove special Index class from the general index class hierarchy ([#8309](https://github.com/rapidsai/cudf/pull/8309)) [@vyasr](https://github.com/vyasr) -- Add first-class dtype utilities ([#8308](https://github.com/rapidsai/cudf/pull/8308)) [@vyasr](https://github.com/vyasr) -- ORC - Support reading multiple orc files/buffers in a single operation ([#8142](https://github.com/rapidsai/cudf/pull/8142)) [@jdye64](https://github.com/jdye64) -- Upgrade arrow to 4.0.1 ([#7495](https://github.com/rapidsai/cudf/pull/7495)) [@galipremsagar](https://github.com/galipremsagar) +- Fix a crash in pack() when being handed tables with no columns. ([#8697](https://github.com/NVIDIA/cudf/pull/8697)) [@nvdbaranec](https://github.com/nvdbaranec) +- Remove unused cudf::strings::create_offsets ([#8663](https://github.com/NVIDIA/cudf/pull/8663)) [@davidwendt](https://github.com/davidwendt) +- Add delimiter parameter to cudf::strings::capitalize() ([#8620](https://github.com/NVIDIA/cudf/pull/8620)) [@davidwendt](https://github.com/davidwendt) +- Change default datetime index resolution to ns to match pandas ([#8611](https://github.com/NVIDIA/cudf/pull/8611)) [@vyasr](https://github.com/vyasr) +- Add sequence_type parameter to cudf::strings::title function ([#8602](https://github.com/NVIDIA/cudf/pull/8602)) [@davidwendt](https://github.com/davidwendt) +- Add `strings::repeat_strings` API that can repeat each string a different number of times ([#8561](https://github.com/NVIDIA/cudf/pull/8561)) [@ttnghia](https://github.com/ttnghia) +- String-to-boolean conversion is different from Pandas ([#8549](https://github.com/NVIDIA/cudf/pull/8549)) [@skirui-source](https://github.com/skirui-source) +- Add accurate hash join size functions ([#8453](https://github.com/NVIDIA/cudf/pull/8453)) [@PointKernel](https://github.com/PointKernel) +- Expose a Decimal32Dtype in cuDF Python ([#8438](https://github.com/NVIDIA/cudf/pull/8438)) [@skirui-source](https://github.com/skirui-source) +- Update dask make_meta changes to be compatible with dask upstream ([#8426](https://github.com/NVIDIA/cudf/pull/8426)) [@galipremsagar](https://github.com/galipremsagar) +- Adapt `cudf::scalar` classes to changes in `rmm::device_scalar` ([#8411](https://github.com/NVIDIA/cudf/pull/8411)) [@harrism](https://github.com/harrism) +- Remove special Index class from the general index class hierarchy ([#8309](https://github.com/NVIDIA/cudf/pull/8309)) [@vyasr](https://github.com/vyasr) +- Add first-class dtype utilities ([#8308](https://github.com/NVIDIA/cudf/pull/8308)) [@vyasr](https://github.com/vyasr) +- ORC - Support reading multiple orc files/buffers in a single operation ([#8142](https://github.com/NVIDIA/cudf/pull/8142)) [@jdye64](https://github.com/jdye64) +- Upgrade arrow to 4.0.1 ([#7495](https://github.com/NVIDIA/cudf/pull/7495)) [@galipremsagar](https://github.com/galipremsagar) ## 🐛 Bug Fixes -- Fix `contains` check in string column ([#8834](https://github.com/rapidsai/cudf/pull/8834)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unused variable from `row_bit_count_test`. ([#8829](https://github.com/rapidsai/cudf/pull/8829)) [@mythrocks](https://github.com/mythrocks) -- Fixes issue with null struct columns in ORC reader ([#8819](https://github.com/rapidsai/cudf/pull/8819)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Set CMake vars for python/parquet support in libarrow builds ([#8808](https://github.com/rapidsai/cudf/pull/8808)) [@vyasr](https://github.com/vyasr) -- Handle empty child columns in row_bit_count() ([#8791](https://github.com/rapidsai/cudf/pull/8791)) [@mythrocks](https://github.com/mythrocks) -- Revert "Remove cudf unneeded build time requirement of the cuda driver" ([#8784](https://github.com/rapidsai/cudf/pull/8784)) [@robertmaynard](https://github.com/robertmaynard) -- Fix isort error in utils.pyx ([#8771](https://github.com/rapidsai/cudf/pull/8771)) [@charlesbluca](https://github.com/charlesbluca) -- Handle sliced struct/list columns properly in concatenate() bounds checking. ([#8760](https://github.com/rapidsai/cudf/pull/8760)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix issues with `_CPackedColumns.serialize()` handling of host and device data ([#8759](https://github.com/rapidsai/cudf/pull/8759)) [@charlesbluca](https://github.com/charlesbluca) -- Fix issues with `MultiIndex` in `dropna`, `stack` & `reset_index` ([#8753](https://github.com/rapidsai/cudf/pull/8753)) [@galipremsagar](https://github.com/galipremsagar) -- Write pandas extension types to parquet file metadata ([#8749](https://github.com/rapidsai/cudf/pull/8749)) [@devavret](https://github.com/devavret) -- Fix `where` to handle `DataFrame` & `Series` input combination ([#8747](https://github.com/rapidsai/cudf/pull/8747)) [@galipremsagar](https://github.com/galipremsagar) -- Fix `replace` to handle null values correctly ([#8744](https://github.com/rapidsai/cudf/pull/8744)) [@galipremsagar](https://github.com/galipremsagar) -- Handle sliced structs properly in pack/contiguous_split. ([#8739](https://github.com/rapidsai/cudf/pull/8739)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix issue in slice() where columns with a positive offset were computing null counts incorrectly. ([#8738](https://github.com/rapidsai/cudf/pull/8738)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix `cudf.Series` constructor to handle list of sequences ([#8735](https://github.com/rapidsai/cudf/pull/8735)) [@galipremsagar](https://github.com/galipremsagar) -- Fix min/max sorted groupby aggregation on string column with nulls (argmin, argmax sentinel value missing on nulls) ([#8731](https://github.com/rapidsai/cudf/pull/8731)) [@karthikeyann](https://github.com/karthikeyann) -- Fix orc reader assert on create data_type in debug ([#8706](https://github.com/rapidsai/cudf/pull/8706)) [@davidwendt](https://github.com/davidwendt) -- Fix min/max inclusive cudf::scan for strings column ([#8705](https://github.com/rapidsai/cudf/pull/8705)) [@davidwendt](https://github.com/davidwendt) -- JNI: Fix driver version assertion logic in testGetCudaRuntimeInfo ([#8701](https://github.com/rapidsai/cudf/pull/8701)) [@sperlingxx](https://github.com/sperlingxx) -- Adding fix for skip_rows and crash in orc reader ([#8700](https://github.com/rapidsai/cudf/pull/8700)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Bug fix: `replace_nulls_policy` functor not returning correct indices for gathermap ([#8699](https://github.com/rapidsai/cudf/pull/8699)) [@isVoid](https://github.com/isVoid) -- Fix a crash in pack() when being handed tables with no columns. ([#8697](https://github.com/rapidsai/cudf/pull/8697)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add post-processing steps to `dask_cudf.groupby.CudfSeriesGroupby.aggregate` ([#8694](https://github.com/rapidsai/cudf/pull/8694)) [@charlesbluca](https://github.com/charlesbluca) -- JNI build no longer looks for Arrow in conda environment ([#8686](https://github.com/rapidsai/cudf/pull/8686)) [@jlowe](https://github.com/jlowe) -- Handle arbitrarily different data in null list column rows when checking for equivalency. ([#8666](https://github.com/rapidsai/cudf/pull/8666)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add ConfigureNVBench to avoid concurrent main() entry points ([#8662](https://github.com/rapidsai/cudf/pull/8662)) [@PointKernel](https://github.com/PointKernel) -- Pin `*arrow` to use `*cuda` in `run` ([#8651](https://github.com/rapidsai/cudf/pull/8651)) [@jakirkham](https://github.com/jakirkham) -- Add proper support for tolerances in testing methods. ([#8649](https://github.com/rapidsai/cudf/pull/8649)) [@vyasr](https://github.com/vyasr) -- Support multi-char case conversion in capitalize function ([#8647](https://github.com/rapidsai/cudf/pull/8647)) [@davidwendt](https://github.com/davidwendt) -- Fix repeated mangled names in read_csv with duplicate column names ([#8645](https://github.com/rapidsai/cudf/pull/8645)) [@karthikeyann](https://github.com/karthikeyann) -- Temporarily disable libcudf example build tests ([#8642](https://github.com/rapidsai/cudf/pull/8642)) [@isVoid](https://github.com/isVoid) -- Use conda-sourced cudf artifacts for libcudf example in CI ([#8638](https://github.com/rapidsai/cudf/pull/8638)) [@isVoid](https://github.com/isVoid) -- Ensure dev environment uses Arrow GPU packages ([#8637](https://github.com/rapidsai/cudf/pull/8637)) [@charlesbluca](https://github.com/charlesbluca) -- Fix bug that columns only initialized once when specified `columns` and `index` in dataframe ctor ([#8628](https://github.com/rapidsai/cudf/pull/8628)) [@isVoid](https://github.com/isVoid) -- Propagate **kwargs through to as_*_column methods ([#8618](https://github.com/rapidsai/cudf/pull/8618)) [@shwina](https://github.com/shwina) -- Fix orc_reader_benchmark.cpp compile error ([#8609](https://github.com/rapidsai/cudf/pull/8609)) [@davidwendt](https://github.com/davidwendt) -- Fix missed renumbering of Aggregation values ([#8600](https://github.com/rapidsai/cudf/pull/8600)) [@revans2](https://github.com/revans2) -- Update cmake to 3.20.5 in the Java Docker image ([#8593](https://github.com/rapidsai/cudf/pull/8593)) [@NvTimLiu](https://github.com/NvTimLiu) -- Fix bug in replace_with_backrefs when group has greedy quantifier ([#8575](https://github.com/rapidsai/cudf/pull/8575)) [@davidwendt](https://github.com/davidwendt) -- Apply metadata to keys before returning in `Frame._encode` ([#8560](https://github.com/rapidsai/cudf/pull/8560)) [@charlesbluca](https://github.com/charlesbluca) -- Fix for strings containing special JSON characters in get_json_object(). ([#8556](https://github.com/rapidsai/cudf/pull/8556)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix debug compile error in gather_struct_tests.cpp ([#8554](https://github.com/rapidsai/cudf/pull/8554)) [@davidwendt](https://github.com/davidwendt) -- String-to-boolean conversion is different from Pandas ([#8549](https://github.com/rapidsai/cudf/pull/8549)) [@skirui-source](https://github.com/skirui-source) -- Fix `__repr__` output with `display.max_rows` is `None` ([#8547](https://github.com/rapidsai/cudf/pull/8547)) [@galipremsagar](https://github.com/galipremsagar) -- Fix size passed to column constructors in _with_type_metadata ([#8539](https://github.com/rapidsai/cudf/pull/8539)) [@shwina](https://github.com/shwina) -- Properly retrieve last column when `-1` is specified for column index ([#8529](https://github.com/rapidsai/cudf/pull/8529)) [@isVoid](https://github.com/isVoid) -- Fix importing `apply` from `dask` ([#8517](https://github.com/rapidsai/cudf/pull/8517)) [@galipremsagar](https://github.com/galipremsagar) -- Fix offset of the string dictionary length stream ([#8515](https://github.com/rapidsai/cudf/pull/8515)) [@vuule](https://github.com/vuule) -- Fix double counting of selected columns in CSV reader ([#8508](https://github.com/rapidsai/cudf/pull/8508)) [@ochan1](https://github.com/ochan1) -- Incorrect map size in scatter_to_gather corrupts struct columns ([#8507](https://github.com/rapidsai/cudf/pull/8507)) [@gerashegalov](https://github.com/gerashegalov) -- replace_nulls properly propagates memory resource to gather calls ([#8500](https://github.com/rapidsai/cudf/pull/8500)) [@robertmaynard](https://github.com/robertmaynard) -- Disallow groupby aggs for `StructColumns` ([#8499](https://github.com/rapidsai/cudf/pull/8499)) [@charlesbluca](https://github.com/charlesbluca) -- Fixes out-of-bounds access for small files in unzip ([#8498](https://github.com/rapidsai/cudf/pull/8498)) [@elstehle](https://github.com/elstehle) -- Adding support for writing empty dataframe ([#8490](https://github.com/rapidsai/cudf/pull/8490)) [@shaneding](https://github.com/shaneding) -- Fix exclusive scan when including nulls and improve testing ([#8478](https://github.com/rapidsai/cudf/pull/8478)) [@harrism](https://github.com/harrism) -- Add workaround for crash in libcudf debug build using output_indexalator in thrust::lower_bound ([#8432](https://github.com/rapidsai/cudf/pull/8432)) [@davidwendt](https://github.com/davidwendt) -- Install only the same Thrust files that Thrust itself installs ([#8420](https://github.com/rapidsai/cudf/pull/8420)) [@robertmaynard](https://github.com/robertmaynard) -- Add nightly version for ucx-py in ci script ([#8419](https://github.com/rapidsai/cudf/pull/8419)) [@galipremsagar](https://github.com/galipremsagar) -- Fix null_equality config of rolling_collect_set ([#8415](https://github.com/rapidsai/cudf/pull/8415)) [@sperlingxx](https://github.com/sperlingxx) -- CollectSetAggregation: implement RollingAggregation interface ([#8406](https://github.com/rapidsai/cudf/pull/8406)) [@sperlingxx](https://github.com/sperlingxx) -- Handle pre-sliced nested columns in contiguous_split. ([#8391](https://github.com/rapidsai/cudf/pull/8391)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix bitmask_tests.cpp host accessing device memory ([#8370](https://github.com/rapidsai/cudf/pull/8370)) [@davidwendt](https://github.com/davidwendt) -- Fix concurrent_unordered_map to prevent accessing padding bits in pair_type ([#8348](https://github.com/rapidsai/cudf/pull/8348)) [@davidwendt](https://github.com/davidwendt) -- BUG FIX: Raise appropriate strings error when concatenating strings column ([#8290](https://github.com/rapidsai/cudf/pull/8290)) [@skirui-source](https://github.com/skirui-source) -- Make gpuCI and pre-commit style configurations consistent ([#8215](https://github.com/rapidsai/cudf/pull/8215)) [@charlesbluca](https://github.com/charlesbluca) -- Add collect list to dask-cudf groupby aggregations ([#8045](https://github.com/rapidsai/cudf/pull/8045)) [@charlesbluca](https://github.com/charlesbluca) +- Fix `contains` check in string column ([#8834](https://github.com/NVIDIA/cudf/pull/8834)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unused variable from `row_bit_count_test`. ([#8829](https://github.com/NVIDIA/cudf/pull/8829)) [@mythrocks](https://github.com/mythrocks) +- Fixes issue with null struct columns in ORC reader ([#8819](https://github.com/NVIDIA/cudf/pull/8819)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Set CMake vars for python/parquet support in libarrow builds ([#8808](https://github.com/NVIDIA/cudf/pull/8808)) [@vyasr](https://github.com/vyasr) +- Handle empty child columns in row_bit_count() ([#8791](https://github.com/NVIDIA/cudf/pull/8791)) [@mythrocks](https://github.com/mythrocks) +- Revert "Remove cudf unneeded build time requirement of the cuda driver" ([#8784](https://github.com/NVIDIA/cudf/pull/8784)) [@robertmaynard](https://github.com/robertmaynard) +- Fix isort error in utils.pyx ([#8771](https://github.com/NVIDIA/cudf/pull/8771)) [@charlesbluca](https://github.com/charlesbluca) +- Handle sliced struct/list columns properly in concatenate() bounds checking. ([#8760](https://github.com/NVIDIA/cudf/pull/8760)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix issues with `_CPackedColumns.serialize()` handling of host and device data ([#8759](https://github.com/NVIDIA/cudf/pull/8759)) [@charlesbluca](https://github.com/charlesbluca) +- Fix issues with `MultiIndex` in `dropna`, `stack` & `reset_index` ([#8753](https://github.com/NVIDIA/cudf/pull/8753)) [@galipremsagar](https://github.com/galipremsagar) +- Write pandas extension types to parquet file metadata ([#8749](https://github.com/NVIDIA/cudf/pull/8749)) [@devavret](https://github.com/devavret) +- Fix `where` to handle `DataFrame` & `Series` input combination ([#8747](https://github.com/NVIDIA/cudf/pull/8747)) [@galipremsagar](https://github.com/galipremsagar) +- Fix `replace` to handle null values correctly ([#8744](https://github.com/NVIDIA/cudf/pull/8744)) [@galipremsagar](https://github.com/galipremsagar) +- Handle sliced structs properly in pack/contiguous_split. ([#8739](https://github.com/NVIDIA/cudf/pull/8739)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix issue in slice() where columns with a positive offset were computing null counts incorrectly. ([#8738](https://github.com/NVIDIA/cudf/pull/8738)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix `cudf.Series` constructor to handle list of sequences ([#8735](https://github.com/NVIDIA/cudf/pull/8735)) [@galipremsagar](https://github.com/galipremsagar) +- Fix min/max sorted groupby aggregation on string column with nulls (argmin, argmax sentinel value missing on nulls) ([#8731](https://github.com/NVIDIA/cudf/pull/8731)) [@karthikeyann](https://github.com/karthikeyann) +- Fix orc reader assert on create data_type in debug ([#8706](https://github.com/NVIDIA/cudf/pull/8706)) [@davidwendt](https://github.com/davidwendt) +- Fix min/max inclusive cudf::scan for strings column ([#8705](https://github.com/NVIDIA/cudf/pull/8705)) [@davidwendt](https://github.com/davidwendt) +- JNI: Fix driver version assertion logic in testGetCudaRuntimeInfo ([#8701](https://github.com/NVIDIA/cudf/pull/8701)) [@sperlingxx](https://github.com/sperlingxx) +- Adding fix for skip_rows and crash in orc reader ([#8700](https://github.com/NVIDIA/cudf/pull/8700)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Bug fix: `replace_nulls_policy` functor not returning correct indices for gathermap ([#8699](https://github.com/NVIDIA/cudf/pull/8699)) [@isVoid](https://github.com/isVoid) +- Fix a crash in pack() when being handed tables with no columns. ([#8697](https://github.com/NVIDIA/cudf/pull/8697)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add post-processing steps to `dask_cudf.groupby.CudfSeriesGroupby.aggregate` ([#8694](https://github.com/NVIDIA/cudf/pull/8694)) [@charlesbluca](https://github.com/charlesbluca) +- JNI build no longer looks for Arrow in conda environment ([#8686](https://github.com/NVIDIA/cudf/pull/8686)) [@jlowe](https://github.com/jlowe) +- Handle arbitrarily different data in null list column rows when checking for equivalency. ([#8666](https://github.com/NVIDIA/cudf/pull/8666)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add ConfigureNVBench to avoid concurrent main() entry points ([#8662](https://github.com/NVIDIA/cudf/pull/8662)) [@PointKernel](https://github.com/PointKernel) +- Pin `*arrow` to use `*cuda` in `run` ([#8651](https://github.com/NVIDIA/cudf/pull/8651)) [@jakirkham](https://github.com/jakirkham) +- Add proper support for tolerances in testing methods. ([#8649](https://github.com/NVIDIA/cudf/pull/8649)) [@vyasr](https://github.com/vyasr) +- Support multi-char case conversion in capitalize function ([#8647](https://github.com/NVIDIA/cudf/pull/8647)) [@davidwendt](https://github.com/davidwendt) +- Fix repeated mangled names in read_csv with duplicate column names ([#8645](https://github.com/NVIDIA/cudf/pull/8645)) [@karthikeyann](https://github.com/karthikeyann) +- Temporarily disable libcudf example build tests ([#8642](https://github.com/NVIDIA/cudf/pull/8642)) [@isVoid](https://github.com/isVoid) +- Use conda-sourced cudf artifacts for libcudf example in CI ([#8638](https://github.com/NVIDIA/cudf/pull/8638)) [@isVoid](https://github.com/isVoid) +- Ensure dev environment uses Arrow GPU packages ([#8637](https://github.com/NVIDIA/cudf/pull/8637)) [@charlesbluca](https://github.com/charlesbluca) +- Fix bug that columns only initialized once when specified `columns` and `index` in dataframe ctor ([#8628](https://github.com/NVIDIA/cudf/pull/8628)) [@isVoid](https://github.com/isVoid) +- Propagate **kwargs through to as_*_column methods ([#8618](https://github.com/NVIDIA/cudf/pull/8618)) [@shwina](https://github.com/shwina) +- Fix orc_reader_benchmark.cpp compile error ([#8609](https://github.com/NVIDIA/cudf/pull/8609)) [@davidwendt](https://github.com/davidwendt) +- Fix missed renumbering of Aggregation values ([#8600](https://github.com/NVIDIA/cudf/pull/8600)) [@revans2](https://github.com/revans2) +- Update cmake to 3.20.5 in the Java Docker image ([#8593](https://github.com/NVIDIA/cudf/pull/8593)) [@NvTimLiu](https://github.com/NvTimLiu) +- Fix bug in replace_with_backrefs when group has greedy quantifier ([#8575](https://github.com/NVIDIA/cudf/pull/8575)) [@davidwendt](https://github.com/davidwendt) +- Apply metadata to keys before returning in `Frame._encode` ([#8560](https://github.com/NVIDIA/cudf/pull/8560)) [@charlesbluca](https://github.com/charlesbluca) +- Fix for strings containing special JSON characters in get_json_object(). ([#8556](https://github.com/NVIDIA/cudf/pull/8556)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix debug compile error in gather_struct_tests.cpp ([#8554](https://github.com/NVIDIA/cudf/pull/8554)) [@davidwendt](https://github.com/davidwendt) +- String-to-boolean conversion is different from Pandas ([#8549](https://github.com/NVIDIA/cudf/pull/8549)) [@skirui-source](https://github.com/skirui-source) +- Fix `__repr__` output with `display.max_rows` is `None` ([#8547](https://github.com/NVIDIA/cudf/pull/8547)) [@galipremsagar](https://github.com/galipremsagar) +- Fix size passed to column constructors in _with_type_metadata ([#8539](https://github.com/NVIDIA/cudf/pull/8539)) [@shwina](https://github.com/shwina) +- Properly retrieve last column when `-1` is specified for column index ([#8529](https://github.com/NVIDIA/cudf/pull/8529)) [@isVoid](https://github.com/isVoid) +- Fix importing `apply` from `dask` ([#8517](https://github.com/NVIDIA/cudf/pull/8517)) [@galipremsagar](https://github.com/galipremsagar) +- Fix offset of the string dictionary length stream ([#8515](https://github.com/NVIDIA/cudf/pull/8515)) [@vuule](https://github.com/vuule) +- Fix double counting of selected columns in CSV reader ([#8508](https://github.com/NVIDIA/cudf/pull/8508)) [@ochan1](https://github.com/ochan1) +- Incorrect map size in scatter_to_gather corrupts struct columns ([#8507](https://github.com/NVIDIA/cudf/pull/8507)) [@gerashegalov](https://github.com/gerashegalov) +- replace_nulls properly propagates memory resource to gather calls ([#8500](https://github.com/NVIDIA/cudf/pull/8500)) [@robertmaynard](https://github.com/robertmaynard) +- Disallow groupby aggs for `StructColumns` ([#8499](https://github.com/NVIDIA/cudf/pull/8499)) [@charlesbluca](https://github.com/charlesbluca) +- Fixes out-of-bounds access for small files in unzip ([#8498](https://github.com/NVIDIA/cudf/pull/8498)) [@elstehle](https://github.com/elstehle) +- Adding support for writing empty dataframe ([#8490](https://github.com/NVIDIA/cudf/pull/8490)) [@shaneding](https://github.com/shaneding) +- Fix exclusive scan when including nulls and improve testing ([#8478](https://github.com/NVIDIA/cudf/pull/8478)) [@harrism](https://github.com/harrism) +- Add workaround for crash in libcudf debug build using output_indexalator in thrust::lower_bound ([#8432](https://github.com/NVIDIA/cudf/pull/8432)) [@davidwendt](https://github.com/davidwendt) +- Install only the same Thrust files that Thrust itself installs ([#8420](https://github.com/NVIDIA/cudf/pull/8420)) [@robertmaynard](https://github.com/robertmaynard) +- Add nightly version for ucx-py in ci script ([#8419](https://github.com/NVIDIA/cudf/pull/8419)) [@galipremsagar](https://github.com/galipremsagar) +- Fix null_equality config of rolling_collect_set ([#8415](https://github.com/NVIDIA/cudf/pull/8415)) [@sperlingxx](https://github.com/sperlingxx) +- CollectSetAggregation: implement RollingAggregation interface ([#8406](https://github.com/NVIDIA/cudf/pull/8406)) [@sperlingxx](https://github.com/sperlingxx) +- Handle pre-sliced nested columns in contiguous_split. ([#8391](https://github.com/NVIDIA/cudf/pull/8391)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix bitmask_tests.cpp host accessing device memory ([#8370](https://github.com/NVIDIA/cudf/pull/8370)) [@davidwendt](https://github.com/davidwendt) +- Fix concurrent_unordered_map to prevent accessing padding bits in pair_type ([#8348](https://github.com/NVIDIA/cudf/pull/8348)) [@davidwendt](https://github.com/davidwendt) +- BUG FIX: Raise appropriate strings error when concatenating strings column ([#8290](https://github.com/NVIDIA/cudf/pull/8290)) [@skirui-source](https://github.com/skirui-source) +- Make gpuCI and pre-commit style configurations consistent ([#8215](https://github.com/NVIDIA/cudf/pull/8215)) [@charlesbluca](https://github.com/charlesbluca) +- Add collect list to dask-cudf groupby aggregations ([#8045](https://github.com/NVIDIA/cudf/pull/8045)) [@charlesbluca](https://github.com/charlesbluca) ## 📖 Documentation -- Update Python UDFs notebook ([#8810](https://github.com/rapidsai/cudf/pull/8810)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix dask.dataframe API docs links after reorg ([#8772](https://github.com/rapidsai/cudf/pull/8772)) [@jsignell](https://github.com/jsignell) -- Fix instructions for running cuDF/dask-cuDF tests in CONTRIBUTING.md ([#8724](https://github.com/rapidsai/cudf/pull/8724)) [@shwina](https://github.com/shwina) -- Translate Markdown documentation to rST and remove recommonmark ([#8698](https://github.com/rapidsai/cudf/pull/8698)) [@vyasr](https://github.com/vyasr) -- Fixed spelling mistakes in libcudf documentation ([#8664](https://github.com/rapidsai/cudf/pull/8664)) [@karthikeyann](https://github.com/karthikeyann) -- Custom Sphinx Extension: `PandasCompat` ([#8643](https://github.com/rapidsai/cudf/pull/8643)) [@isVoid](https://github.com/isVoid) -- Fix README.md ([#8535](https://github.com/rapidsai/cudf/pull/8535)) [@ajschmidt8](https://github.com/ajschmidt8) -- Change namespace contains_nulls to struct ([#8523](https://github.com/rapidsai/cudf/pull/8523)) [@davidwendt](https://github.com/davidwendt) -- Add info about NVTX ranges to dev guide ([#8461](https://github.com/rapidsai/cudf/pull/8461)) [@jrhemstad](https://github.com/jrhemstad) -- Fixed documentation bug in groupby agg method ([#8325](https://github.com/rapidsai/cudf/pull/8325)) [@ahmet-uyar](https://github.com/ahmet-uyar) +- Update Python UDFs notebook ([#8810](https://github.com/NVIDIA/cudf/pull/8810)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix dask.dataframe API docs links after reorg ([#8772](https://github.com/NVIDIA/cudf/pull/8772)) [@jsignell](https://github.com/jsignell) +- Fix instructions for running cuDF/dask-cuDF tests in CONTRIBUTING.md ([#8724](https://github.com/NVIDIA/cudf/pull/8724)) [@shwina](https://github.com/shwina) +- Translate Markdown documentation to rST and remove recommonmark ([#8698](https://github.com/NVIDIA/cudf/pull/8698)) [@vyasr](https://github.com/vyasr) +- Fixed spelling mistakes in libcudf documentation ([#8664](https://github.com/NVIDIA/cudf/pull/8664)) [@karthikeyann](https://github.com/karthikeyann) +- Custom Sphinx Extension: `PandasCompat` ([#8643](https://github.com/NVIDIA/cudf/pull/8643)) [@isVoid](https://github.com/isVoid) +- Fix README.md ([#8535](https://github.com/NVIDIA/cudf/pull/8535)) [@ajschmidt8](https://github.com/ajschmidt8) +- Change namespace contains_nulls to struct ([#8523](https://github.com/NVIDIA/cudf/pull/8523)) [@davidwendt](https://github.com/davidwendt) +- Add info about NVTX ranges to dev guide ([#8461](https://github.com/NVIDIA/cudf/pull/8461)) [@jrhemstad](https://github.com/jrhemstad) +- Fixed documentation bug in groupby agg method ([#8325](https://github.com/NVIDIA/cudf/pull/8325)) [@ahmet-uyar](https://github.com/ahmet-uyar) ## 🚀 New Features -- Fix concatenating structs ([#8811](https://github.com/rapidsai/cudf/pull/8811)) [@shaneding](https://github.com/shaneding) -- Implement JNI for groupby aggregations `M2` and `MERGE_M2` ([#8763](https://github.com/rapidsai/cudf/pull/8763)) [@ttnghia](https://github.com/ttnghia) -- Bump `isort` to `5.6.4` and remove `isort` overrides made for 5.0.7 ([#8755](https://github.com/rapidsai/cudf/pull/8755)) [@charlesbluca](https://github.com/charlesbluca) -- Implement `__setitem__` for `StructColumn` ([#8737](https://github.com/rapidsai/cudf/pull/8737)) [@shaneding](https://github.com/shaneding) -- Add `is_leap_year` to `DateTimeProperties` and `DatetimeIndex` ([#8736](https://github.com/rapidsai/cudf/pull/8736)) [@isVoid](https://github.com/isVoid) -- Add `struct.explode()` method ([#8729](https://github.com/rapidsai/cudf/pull/8729)) [@shwina](https://github.com/shwina) -- Add `DataFrame.to_struct()` method to convert a DataFrame to a struct Series ([#8728](https://github.com/rapidsai/cudf/pull/8728)) [@shwina](https://github.com/shwina) -- Add support for list type in ORC writer ([#8723](https://github.com/rapidsai/cudf/pull/8723)) [@vuule](https://github.com/vuule) -- Fix slicing from struct columns and accessing struct columns ([#8719](https://github.com/rapidsai/cudf/pull/8719)) [@shaneding](https://github.com/shaneding) -- Add `datetime::is_leap_year` ([#8711](https://github.com/rapidsai/cudf/pull/8711)) [@isVoid](https://github.com/isVoid) -- Accessing struct columns from `dask_cudf` ([#8675](https://github.com/rapidsai/cudf/pull/8675)) [@shaneding](https://github.com/shaneding) -- Added pct_change to Series ([#8650](https://github.com/rapidsai/cudf/pull/8650)) [@TravisHester](https://github.com/TravisHester) -- Add strings support to cudf::shift function ([#8648](https://github.com/rapidsai/cudf/pull/8648)) [@davidwendt](https://github.com/davidwendt) -- Support Scatter `struct_scalar` ([#8630](https://github.com/rapidsai/cudf/pull/8630)) [@isVoid](https://github.com/isVoid) -- Struct scalar from host dictionary ([#8629](https://github.com/rapidsai/cudf/pull/8629)) [@shaneding](https://github.com/shaneding) -- Add dayofyear and day_of_year to Series, DatetimeColumn, and DatetimeIndex ([#8626](https://github.com/rapidsai/cudf/pull/8626)) [@beckernick](https://github.com/beckernick) -- JNI support for capitalize ([#8624](https://github.com/rapidsai/cudf/pull/8624)) [@firestarman](https://github.com/firestarman) -- Add delimiter parameter to cudf::strings::capitalize() ([#8620](https://github.com/rapidsai/cudf/pull/8620)) [@davidwendt](https://github.com/davidwendt) -- Add NVBench in CMake ([#8619](https://github.com/rapidsai/cudf/pull/8619)) [@PointKernel](https://github.com/PointKernel) -- Change default datetime index resolution to ns to match pandas ([#8611](https://github.com/rapidsai/cudf/pull/8611)) [@vyasr](https://github.com/vyasr) -- ListColumn `__setitem__` ([#8606](https://github.com/rapidsai/cudf/pull/8606)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Implement groupby aggregations `M2` and `MERGE_M2` ([#8605](https://github.com/rapidsai/cudf/pull/8605)) [@ttnghia](https://github.com/ttnghia) -- Add sequence_type parameter to cudf::strings::title function ([#8602](https://github.com/rapidsai/cudf/pull/8602)) [@davidwendt](https://github.com/davidwendt) -- Adding support for list and struct type in ORC Reader ([#8599](https://github.com/rapidsai/cudf/pull/8599)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Benchmark for `strings::repeat_strings` APIs ([#8589](https://github.com/rapidsai/cudf/pull/8589)) [@ttnghia](https://github.com/ttnghia) -- Nested scalar support for copy if else ([#8588](https://github.com/rapidsai/cudf/pull/8588)) [@gerashegalov](https://github.com/gerashegalov) -- User specified decimal columns to float64 ([#8587](https://github.com/rapidsai/cudf/pull/8587)) [@jdye64](https://github.com/jdye64) -- Add `get_element` for struct column ([#8578](https://github.com/rapidsai/cudf/pull/8578)) [@isVoid](https://github.com/isVoid) -- Python changes for adding `__getitem__` for `struct` ([#8577](https://github.com/rapidsai/cudf/pull/8577)) [@shaneding](https://github.com/shaneding) -- Add `strings::repeat_strings` API that can repeat each string a different number of times ([#8561](https://github.com/rapidsai/cudf/pull/8561)) [@ttnghia](https://github.com/ttnghia) -- Refactor `tests/iterator_utilities.hpp` functions ([#8540](https://github.com/rapidsai/cudf/pull/8540)) [@ttnghia](https://github.com/ttnghia) -- Support MERGE_LISTS and MERGE_SETS in Java package ([#8516](https://github.com/rapidsai/cudf/pull/8516)) [@sperlingxx](https://github.com/sperlingxx) -- Decimal support csv reader ([#8511](https://github.com/rapidsai/cudf/pull/8511)) [@elstehle](https://github.com/elstehle) -- Add column type tests ([#8505](https://github.com/rapidsai/cudf/pull/8505)) [@isVoid](https://github.com/isVoid) -- Warn when downscaling decimal columns ([#8492](https://github.com/rapidsai/cudf/pull/8492)) [@ChrisJar](https://github.com/ChrisJar) -- Add JNI for `strings::repeat_strings` ([#8491](https://github.com/rapidsai/cudf/pull/8491)) [@ttnghia](https://github.com/ttnghia) -- Add `Index.get_loc` for Numerical, String Index support ([#8489](https://github.com/rapidsai/cudf/pull/8489)) [@isVoid](https://github.com/isVoid) -- Expose half_up rounding in cuDF ([#8477](https://github.com/rapidsai/cudf/pull/8477)) [@shwina](https://github.com/shwina) -- Java APIs to fetch CUDA runtime info ([#8465](https://github.com/rapidsai/cudf/pull/8465)) [@sperlingxx](https://github.com/sperlingxx) -- Add `str.edit_distance_matrix` ([#8463](https://github.com/rapidsai/cudf/pull/8463)) [@isVoid](https://github.com/isVoid) -- Support constructing `cudf.Scalar` objects from host side lists ([#8459](https://github.com/rapidsai/cudf/pull/8459)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add accurate hash join size functions ([#8453](https://github.com/rapidsai/cudf/pull/8453)) [@PointKernel](https://github.com/PointKernel) -- Add cudf::strings::integer_to_hex convert API ([#8450](https://github.com/rapidsai/cudf/pull/8450)) [@davidwendt](https://github.com/davidwendt) -- Create objects from iterables that contain cudf.NA ([#8442](https://github.com/rapidsai/cudf/pull/8442)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- JNI bindings for sort_lists ([#8439](https://github.com/rapidsai/cudf/pull/8439)) [@sperlingxx](https://github.com/sperlingxx) -- Expose a Decimal32Dtype in cuDF Python ([#8438](https://github.com/rapidsai/cudf/pull/8438)) [@skirui-source](https://github.com/skirui-source) -- Replace `all_null()` and `all_valid()` by `iterator_all_nulls()` and `iterator_no_null()` in tests ([#8437](https://github.com/rapidsai/cudf/pull/8437)) [@ttnghia](https://github.com/ttnghia) -- Implement groupby `MERGE_LISTS` and `MERGE_SETS` aggregates ([#8436](https://github.com/rapidsai/cudf/pull/8436)) [@ttnghia](https://github.com/ttnghia) -- Add public libcudf match_dictionaries API ([#8429](https://github.com/rapidsai/cudf/pull/8429)) [@davidwendt](https://github.com/davidwendt) -- Add move constructors for `string_scalar` and `struct_scalar` ([#8428](https://github.com/rapidsai/cudf/pull/8428)) [@ttnghia](https://github.com/ttnghia) -- Implement `strings::repeat_strings` ([#8423](https://github.com/rapidsai/cudf/pull/8423)) [@ttnghia](https://github.com/ttnghia) -- STRUCT column support for cudf::merge. ([#8422](https://github.com/rapidsai/cudf/pull/8422)) [@nvdbaranec](https://github.com/nvdbaranec) -- Implement reverse in libcudf ([#8410](https://github.com/rapidsai/cudf/pull/8410)) [@shaneding](https://github.com/shaneding) -- Support multiple input files/buffers for read_json ([#8403](https://github.com/rapidsai/cudf/pull/8403)) [@jdye64](https://github.com/jdye64) -- Improve test coverage for struct search ([#8396](https://github.com/rapidsai/cudf/pull/8396)) [@ttnghia](https://github.com/ttnghia) -- Add `groupby.fillna` ([#8362](https://github.com/rapidsai/cudf/pull/8362)) [@isVoid](https://github.com/isVoid) -- Enable AST-based joining ([#8214](https://github.com/rapidsai/cudf/pull/8214)) [@vyasr](https://github.com/vyasr) -- Generalized null support in user defined functions ([#8213](https://github.com/rapidsai/cudf/pull/8213)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add compiled binary operation ([#8192](https://github.com/rapidsai/cudf/pull/8192)) [@karthikeyann](https://github.com/karthikeyann) -- Implement `.describe() ` for `DataFrameGroupBy` ([#8179](https://github.com/rapidsai/cudf/pull/8179)) [@skirui-source](https://github.com/skirui-source) -- ORC - Support reading multiple orc files/buffers in a single operation ([#8142](https://github.com/rapidsai/cudf/pull/8142)) [@jdye64](https://github.com/jdye64) -- Add Python bindings for `lists::concatenate_list_elements` and expose them as `.list.concat()` ([#8006](https://github.com/rapidsai/cudf/pull/8006)) [@shwina](https://github.com/shwina) -- Use Arrow URI FileSystem backed instance to retrieve remote files ([#7709](https://github.com/rapidsai/cudf/pull/7709)) [@jdye64](https://github.com/jdye64) -- Example to build custom application and link to libcudf ([#7671](https://github.com/rapidsai/cudf/pull/7671)) [@isVoid](https://github.com/isVoid) -- Upgrade arrow to 4.0.1 ([#7495](https://github.com/rapidsai/cudf/pull/7495)) [@galipremsagar](https://github.com/galipremsagar) +- Fix concatenating structs ([#8811](https://github.com/NVIDIA/cudf/pull/8811)) [@shaneding](https://github.com/shaneding) +- Implement JNI for groupby aggregations `M2` and `MERGE_M2` ([#8763](https://github.com/NVIDIA/cudf/pull/8763)) [@ttnghia](https://github.com/ttnghia) +- Bump `isort` to `5.6.4` and remove `isort` overrides made for 5.0.7 ([#8755](https://github.com/NVIDIA/cudf/pull/8755)) [@charlesbluca](https://github.com/charlesbluca) +- Implement `__setitem__` for `StructColumn` ([#8737](https://github.com/NVIDIA/cudf/pull/8737)) [@shaneding](https://github.com/shaneding) +- Add `is_leap_year` to `DateTimeProperties` and `DatetimeIndex` ([#8736](https://github.com/NVIDIA/cudf/pull/8736)) [@isVoid](https://github.com/isVoid) +- Add `struct.explode()` method ([#8729](https://github.com/NVIDIA/cudf/pull/8729)) [@shwina](https://github.com/shwina) +- Add `DataFrame.to_struct()` method to convert a DataFrame to a struct Series ([#8728](https://github.com/NVIDIA/cudf/pull/8728)) [@shwina](https://github.com/shwina) +- Add support for list type in ORC writer ([#8723](https://github.com/NVIDIA/cudf/pull/8723)) [@vuule](https://github.com/vuule) +- Fix slicing from struct columns and accessing struct columns ([#8719](https://github.com/NVIDIA/cudf/pull/8719)) [@shaneding](https://github.com/shaneding) +- Add `datetime::is_leap_year` ([#8711](https://github.com/NVIDIA/cudf/pull/8711)) [@isVoid](https://github.com/isVoid) +- Accessing struct columns from `dask_cudf` ([#8675](https://github.com/NVIDIA/cudf/pull/8675)) [@shaneding](https://github.com/shaneding) +- Added pct_change to Series ([#8650](https://github.com/NVIDIA/cudf/pull/8650)) [@TravisHester](https://github.com/TravisHester) +- Add strings support to cudf::shift function ([#8648](https://github.com/NVIDIA/cudf/pull/8648)) [@davidwendt](https://github.com/davidwendt) +- Support Scatter `struct_scalar` ([#8630](https://github.com/NVIDIA/cudf/pull/8630)) [@isVoid](https://github.com/isVoid) +- Struct scalar from host dictionary ([#8629](https://github.com/NVIDIA/cudf/pull/8629)) [@shaneding](https://github.com/shaneding) +- Add dayofyear and day_of_year to Series, DatetimeColumn, and DatetimeIndex ([#8626](https://github.com/NVIDIA/cudf/pull/8626)) [@beckernick](https://github.com/beckernick) +- JNI support for capitalize ([#8624](https://github.com/NVIDIA/cudf/pull/8624)) [@firestarman](https://github.com/firestarman) +- Add delimiter parameter to cudf::strings::capitalize() ([#8620](https://github.com/NVIDIA/cudf/pull/8620)) [@davidwendt](https://github.com/davidwendt) +- Add NVBench in CMake ([#8619](https://github.com/NVIDIA/cudf/pull/8619)) [@PointKernel](https://github.com/PointKernel) +- Change default datetime index resolution to ns to match pandas ([#8611](https://github.com/NVIDIA/cudf/pull/8611)) [@vyasr](https://github.com/vyasr) +- ListColumn `__setitem__` ([#8606](https://github.com/NVIDIA/cudf/pull/8606)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Implement groupby aggregations `M2` and `MERGE_M2` ([#8605](https://github.com/NVIDIA/cudf/pull/8605)) [@ttnghia](https://github.com/ttnghia) +- Add sequence_type parameter to cudf::strings::title function ([#8602](https://github.com/NVIDIA/cudf/pull/8602)) [@davidwendt](https://github.com/davidwendt) +- Adding support for list and struct type in ORC Reader ([#8599](https://github.com/NVIDIA/cudf/pull/8599)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Benchmark for `strings::repeat_strings` APIs ([#8589](https://github.com/NVIDIA/cudf/pull/8589)) [@ttnghia](https://github.com/ttnghia) +- Nested scalar support for copy if else ([#8588](https://github.com/NVIDIA/cudf/pull/8588)) [@gerashegalov](https://github.com/gerashegalov) +- User specified decimal columns to float64 ([#8587](https://github.com/NVIDIA/cudf/pull/8587)) [@jdye64](https://github.com/jdye64) +- Add `get_element` for struct column ([#8578](https://github.com/NVIDIA/cudf/pull/8578)) [@isVoid](https://github.com/isVoid) +- Python changes for adding `__getitem__` for `struct` ([#8577](https://github.com/NVIDIA/cudf/pull/8577)) [@shaneding](https://github.com/shaneding) +- Add `strings::repeat_strings` API that can repeat each string a different number of times ([#8561](https://github.com/NVIDIA/cudf/pull/8561)) [@ttnghia](https://github.com/ttnghia) +- Refactor `tests/iterator_utilities.hpp` functions ([#8540](https://github.com/NVIDIA/cudf/pull/8540)) [@ttnghia](https://github.com/ttnghia) +- Support MERGE_LISTS and MERGE_SETS in Java package ([#8516](https://github.com/NVIDIA/cudf/pull/8516)) [@sperlingxx](https://github.com/sperlingxx) +- Decimal support csv reader ([#8511](https://github.com/NVIDIA/cudf/pull/8511)) [@elstehle](https://github.com/elstehle) +- Add column type tests ([#8505](https://github.com/NVIDIA/cudf/pull/8505)) [@isVoid](https://github.com/isVoid) +- Warn when downscaling decimal columns ([#8492](https://github.com/NVIDIA/cudf/pull/8492)) [@ChrisJar](https://github.com/ChrisJar) +- Add JNI for `strings::repeat_strings` ([#8491](https://github.com/NVIDIA/cudf/pull/8491)) [@ttnghia](https://github.com/ttnghia) +- Add `Index.get_loc` for Numerical, String Index support ([#8489](https://github.com/NVIDIA/cudf/pull/8489)) [@isVoid](https://github.com/isVoid) +- Expose half_up rounding in cuDF ([#8477](https://github.com/NVIDIA/cudf/pull/8477)) [@shwina](https://github.com/shwina) +- Java APIs to fetch CUDA runtime info ([#8465](https://github.com/NVIDIA/cudf/pull/8465)) [@sperlingxx](https://github.com/sperlingxx) +- Add `str.edit_distance_matrix` ([#8463](https://github.com/NVIDIA/cudf/pull/8463)) [@isVoid](https://github.com/isVoid) +- Support constructing `cudf.Scalar` objects from host side lists ([#8459](https://github.com/NVIDIA/cudf/pull/8459)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add accurate hash join size functions ([#8453](https://github.com/NVIDIA/cudf/pull/8453)) [@PointKernel](https://github.com/PointKernel) +- Add cudf::strings::integer_to_hex convert API ([#8450](https://github.com/NVIDIA/cudf/pull/8450)) [@davidwendt](https://github.com/davidwendt) +- Create objects from iterables that contain cudf.NA ([#8442](https://github.com/NVIDIA/cudf/pull/8442)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- JNI bindings for sort_lists ([#8439](https://github.com/NVIDIA/cudf/pull/8439)) [@sperlingxx](https://github.com/sperlingxx) +- Expose a Decimal32Dtype in cuDF Python ([#8438](https://github.com/NVIDIA/cudf/pull/8438)) [@skirui-source](https://github.com/skirui-source) +- Replace `all_null()` and `all_valid()` by `iterator_all_nulls()` and `iterator_no_null()` in tests ([#8437](https://github.com/NVIDIA/cudf/pull/8437)) [@ttnghia](https://github.com/ttnghia) +- Implement groupby `MERGE_LISTS` and `MERGE_SETS` aggregates ([#8436](https://github.com/NVIDIA/cudf/pull/8436)) [@ttnghia](https://github.com/ttnghia) +- Add public libcudf match_dictionaries API ([#8429](https://github.com/NVIDIA/cudf/pull/8429)) [@davidwendt](https://github.com/davidwendt) +- Add move constructors for `string_scalar` and `struct_scalar` ([#8428](https://github.com/NVIDIA/cudf/pull/8428)) [@ttnghia](https://github.com/ttnghia) +- Implement `strings::repeat_strings` ([#8423](https://github.com/NVIDIA/cudf/pull/8423)) [@ttnghia](https://github.com/ttnghia) +- STRUCT column support for cudf::merge. ([#8422](https://github.com/NVIDIA/cudf/pull/8422)) [@nvdbaranec](https://github.com/nvdbaranec) +- Implement reverse in libcudf ([#8410](https://github.com/NVIDIA/cudf/pull/8410)) [@shaneding](https://github.com/shaneding) +- Support multiple input files/buffers for read_json ([#8403](https://github.com/NVIDIA/cudf/pull/8403)) [@jdye64](https://github.com/jdye64) +- Improve test coverage for struct search ([#8396](https://github.com/NVIDIA/cudf/pull/8396)) [@ttnghia](https://github.com/ttnghia) +- Add `groupby.fillna` ([#8362](https://github.com/NVIDIA/cudf/pull/8362)) [@isVoid](https://github.com/isVoid) +- Enable AST-based joining ([#8214](https://github.com/NVIDIA/cudf/pull/8214)) [@vyasr](https://github.com/vyasr) +- Generalized null support in user defined functions ([#8213](https://github.com/NVIDIA/cudf/pull/8213)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add compiled binary operation ([#8192](https://github.com/NVIDIA/cudf/pull/8192)) [@karthikeyann](https://github.com/karthikeyann) +- Implement `.describe() ` for `DataFrameGroupBy` ([#8179](https://github.com/NVIDIA/cudf/pull/8179)) [@skirui-source](https://github.com/skirui-source) +- ORC - Support reading multiple orc files/buffers in a single operation ([#8142](https://github.com/NVIDIA/cudf/pull/8142)) [@jdye64](https://github.com/jdye64) +- Add Python bindings for `lists::concatenate_list_elements` and expose them as `.list.concat()` ([#8006](https://github.com/NVIDIA/cudf/pull/8006)) [@shwina](https://github.com/shwina) +- Use Arrow URI FileSystem backed instance to retrieve remote files ([#7709](https://github.com/NVIDIA/cudf/pull/7709)) [@jdye64](https://github.com/jdye64) +- Example to build custom application and link to libcudf ([#7671](https://github.com/NVIDIA/cudf/pull/7671)) [@isVoid](https://github.com/isVoid) +- Upgrade arrow to 4.0.1 ([#7495](https://github.com/NVIDIA/cudf/pull/7495)) [@galipremsagar](https://github.com/galipremsagar) ## 🛠️ Improvements -- Provide a better error message when `CUDA::cuda_driver` not found ([#8794](https://github.com/rapidsai/cudf/pull/8794)) [@robertmaynard](https://github.com/robertmaynard) -- Remove anonymous namespace from null_mask.cuh ([#8786](https://github.com/rapidsai/cudf/pull/8786)) [@nvdbaranec](https://github.com/nvdbaranec) -- Allow cudf to be built without libcuda.so existing ([#8751](https://github.com/rapidsai/cudf/pull/8751)) [@robertmaynard](https://github.com/robertmaynard) -- Pin `mimesis` to `<4.1` ([#8745](https://github.com/rapidsai/cudf/pull/8745)) [@galipremsagar](https://github.com/galipremsagar) -- Update `conda` environment name for CI ([#8692](https://github.com/rapidsai/cudf/pull/8692)) [@ajschmidt8](https://github.com/ajschmidt8) -- Remove flatbuffers dependency ([#8671](https://github.com/rapidsai/cudf/pull/8671)) [@Ethyling](https://github.com/Ethyling) -- Add options to build Arrow with Python and Parquet support ([#8670](https://github.com/rapidsai/cudf/pull/8670)) [@trxcllnt](https://github.com/trxcllnt) -- Remove unused cudf::strings::create_offsets ([#8663](https://github.com/rapidsai/cudf/pull/8663)) [@davidwendt](https://github.com/davidwendt) -- Update GDS lib version to 1.0.0 ([#8654](https://github.com/rapidsai/cudf/pull/8654)) [@pxLi](https://github.com/pxLi) -- Support for groupby/scan rank and dense_rank aggregations ([#8652](https://github.com/rapidsai/cudf/pull/8652)) [@rwlee](https://github.com/rwlee) -- Fix usage of deprecated arrow ipc API ([#8632](https://github.com/rapidsai/cudf/pull/8632)) [@revans2](https://github.com/revans2) -- Use absolute imports in `cudf` ([#8631](https://github.com/rapidsai/cudf/pull/8631)) [@galipremsagar](https://github.com/galipremsagar) -- ENH Add Java CI build script ([#8627](https://github.com/rapidsai/cudf/pull/8627)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- Add DeprecationWarning to `ser.str.subword_tokenize` ([#8603](https://github.com/rapidsai/cudf/pull/8603)) [@VibhuJawa](https://github.com/VibhuJawa) -- Rewrite binary operations for improved performance and additional type support ([#8598](https://github.com/rapidsai/cudf/pull/8598)) [@vyasr](https://github.com/vyasr) -- Fix `mypy` errors surfacing because of `numpy-1.21.0` ([#8595](https://github.com/rapidsai/cudf/pull/8595)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unneeded includes from cudf::string_view headers ([#8594](https://github.com/rapidsai/cudf/pull/8594)) [@davidwendt](https://github.com/davidwendt) -- Use cmake 3.20.1 as it is now required by rmm ([#8586](https://github.com/rapidsai/cudf/pull/8586)) [@robertmaynard](https://github.com/robertmaynard) -- Remove device debug symbols from cmake CUDF_CUDA_FLAGS ([#8584](https://github.com/rapidsai/cudf/pull/8584)) [@davidwendt](https://github.com/davidwendt) -- Dask-CuDF: use default Dask Dataframe optimizer ([#8581](https://github.com/rapidsai/cudf/pull/8581)) [@madsbk](https://github.com/madsbk) -- Remove checking if an unsigned value is less than zero ([#8579](https://github.com/rapidsai/cudf/pull/8579)) [@robertmaynard](https://github.com/robertmaynard) -- Remove strings_count parameter from cudf::strings::detail::create_chars_child_column ([#8576](https://github.com/rapidsai/cudf/pull/8576)) [@davidwendt](https://github.com/davidwendt) -- Make `cudf.api.types` imports consistent ([#8571](https://github.com/rapidsai/cudf/pull/8571)) [@galipremsagar](https://github.com/galipremsagar) -- Modernize libcudf basic example CMakeFile; updates CI build tests ([#8568](https://github.com/rapidsai/cudf/pull/8568)) [@isVoid](https://github.com/isVoid) -- Rename concatenate_tests.cu to .cpp ([#8555](https://github.com/rapidsai/cudf/pull/8555)) [@davidwendt](https://github.com/davidwendt) -- enable window lead/lag test on struct ([#8548](https://github.com/rapidsai/cudf/pull/8548)) [@wbo4958](https://github.com/wbo4958) -- Add Java methods to split and write column views ([#8546](https://github.com/rapidsai/cudf/pull/8546)) [@razajafri](https://github.com/razajafri) -- Small cleanup ([#8534](https://github.com/rapidsai/cudf/pull/8534)) [@codereport](https://github.com/codereport) -- Unpin `dask` version in CI ([#8533](https://github.com/rapidsai/cudf/pull/8533)) [@galipremsagar](https://github.com/galipremsagar) -- Added optional flag for building Arrow with S3 filesystem support ([#8531](https://github.com/rapidsai/cudf/pull/8531)) [@jdye64](https://github.com/jdye64) -- Minor clean up of various internal column and frame utilities ([#8528](https://github.com/rapidsai/cudf/pull/8528)) [@vyasr](https://github.com/vyasr) -- Rename some copying_test source files .cu to .cpp ([#8527](https://github.com/rapidsai/cudf/pull/8527)) [@davidwendt](https://github.com/davidwendt) -- Correct the last warnings and issues when using newer cuda versions ([#8525](https://github.com/rapidsai/cudf/pull/8525)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in transform and unary ops ([#8521](https://github.com/rapidsai/cudf/pull/8521)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in string algorithms ([#8509](https://github.com/rapidsai/cudf/pull/8509)) [@robertmaynard](https://github.com/robertmaynard) -- Add in JNI APIs for scan, replace_nulls, group_by.scan, and group_by.replace_nulls ([#8503](https://github.com/rapidsai/cudf/pull/8503)) [@revans2](https://github.com/revans2) -- Fix `21.08` forward-merge conflicts ([#8502](https://github.com/rapidsai/cudf/pull/8502)) [@ajschmidt8](https://github.com/ajschmidt8) -- Fix Cython formatting command in Contributing.md. ([#8496](https://github.com/rapidsai/cudf/pull/8496)) [@marlenezw](https://github.com/marlenezw) -- Bug/correct unused parameters in reshape and text ([#8495](https://github.com/rapidsai/cudf/pull/8495)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in partitioning and stream compact ([#8494](https://github.com/rapidsai/cudf/pull/8494)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in labelling and list algorithms ([#8493](https://github.com/rapidsai/cudf/pull/8493)) [@robertmaynard](https://github.com/robertmaynard) -- Refactor index construction ([#8485](https://github.com/rapidsai/cudf/pull/8485)) [@vyasr](https://github.com/vyasr) -- Correct unused parameter warnings in replace algorithms ([#8483](https://github.com/rapidsai/cudf/pull/8483)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in reduction algorithms ([#8481](https://github.com/rapidsai/cudf/pull/8481)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in io algorithms ([#8480](https://github.com/rapidsai/cudf/pull/8480)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in interop algorithms ([#8479](https://github.com/rapidsai/cudf/pull/8479)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in filling algorithms ([#8468](https://github.com/rapidsai/cudf/pull/8468)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameter warnings in groupby ([#8467](https://github.com/rapidsai/cudf/pull/8467)) [@robertmaynard](https://github.com/robertmaynard) -- use libcu++ time_point as timestamp ([#8466](https://github.com/rapidsai/cudf/pull/8466)) [@karthikeyann](https://github.com/karthikeyann) -- Modify reprog_device::extract to return groups in a single pass ([#8460](https://github.com/rapidsai/cudf/pull/8460)) [@davidwendt](https://github.com/davidwendt) -- Update minimum Dask requirement to 2021.6.0 ([#8458](https://github.com/rapidsai/cudf/pull/8458)) [@pentschev](https://github.com/pentschev) -- Fix failures when performing binary operations on DataFrames with empty columns ([#8452](https://github.com/rapidsai/cudf/pull/8452)) [@ChrisJar](https://github.com/ChrisJar) -- Fix conflicts in `8447` ([#8448](https://github.com/rapidsai/cudf/pull/8448)) [@ajschmidt8](https://github.com/ajschmidt8) -- Add serialization methods for `List` and `StructDtype` ([#8441](https://github.com/rapidsai/cudf/pull/8441)) [@charlesbluca](https://github.com/charlesbluca) -- Replace make_empty_strings_column with make_empty_column ([#8435](https://github.com/rapidsai/cudf/pull/8435)) [@davidwendt](https://github.com/davidwendt) -- JNI bindings for get_element ([#8433](https://github.com/rapidsai/cudf/pull/8433)) [@revans2](https://github.com/revans2) -- Update dask make_meta changes to be compatible with dask upstream ([#8426](https://github.com/rapidsai/cudf/pull/8426)) [@galipremsagar](https://github.com/galipremsagar) -- Unpin dask version on CI ([#8425](https://github.com/rapidsai/cudf/pull/8425)) [@galipremsagar](https://github.com/galipremsagar) -- Add benchmark for strings/fixed_point convert APIs ([#8417](https://github.com/rapidsai/cudf/pull/8417)) [@davidwendt](https://github.com/davidwendt) -- Adapt `cudf::scalar` classes to changes in `rmm::device_scalar` ([#8411](https://github.com/rapidsai/cudf/pull/8411)) [@harrism](https://github.com/harrism) -- Add benchmark for strings/integers convert APIs ([#8402](https://github.com/rapidsai/cudf/pull/8402)) [@davidwendt](https://github.com/davidwendt) -- Enable multi-file partitioning in dask_cudf.read_parquet ([#8393](https://github.com/rapidsai/cudf/pull/8393)) [@rjzamora](https://github.com/rjzamora) -- Correct unused parameter warnings in rolling algorithms ([#8390](https://github.com/rapidsai/cudf/pull/8390)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameters in column round and search ([#8389](https://github.com/rapidsai/cudf/pull/8389)) [@robertmaynard](https://github.com/robertmaynard) -- Add functionality to apply `Dtype` metadata to `ColumnBase` ([#8373](https://github.com/rapidsai/cudf/pull/8373)) [@charlesbluca](https://github.com/charlesbluca) -- Refactor setting stack size in regex code ([#8358](https://github.com/rapidsai/cudf/pull/8358)) [@davidwendt](https://github.com/davidwendt) -- Update Java bindings to 21.08-SNAPSHOT ([#8344](https://github.com/rapidsai/cudf/pull/8344)) [@pxLi](https://github.com/pxLi) -- Replace remaining uses of device_vector ([#8343](https://github.com/rapidsai/cudf/pull/8343)) [@harrism](https://github.com/harrism) -- Statically link libnvcomp into libcudfjni ([#8334](https://github.com/rapidsai/cudf/pull/8334)) [@jlowe](https://github.com/jlowe) -- Resolve auto merge conflicts for Branch 21.08 from branch 21.06 ([#8329](https://github.com/rapidsai/cudf/pull/8329)) [@galipremsagar](https://github.com/galipremsagar) -- Minor code refactor for sorted_order ([#8326](https://github.com/rapidsai/cudf/pull/8326)) [@wbo4958](https://github.com/wbo4958) -- Remove special Index class from the general index class hierarchy ([#8309](https://github.com/rapidsai/cudf/pull/8309)) [@vyasr](https://github.com/vyasr) -- Add first-class dtype utilities ([#8308](https://github.com/rapidsai/cudf/pull/8308)) [@vyasr](https://github.com/vyasr) -- Add option to link Java bindings with Arrow dynamically ([#8307](https://github.com/rapidsai/cudf/pull/8307)) [@jlowe](https://github.com/jlowe) -- Refactor ColumnMethods and its subclasses to remove `column` argument and require `parent` argument ([#8306](https://github.com/rapidsai/cudf/pull/8306)) [@shwina](https://github.com/shwina) -- Refactor `scatter` for list columns ([#8255](https://github.com/rapidsai/cudf/pull/8255)) [@isVoid](https://github.com/isVoid) -- Expose pack/unpack API to Python ([#8153](https://github.com/rapidsai/cudf/pull/8153)) [@charlesbluca](https://github.com/charlesbluca) -- Adding cudf.cut method ([#8002](https://github.com/rapidsai/cudf/pull/8002)) [@marlenezw](https://github.com/marlenezw) -- Optimize string gather performance for large strings ([#7980](https://github.com/rapidsai/cudf/pull/7980)) [@gaohao95](https://github.com/gaohao95) -- Add peak memory usage tracking to cuIO benchmarks ([#7770](https://github.com/rapidsai/cudf/pull/7770)) [@devavret](https://github.com/devavret) -- Updating Clang Version to 11.0.0 ([#6695](https://github.com/rapidsai/cudf/pull/6695)) [@codereport](https://github.com/codereport) +- Provide a better error message when `CUDA::cuda_driver` not found ([#8794](https://github.com/NVIDIA/cudf/pull/8794)) [@robertmaynard](https://github.com/robertmaynard) +- Remove anonymous namespace from null_mask.cuh ([#8786](https://github.com/NVIDIA/cudf/pull/8786)) [@nvdbaranec](https://github.com/nvdbaranec) +- Allow cudf to be built without libcuda.so existing ([#8751](https://github.com/NVIDIA/cudf/pull/8751)) [@robertmaynard](https://github.com/robertmaynard) +- Pin `mimesis` to `<4.1` ([#8745](https://github.com/NVIDIA/cudf/pull/8745)) [@galipremsagar](https://github.com/galipremsagar) +- Update `conda` environment name for CI ([#8692](https://github.com/NVIDIA/cudf/pull/8692)) [@ajschmidt8](https://github.com/ajschmidt8) +- Remove flatbuffers dependency ([#8671](https://github.com/NVIDIA/cudf/pull/8671)) [@Ethyling](https://github.com/Ethyling) +- Add options to build Arrow with Python and Parquet support ([#8670](https://github.com/NVIDIA/cudf/pull/8670)) [@trxcllnt](https://github.com/trxcllnt) +- Remove unused cudf::strings::create_offsets ([#8663](https://github.com/NVIDIA/cudf/pull/8663)) [@davidwendt](https://github.com/davidwendt) +- Update GDS lib version to 1.0.0 ([#8654](https://github.com/NVIDIA/cudf/pull/8654)) [@pxLi](https://github.com/pxLi) +- Support for groupby/scan rank and dense_rank aggregations ([#8652](https://github.com/NVIDIA/cudf/pull/8652)) [@rwlee](https://github.com/rwlee) +- Fix usage of deprecated arrow ipc API ([#8632](https://github.com/NVIDIA/cudf/pull/8632)) [@revans2](https://github.com/revans2) +- Use absolute imports in `cudf` ([#8631](https://github.com/NVIDIA/cudf/pull/8631)) [@galipremsagar](https://github.com/galipremsagar) +- ENH Add Java CI build script ([#8627](https://github.com/NVIDIA/cudf/pull/8627)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- Add DeprecationWarning to `ser.str.subword_tokenize` ([#8603](https://github.com/NVIDIA/cudf/pull/8603)) [@VibhuJawa](https://github.com/VibhuJawa) +- Rewrite binary operations for improved performance and additional type support ([#8598](https://github.com/NVIDIA/cudf/pull/8598)) [@vyasr](https://github.com/vyasr) +- Fix `mypy` errors surfacing because of `numpy-1.21.0` ([#8595](https://github.com/NVIDIA/cudf/pull/8595)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unneeded includes from cudf::string_view headers ([#8594](https://github.com/NVIDIA/cudf/pull/8594)) [@davidwendt](https://github.com/davidwendt) +- Use cmake 3.20.1 as it is now required by rmm ([#8586](https://github.com/NVIDIA/cudf/pull/8586)) [@robertmaynard](https://github.com/robertmaynard) +- Remove device debug symbols from cmake CUDF_CUDA_FLAGS ([#8584](https://github.com/NVIDIA/cudf/pull/8584)) [@davidwendt](https://github.com/davidwendt) +- Dask-CuDF: use default Dask Dataframe optimizer ([#8581](https://github.com/NVIDIA/cudf/pull/8581)) [@madsbk](https://github.com/madsbk) +- Remove checking if an unsigned value is less than zero ([#8579](https://github.com/NVIDIA/cudf/pull/8579)) [@robertmaynard](https://github.com/robertmaynard) +- Remove strings_count parameter from cudf::strings::detail::create_chars_child_column ([#8576](https://github.com/NVIDIA/cudf/pull/8576)) [@davidwendt](https://github.com/davidwendt) +- Make `cudf.api.types` imports consistent ([#8571](https://github.com/NVIDIA/cudf/pull/8571)) [@galipremsagar](https://github.com/galipremsagar) +- Modernize libcudf basic example CMakeFile; updates CI build tests ([#8568](https://github.com/NVIDIA/cudf/pull/8568)) [@isVoid](https://github.com/isVoid) +- Rename concatenate_tests.cu to .cpp ([#8555](https://github.com/NVIDIA/cudf/pull/8555)) [@davidwendt](https://github.com/davidwendt) +- enable window lead/lag test on struct ([#8548](https://github.com/NVIDIA/cudf/pull/8548)) [@wbo4958](https://github.com/wbo4958) +- Add Java methods to split and write column views ([#8546](https://github.com/NVIDIA/cudf/pull/8546)) [@razajafri](https://github.com/razajafri) +- Small cleanup ([#8534](https://github.com/NVIDIA/cudf/pull/8534)) [@codereport](https://github.com/codereport) +- Unpin `dask` version in CI ([#8533](https://github.com/NVIDIA/cudf/pull/8533)) [@galipremsagar](https://github.com/galipremsagar) +- Added optional flag for building Arrow with S3 filesystem support ([#8531](https://github.com/NVIDIA/cudf/pull/8531)) [@jdye64](https://github.com/jdye64) +- Minor clean up of various internal column and frame utilities ([#8528](https://github.com/NVIDIA/cudf/pull/8528)) [@vyasr](https://github.com/vyasr) +- Rename some copying_test source files .cu to .cpp ([#8527](https://github.com/NVIDIA/cudf/pull/8527)) [@davidwendt](https://github.com/davidwendt) +- Correct the last warnings and issues when using newer cuda versions ([#8525](https://github.com/NVIDIA/cudf/pull/8525)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in transform and unary ops ([#8521](https://github.com/NVIDIA/cudf/pull/8521)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in string algorithms ([#8509](https://github.com/NVIDIA/cudf/pull/8509)) [@robertmaynard](https://github.com/robertmaynard) +- Add in JNI APIs for scan, replace_nulls, group_by.scan, and group_by.replace_nulls ([#8503](https://github.com/NVIDIA/cudf/pull/8503)) [@revans2](https://github.com/revans2) +- Fix `21.08` forward-merge conflicts ([#8502](https://github.com/NVIDIA/cudf/pull/8502)) [@ajschmidt8](https://github.com/ajschmidt8) +- Fix Cython formatting command in Contributing.md. ([#8496](https://github.com/NVIDIA/cudf/pull/8496)) [@marlenezw](https://github.com/marlenezw) +- Bug/correct unused parameters in reshape and text ([#8495](https://github.com/NVIDIA/cudf/pull/8495)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in partitioning and stream compact ([#8494](https://github.com/NVIDIA/cudf/pull/8494)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in labelling and list algorithms ([#8493](https://github.com/NVIDIA/cudf/pull/8493)) [@robertmaynard](https://github.com/robertmaynard) +- Refactor index construction ([#8485](https://github.com/NVIDIA/cudf/pull/8485)) [@vyasr](https://github.com/vyasr) +- Correct unused parameter warnings in replace algorithms ([#8483](https://github.com/NVIDIA/cudf/pull/8483)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in reduction algorithms ([#8481](https://github.com/NVIDIA/cudf/pull/8481)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in io algorithms ([#8480](https://github.com/NVIDIA/cudf/pull/8480)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in interop algorithms ([#8479](https://github.com/NVIDIA/cudf/pull/8479)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in filling algorithms ([#8468](https://github.com/NVIDIA/cudf/pull/8468)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameter warnings in groupby ([#8467](https://github.com/NVIDIA/cudf/pull/8467)) [@robertmaynard](https://github.com/robertmaynard) +- use libcu++ time_point as timestamp ([#8466](https://github.com/NVIDIA/cudf/pull/8466)) [@karthikeyann](https://github.com/karthikeyann) +- Modify reprog_device::extract to return groups in a single pass ([#8460](https://github.com/NVIDIA/cudf/pull/8460)) [@davidwendt](https://github.com/davidwendt) +- Update minimum Dask requirement to 2021.6.0 ([#8458](https://github.com/NVIDIA/cudf/pull/8458)) [@pentschev](https://github.com/pentschev) +- Fix failures when performing binary operations on DataFrames with empty columns ([#8452](https://github.com/NVIDIA/cudf/pull/8452)) [@ChrisJar](https://github.com/ChrisJar) +- Fix conflicts in `8447` ([#8448](https://github.com/NVIDIA/cudf/pull/8448)) [@ajschmidt8](https://github.com/ajschmidt8) +- Add serialization methods for `List` and `StructDtype` ([#8441](https://github.com/NVIDIA/cudf/pull/8441)) [@charlesbluca](https://github.com/charlesbluca) +- Replace make_empty_strings_column with make_empty_column ([#8435](https://github.com/NVIDIA/cudf/pull/8435)) [@davidwendt](https://github.com/davidwendt) +- JNI bindings for get_element ([#8433](https://github.com/NVIDIA/cudf/pull/8433)) [@revans2](https://github.com/revans2) +- Update dask make_meta changes to be compatible with dask upstream ([#8426](https://github.com/NVIDIA/cudf/pull/8426)) [@galipremsagar](https://github.com/galipremsagar) +- Unpin dask version on CI ([#8425](https://github.com/NVIDIA/cudf/pull/8425)) [@galipremsagar](https://github.com/galipremsagar) +- Add benchmark for strings/fixed_point convert APIs ([#8417](https://github.com/NVIDIA/cudf/pull/8417)) [@davidwendt](https://github.com/davidwendt) +- Adapt `cudf::scalar` classes to changes in `rmm::device_scalar` ([#8411](https://github.com/NVIDIA/cudf/pull/8411)) [@harrism](https://github.com/harrism) +- Add benchmark for strings/integers convert APIs ([#8402](https://github.com/NVIDIA/cudf/pull/8402)) [@davidwendt](https://github.com/davidwendt) +- Enable multi-file partitioning in dask_cudf.read_parquet ([#8393](https://github.com/NVIDIA/cudf/pull/8393)) [@rjzamora](https://github.com/rjzamora) +- Correct unused parameter warnings in rolling algorithms ([#8390](https://github.com/NVIDIA/cudf/pull/8390)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameters in column round and search ([#8389](https://github.com/NVIDIA/cudf/pull/8389)) [@robertmaynard](https://github.com/robertmaynard) +- Add functionality to apply `Dtype` metadata to `ColumnBase` ([#8373](https://github.com/NVIDIA/cudf/pull/8373)) [@charlesbluca](https://github.com/charlesbluca) +- Refactor setting stack size in regex code ([#8358](https://github.com/NVIDIA/cudf/pull/8358)) [@davidwendt](https://github.com/davidwendt) +- Update Java bindings to 21.08-SNAPSHOT ([#8344](https://github.com/NVIDIA/cudf/pull/8344)) [@pxLi](https://github.com/pxLi) +- Replace remaining uses of device_vector ([#8343](https://github.com/NVIDIA/cudf/pull/8343)) [@harrism](https://github.com/harrism) +- Statically link libnvcomp into libcudfjni ([#8334](https://github.com/NVIDIA/cudf/pull/8334)) [@jlowe](https://github.com/jlowe) +- Resolve auto merge conflicts for Branch 21.08 from branch 21.06 ([#8329](https://github.com/NVIDIA/cudf/pull/8329)) [@galipremsagar](https://github.com/galipremsagar) +- Minor code refactor for sorted_order ([#8326](https://github.com/NVIDIA/cudf/pull/8326)) [@wbo4958](https://github.com/wbo4958) +- Remove special Index class from the general index class hierarchy ([#8309](https://github.com/NVIDIA/cudf/pull/8309)) [@vyasr](https://github.com/vyasr) +- Add first-class dtype utilities ([#8308](https://github.com/NVIDIA/cudf/pull/8308)) [@vyasr](https://github.com/vyasr) +- Add option to link Java bindings with Arrow dynamically ([#8307](https://github.com/NVIDIA/cudf/pull/8307)) [@jlowe](https://github.com/jlowe) +- Refactor ColumnMethods and its subclasses to remove `column` argument and require `parent` argument ([#8306](https://github.com/NVIDIA/cudf/pull/8306)) [@shwina](https://github.com/shwina) +- Refactor `scatter` for list columns ([#8255](https://github.com/NVIDIA/cudf/pull/8255)) [@isVoid](https://github.com/isVoid) +- Expose pack/unpack API to Python ([#8153](https://github.com/NVIDIA/cudf/pull/8153)) [@charlesbluca](https://github.com/charlesbluca) +- Adding cudf.cut method ([#8002](https://github.com/NVIDIA/cudf/pull/8002)) [@marlenezw](https://github.com/marlenezw) +- Optimize string gather performance for large strings ([#7980](https://github.com/NVIDIA/cudf/pull/7980)) [@gaohao95](https://github.com/gaohao95) +- Add peak memory usage tracking to cuIO benchmarks ([#7770](https://github.com/NVIDIA/cudf/pull/7770)) [@devavret](https://github.com/devavret) +- Updating Clang Version to 11.0.0 ([#6695](https://github.com/NVIDIA/cudf/pull/6695)) [@codereport](https://github.com/codereport) # cuDF 21.06.00 (9 Jun 2021) ## 🚨 Breaking Changes -- Add support for `make_meta_obj` dispatch in `dask-cudf` ([#8342](https://github.com/rapidsai/cudf/pull/8342)) [@galipremsagar](https://github.com/galipremsagar) -- Add separator-on-null parameter to strings concatenate APIs ([#8282](https://github.com/rapidsai/cudf/pull/8282)) [@davidwendt](https://github.com/davidwendt) -- Introduce a common parent class for NumericalColumn and DecimalColumn ([#8278](https://github.com/rapidsai/cudf/pull/8278)) [@vyasr](https://github.com/vyasr) -- Update ORC statistics API to use C++17 standard library ([#8241](https://github.com/rapidsai/cudf/pull/8241)) [@vuule](https://github.com/vuule) -- Preserve column hierarchy when getting NULL row from `LIST` column ([#8206](https://github.com/rapidsai/cudf/pull/8206)) [@isVoid](https://github.com/isVoid) -- `Groupby.shift` c++ API refactor and python binding ([#8131](https://github.com/rapidsai/cudf/pull/8131)) [@isVoid](https://github.com/isVoid) +- Add support for `make_meta_obj` dispatch in `dask-cudf` ([#8342](https://github.com/NVIDIA/cudf/pull/8342)) [@galipremsagar](https://github.com/galipremsagar) +- Add separator-on-null parameter to strings concatenate APIs ([#8282](https://github.com/NVIDIA/cudf/pull/8282)) [@davidwendt](https://github.com/davidwendt) +- Introduce a common parent class for NumericalColumn and DecimalColumn ([#8278](https://github.com/NVIDIA/cudf/pull/8278)) [@vyasr](https://github.com/vyasr) +- Update ORC statistics API to use C++17 standard library ([#8241](https://github.com/NVIDIA/cudf/pull/8241)) [@vuule](https://github.com/vuule) +- Preserve column hierarchy when getting NULL row from `LIST` column ([#8206](https://github.com/NVIDIA/cudf/pull/8206)) [@isVoid](https://github.com/isVoid) +- `Groupby.shift` c++ API refactor and python binding ([#8131](https://github.com/NVIDIA/cudf/pull/8131)) [@isVoid](https://github.com/isVoid) ## 🐛 Bug Fixes -- Fix struct flattening to add a validity column only when the input column has null element ([#8374](https://github.com/rapidsai/cudf/pull/8374)) [@ttnghia](https://github.com/ttnghia) -- Compilation fix: Remove redefinition for `std::is_same_v()` ([#8369](https://github.com/rapidsai/cudf/pull/8369)) [@mythrocks](https://github.com/mythrocks) -- Add backward compatibility for `dask-cudf` to work with other versions of `dask` ([#8368](https://github.com/rapidsai/cudf/pull/8368)) [@galipremsagar](https://github.com/galipremsagar) -- Handle empty results with nested types in copy_if_else ([#8359](https://github.com/rapidsai/cudf/pull/8359)) [@nvdbaranec](https://github.com/nvdbaranec) -- Handle nested column types properly for empty parquet files. ([#8350](https://github.com/rapidsai/cudf/pull/8350)) [@nvdbaranec](https://github.com/nvdbaranec) -- Raise error when unsupported arguments are passed to `dask_cudf.DataFrame.sort_values` ([#8349](https://github.com/rapidsai/cudf/pull/8349)) [@galipremsagar](https://github.com/galipremsagar) -- Raise `NotImplementedError` for axis=1 in `rank` ([#8347](https://github.com/rapidsai/cudf/pull/8347)) [@galipremsagar](https://github.com/galipremsagar) -- Add support for `make_meta_obj` dispatch in `dask-cudf` ([#8342](https://github.com/rapidsai/cudf/pull/8342)) [@galipremsagar](https://github.com/galipremsagar) -- Update Java string concatenate test for single column ([#8330](https://github.com/rapidsai/cudf/pull/8330)) [@tgravescs](https://github.com/tgravescs) -- Use empty_like in scatter ([#8314](https://github.com/rapidsai/cudf/pull/8314)) [@revans2](https://github.com/revans2) -- Fix concatenate_lists_ignore_null on rows of all_nulls ([#8312](https://github.com/rapidsai/cudf/pull/8312)) [@sperlingxx](https://github.com/sperlingxx) -- Add separator-on-null parameter to strings concatenate APIs ([#8282](https://github.com/rapidsai/cudf/pull/8282)) [@davidwendt](https://github.com/davidwendt) -- COLLECT_LIST support returning empty output columns. ([#8279](https://github.com/rapidsai/cudf/pull/8279)) [@mythrocks](https://github.com/mythrocks) -- Update io util to convert path like object to string ([#8275](https://github.com/rapidsai/cudf/pull/8275)) [@ayushdg](https://github.com/ayushdg) -- Fix result column types for empty inputs to rolling window ([#8274](https://github.com/rapidsai/cudf/pull/8274)) [@mythrocks](https://github.com/mythrocks) -- Actually test equality in assert_groupby_results_equal ([#8272](https://github.com/rapidsai/cudf/pull/8272)) [@shwina](https://github.com/shwina) -- CMake always explicitly specify a source files extension ([#8270](https://github.com/rapidsai/cudf/pull/8270)) [@robertmaynard](https://github.com/robertmaynard) -- Fix struct binary search and struct flattening ([#8268](https://github.com/rapidsai/cudf/pull/8268)) [@ttnghia](https://github.com/ttnghia) -- Revert "patch thrust to fix intmax num elements limitation in scan_by_key" ([#8263](https://github.com/rapidsai/cudf/pull/8263)) [@cwharris](https://github.com/cwharris) -- upgrade dlpack to 0.5 ([#8262](https://github.com/rapidsai/cudf/pull/8262)) [@cwharris](https://github.com/cwharris) -- Fixes CSV-reader type inference for thousands separator and decimal point ([#8261](https://github.com/rapidsai/cudf/pull/8261)) [@elstehle](https://github.com/elstehle) -- Fix incorrect assertion in Java concat ([#8258](https://github.com/rapidsai/cudf/pull/8258)) [@sperlingxx](https://github.com/sperlingxx) -- Copy nested types upon construction ([#8244](https://github.com/rapidsai/cudf/pull/8244)) [@isVoid](https://github.com/isVoid) -- Preserve column hierarchy when getting NULL row from `LIST` column ([#8206](https://github.com/rapidsai/cudf/pull/8206)) [@isVoid](https://github.com/isVoid) -- Clip decimal binary op precision at max precision ([#8194](https://github.com/rapidsai/cudf/pull/8194)) [@ChrisJar](https://github.com/ChrisJar) +- Fix struct flattening to add a validity column only when the input column has null element ([#8374](https://github.com/NVIDIA/cudf/pull/8374)) [@ttnghia](https://github.com/ttnghia) +- Compilation fix: Remove redefinition for `std::is_same_v()` ([#8369](https://github.com/NVIDIA/cudf/pull/8369)) [@mythrocks](https://github.com/mythrocks) +- Add backward compatibility for `dask-cudf` to work with other versions of `dask` ([#8368](https://github.com/NVIDIA/cudf/pull/8368)) [@galipremsagar](https://github.com/galipremsagar) +- Handle empty results with nested types in copy_if_else ([#8359](https://github.com/NVIDIA/cudf/pull/8359)) [@nvdbaranec](https://github.com/nvdbaranec) +- Handle nested column types properly for empty parquet files. ([#8350](https://github.com/NVIDIA/cudf/pull/8350)) [@nvdbaranec](https://github.com/nvdbaranec) +- Raise error when unsupported arguments are passed to `dask_cudf.DataFrame.sort_values` ([#8349](https://github.com/NVIDIA/cudf/pull/8349)) [@galipremsagar](https://github.com/galipremsagar) +- Raise `NotImplementedError` for axis=1 in `rank` ([#8347](https://github.com/NVIDIA/cudf/pull/8347)) [@galipremsagar](https://github.com/galipremsagar) +- Add support for `make_meta_obj` dispatch in `dask-cudf` ([#8342](https://github.com/NVIDIA/cudf/pull/8342)) [@galipremsagar](https://github.com/galipremsagar) +- Update Java string concatenate test for single column ([#8330](https://github.com/NVIDIA/cudf/pull/8330)) [@tgravescs](https://github.com/tgravescs) +- Use empty_like in scatter ([#8314](https://github.com/NVIDIA/cudf/pull/8314)) [@revans2](https://github.com/revans2) +- Fix concatenate_lists_ignore_null on rows of all_nulls ([#8312](https://github.com/NVIDIA/cudf/pull/8312)) [@sperlingxx](https://github.com/sperlingxx) +- Add separator-on-null parameter to strings concatenate APIs ([#8282](https://github.com/NVIDIA/cudf/pull/8282)) [@davidwendt](https://github.com/davidwendt) +- COLLECT_LIST support returning empty output columns. ([#8279](https://github.com/NVIDIA/cudf/pull/8279)) [@mythrocks](https://github.com/mythrocks) +- Update io util to convert path like object to string ([#8275](https://github.com/NVIDIA/cudf/pull/8275)) [@ayushdg](https://github.com/ayushdg) +- Fix result column types for empty inputs to rolling window ([#8274](https://github.com/NVIDIA/cudf/pull/8274)) [@mythrocks](https://github.com/mythrocks) +- Actually test equality in assert_groupby_results_equal ([#8272](https://github.com/NVIDIA/cudf/pull/8272)) [@shwina](https://github.com/shwina) +- CMake always explicitly specify a source files extension ([#8270](https://github.com/NVIDIA/cudf/pull/8270)) [@robertmaynard](https://github.com/robertmaynard) +- Fix struct binary search and struct flattening ([#8268](https://github.com/NVIDIA/cudf/pull/8268)) [@ttnghia](https://github.com/ttnghia) +- Revert "patch thrust to fix intmax num elements limitation in scan_by_key" ([#8263](https://github.com/NVIDIA/cudf/pull/8263)) [@cwharris](https://github.com/cwharris) +- upgrade dlpack to 0.5 ([#8262](https://github.com/NVIDIA/cudf/pull/8262)) [@cwharris](https://github.com/cwharris) +- Fixes CSV-reader type inference for thousands separator and decimal point ([#8261](https://github.com/NVIDIA/cudf/pull/8261)) [@elstehle](https://github.com/elstehle) +- Fix incorrect assertion in Java concat ([#8258](https://github.com/NVIDIA/cudf/pull/8258)) [@sperlingxx](https://github.com/sperlingxx) +- Copy nested types upon construction ([#8244](https://github.com/NVIDIA/cudf/pull/8244)) [@isVoid](https://github.com/isVoid) +- Preserve column hierarchy when getting NULL row from `LIST` column ([#8206](https://github.com/NVIDIA/cudf/pull/8206)) [@isVoid](https://github.com/isVoid) +- Clip decimal binary op precision at max precision ([#8194](https://github.com/NVIDIA/cudf/pull/8194)) [@ChrisJar](https://github.com/ChrisJar) ## 📖 Documentation -- Add docstring for `dask_cudf.read_csv` ([#8355](https://github.com/rapidsai/cudf/pull/8355)) [@galipremsagar](https://github.com/galipremsagar) -- Fix cudf release version in readme ([#8331](https://github.com/rapidsai/cudf/pull/8331)) [@galipremsagar](https://github.com/galipremsagar) -- Fix structs column description in dev docs ([#8318](https://github.com/rapidsai/cudf/pull/8318)) [@isVoid](https://github.com/isVoid) -- Update readme with correct CUDA versions ([#8315](https://github.com/rapidsai/cudf/pull/8315)) [@raydouglass](https://github.com/raydouglass) -- Add description of the cuIO GDS integration ([#8293](https://github.com/rapidsai/cudf/pull/8293)) [@vuule](https://github.com/vuule) -- Remove unused parameter from copy_partition kernel documentation ([#8283](https://github.com/rapidsai/cudf/pull/8283)) [@robertmaynard](https://github.com/robertmaynard) +- Add docstring for `dask_cudf.read_csv` ([#8355](https://github.com/NVIDIA/cudf/pull/8355)) [@galipremsagar](https://github.com/galipremsagar) +- Fix cudf release version in readme ([#8331](https://github.com/NVIDIA/cudf/pull/8331)) [@galipremsagar](https://github.com/galipremsagar) +- Fix structs column description in dev docs ([#8318](https://github.com/NVIDIA/cudf/pull/8318)) [@isVoid](https://github.com/isVoid) +- Update readme with correct CUDA versions ([#8315](https://github.com/NVIDIA/cudf/pull/8315)) [@raydouglass](https://github.com/raydouglass) +- Add description of the cuIO GDS integration ([#8293](https://github.com/NVIDIA/cudf/pull/8293)) [@vuule](https://github.com/vuule) +- Remove unused parameter from copy_partition kernel documentation ([#8283](https://github.com/NVIDIA/cudf/pull/8283)) [@robertmaynard](https://github.com/robertmaynard) ## 🚀 New Features -- Add support merging b/w categorical data ([#8332](https://github.com/rapidsai/cudf/pull/8332)) [@galipremsagar](https://github.com/galipremsagar) -- Java: Support struct scalar ([#8327](https://github.com/rapidsai/cudf/pull/8327)) [@sperlingxx](https://github.com/sperlingxx) -- added _is_homogeneous property ([#8299](https://github.com/rapidsai/cudf/pull/8299)) [@shaneding](https://github.com/shaneding) -- Added decimal writing for CSV writer ([#8296](https://github.com/rapidsai/cudf/pull/8296)) [@kaatish](https://github.com/kaatish) -- Java: Support creating a scalar from utf8 string ([#8294](https://github.com/rapidsai/cudf/pull/8294)) [@firestarman](https://github.com/firestarman) -- Add Java API for Concatenate strings with separator ([#8289](https://github.com/rapidsai/cudf/pull/8289)) [@tgravescs](https://github.com/tgravescs) -- `strings::join_list_elements` options for empty list inputs ([#8285](https://github.com/rapidsai/cudf/pull/8285)) [@ttnghia](https://github.com/ttnghia) -- Return python lists for __getitem__ calls to list type series ([#8265](https://github.com/rapidsai/cudf/pull/8265)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- add unit tests for lead/lag on list for row window ([#8259](https://github.com/rapidsai/cudf/pull/8259)) [@wbo4958](https://github.com/wbo4958) -- Create a String column from UTF8 String byte arrays ([#8257](https://github.com/rapidsai/cudf/pull/8257)) [@firestarman](https://github.com/firestarman) -- Support scattering `list_scalar` ([#8256](https://github.com/rapidsai/cudf/pull/8256)) [@isVoid](https://github.com/isVoid) -- Implement `lists::concatenate_list_elements` ([#8231](https://github.com/rapidsai/cudf/pull/8231)) [@ttnghia](https://github.com/ttnghia) -- Support for struct scalars. ([#8220](https://github.com/rapidsai/cudf/pull/8220)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add support for decimal types in ORC writer ([#8198](https://github.com/rapidsai/cudf/pull/8198)) [@vuule](https://github.com/vuule) -- Support create lists column from a `list_scalar` ([#8185](https://github.com/rapidsai/cudf/pull/8185)) [@isVoid](https://github.com/isVoid) -- `Groupby.shift` c++ API refactor and python binding ([#8131](https://github.com/rapidsai/cudf/pull/8131)) [@isVoid](https://github.com/isVoid) -- Add `groupby::replace_nulls(replace_policy)` api ([#7118](https://github.com/rapidsai/cudf/pull/7118)) [@isVoid](https://github.com/isVoid) +- Add support merging b/w categorical data ([#8332](https://github.com/NVIDIA/cudf/pull/8332)) [@galipremsagar](https://github.com/galipremsagar) +- Java: Support struct scalar ([#8327](https://github.com/NVIDIA/cudf/pull/8327)) [@sperlingxx](https://github.com/sperlingxx) +- added _is_homogeneous property ([#8299](https://github.com/NVIDIA/cudf/pull/8299)) [@shaneding](https://github.com/shaneding) +- Added decimal writing for CSV writer ([#8296](https://github.com/NVIDIA/cudf/pull/8296)) [@kaatish](https://github.com/kaatish) +- Java: Support creating a scalar from utf8 string ([#8294](https://github.com/NVIDIA/cudf/pull/8294)) [@firestarman](https://github.com/firestarman) +- Add Java API for Concatenate strings with separator ([#8289](https://github.com/NVIDIA/cudf/pull/8289)) [@tgravescs](https://github.com/tgravescs) +- `strings::join_list_elements` options for empty list inputs ([#8285](https://github.com/NVIDIA/cudf/pull/8285)) [@ttnghia](https://github.com/ttnghia) +- Return python lists for __getitem__ calls to list type series ([#8265](https://github.com/NVIDIA/cudf/pull/8265)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- add unit tests for lead/lag on list for row window ([#8259](https://github.com/NVIDIA/cudf/pull/8259)) [@wbo4958](https://github.com/wbo4958) +- Create a String column from UTF8 String byte arrays ([#8257](https://github.com/NVIDIA/cudf/pull/8257)) [@firestarman](https://github.com/firestarman) +- Support scattering `list_scalar` ([#8256](https://github.com/NVIDIA/cudf/pull/8256)) [@isVoid](https://github.com/isVoid) +- Implement `lists::concatenate_list_elements` ([#8231](https://github.com/NVIDIA/cudf/pull/8231)) [@ttnghia](https://github.com/ttnghia) +- Support for struct scalars. ([#8220](https://github.com/NVIDIA/cudf/pull/8220)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add support for decimal types in ORC writer ([#8198](https://github.com/NVIDIA/cudf/pull/8198)) [@vuule](https://github.com/vuule) +- Support create lists column from a `list_scalar` ([#8185](https://github.com/NVIDIA/cudf/pull/8185)) [@isVoid](https://github.com/isVoid) +- `Groupby.shift` c++ API refactor and python binding ([#8131](https://github.com/NVIDIA/cudf/pull/8131)) [@isVoid](https://github.com/isVoid) +- Add `groupby::replace_nulls(replace_policy)` api ([#7118](https://github.com/NVIDIA/cudf/pull/7118)) [@isVoid](https://github.com/isVoid) ## 🛠️ Improvements -- Support Dask + Distributed 2021.05.1 ([#8392](https://github.com/rapidsai/cudf/pull/8392)) [@jakirkham](https://github.com/jakirkham) -- Add aliases for string methods ([#8353](https://github.com/rapidsai/cudf/pull/8353)) [@shwina](https://github.com/shwina) -- Update environment variable used to determine `cuda_version` ([#8321](https://github.com/rapidsai/cudf/pull/8321)) [@ajschmidt8](https://github.com/ajschmidt8) -- JNI: Refactor the code of making column from scalar ([#8310](https://github.com/rapidsai/cudf/pull/8310)) [@firestarman](https://github.com/firestarman) -- Update `CHANGELOG.md` links for calver ([#8303](https://github.com/rapidsai/cudf/pull/8303)) [@ajschmidt8](https://github.com/ajschmidt8) -- Merge `branch-0.19` into `branch-21.06` ([#8302](https://github.com/rapidsai/cudf/pull/8302)) [@ajschmidt8](https://github.com/ajschmidt8) -- use address and length for GDS reads/writes ([#8301](https://github.com/rapidsai/cudf/pull/8301)) [@rongou](https://github.com/rongou) -- Update cudfjni version to 21.06.0 ([#8292](https://github.com/rapidsai/cudf/pull/8292)) [@pxLi](https://github.com/pxLi) -- Update docs build script ([#8284](https://github.com/rapidsai/cudf/pull/8284)) [@ajschmidt8](https://github.com/ajschmidt8) -- Make device_buffer streams explicit and enforce move construction ([#8280](https://github.com/rapidsai/cudf/pull/8280)) [@harrism](https://github.com/harrism) -- Introduce a common parent class for NumericalColumn and DecimalColumn ([#8278](https://github.com/rapidsai/cudf/pull/8278)) [@vyasr](https://github.com/vyasr) -- Do not add nulls to the hash table when null_equality::NOT_EQUAL is passed to left_semi_join and left_anti_join ([#8277](https://github.com/rapidsai/cudf/pull/8277)) [@nvdbaranec](https://github.com/nvdbaranec) -- Enable implicit casting when concatenating mixed types ([#8276](https://github.com/rapidsai/cudf/pull/8276)) [@ChrisJar](https://github.com/ChrisJar) -- Fix CMake FindPackage rmm, pin dev envs' dlpack to v0.3 ([#8271](https://github.com/rapidsai/cudf/pull/8271)) [@trxcllnt](https://github.com/trxcllnt) -- Update cudfjni version to 21.06 ([#8267](https://github.com/rapidsai/cudf/pull/8267)) [@pxLi](https://github.com/pxLi) -- support RMM aligned resource adapter in JNI ([#8266](https://github.com/rapidsai/cudf/pull/8266)) [@rongou](https://github.com/rongou) -- Pass compiler environment variables to conda python build ([#8260](https://github.com/rapidsai/cudf/pull/8260)) [@Ethyling](https://github.com/Ethyling) -- Remove abc inheritance from Serializable ([#8254](https://github.com/rapidsai/cudf/pull/8254)) [@vyasr](https://github.com/vyasr) -- Move more methods into SingleColumnFrame ([#8253](https://github.com/rapidsai/cudf/pull/8253)) [@vyasr](https://github.com/vyasr) -- Update ORC statistics API to use C++17 standard library ([#8241](https://github.com/rapidsai/cudf/pull/8241)) [@vuule](https://github.com/vuule) -- Correct unused parameter warnings in dictionary algorithms ([#8239](https://github.com/rapidsai/cudf/pull/8239)) [@robertmaynard](https://github.com/robertmaynard) -- Correct unused parameters in the copying algorithms ([#8232](https://github.com/rapidsai/cudf/pull/8232)) [@robertmaynard](https://github.com/robertmaynard) -- IO statistics cleanup ([#8191](https://github.com/rapidsai/cudf/pull/8191)) [@kaatish](https://github.com/kaatish) -- Refactor of rolling_window implementation. ([#8158](https://github.com/rapidsai/cudf/pull/8158)) [@nvdbaranec](https://github.com/nvdbaranec) -- Add a flag for allowing single quotes in JSON strings. ([#8144](https://github.com/rapidsai/cudf/pull/8144)) [@nvdbaranec](https://github.com/nvdbaranec) -- Column refactoring 2 ([#8130](https://github.com/rapidsai/cudf/pull/8130)) [@vyasr](https://github.com/vyasr) -- support space in workspace ([#7956](https://github.com/rapidsai/cudf/pull/7956)) [@jolorunyomi](https://github.com/jolorunyomi) -- Support collect_set on rolling window ([#7881](https://github.com/rapidsai/cudf/pull/7881)) [@sperlingxx](https://github.com/sperlingxx) +- Support Dask + Distributed 2021.05.1 ([#8392](https://github.com/NVIDIA/cudf/pull/8392)) [@jakirkham](https://github.com/jakirkham) +- Add aliases for string methods ([#8353](https://github.com/NVIDIA/cudf/pull/8353)) [@shwina](https://github.com/shwina) +- Update environment variable used to determine `cuda_version` ([#8321](https://github.com/NVIDIA/cudf/pull/8321)) [@ajschmidt8](https://github.com/ajschmidt8) +- JNI: Refactor the code of making column from scalar ([#8310](https://github.com/NVIDIA/cudf/pull/8310)) [@firestarman](https://github.com/firestarman) +- Update `CHANGELOG.md` links for calver ([#8303](https://github.com/NVIDIA/cudf/pull/8303)) [@ajschmidt8](https://github.com/ajschmidt8) +- Merge `branch-0.19` into `branch-21.06` ([#8302](https://github.com/NVIDIA/cudf/pull/8302)) [@ajschmidt8](https://github.com/ajschmidt8) +- use address and length for GDS reads/writes ([#8301](https://github.com/NVIDIA/cudf/pull/8301)) [@rongou](https://github.com/rongou) +- Update cudfjni version to 21.06.0 ([#8292](https://github.com/NVIDIA/cudf/pull/8292)) [@pxLi](https://github.com/pxLi) +- Update docs build script ([#8284](https://github.com/NVIDIA/cudf/pull/8284)) [@ajschmidt8](https://github.com/ajschmidt8) +- Make device_buffer streams explicit and enforce move construction ([#8280](https://github.com/NVIDIA/cudf/pull/8280)) [@harrism](https://github.com/harrism) +- Introduce a common parent class for NumericalColumn and DecimalColumn ([#8278](https://github.com/NVIDIA/cudf/pull/8278)) [@vyasr](https://github.com/vyasr) +- Do not add nulls to the hash table when null_equality::NOT_EQUAL is passed to left_semi_join and left_anti_join ([#8277](https://github.com/NVIDIA/cudf/pull/8277)) [@nvdbaranec](https://github.com/nvdbaranec) +- Enable implicit casting when concatenating mixed types ([#8276](https://github.com/NVIDIA/cudf/pull/8276)) [@ChrisJar](https://github.com/ChrisJar) +- Fix CMake FindPackage rmm, pin dev envs' dlpack to v0.3 ([#8271](https://github.com/NVIDIA/cudf/pull/8271)) [@trxcllnt](https://github.com/trxcllnt) +- Update cudfjni version to 21.06 ([#8267](https://github.com/NVIDIA/cudf/pull/8267)) [@pxLi](https://github.com/pxLi) +- support RMM aligned resource adapter in JNI ([#8266](https://github.com/NVIDIA/cudf/pull/8266)) [@rongou](https://github.com/rongou) +- Pass compiler environment variables to conda python build ([#8260](https://github.com/NVIDIA/cudf/pull/8260)) [@Ethyling](https://github.com/Ethyling) +- Remove abc inheritance from Serializable ([#8254](https://github.com/NVIDIA/cudf/pull/8254)) [@vyasr](https://github.com/vyasr) +- Move more methods into SingleColumnFrame ([#8253](https://github.com/NVIDIA/cudf/pull/8253)) [@vyasr](https://github.com/vyasr) +- Update ORC statistics API to use C++17 standard library ([#8241](https://github.com/NVIDIA/cudf/pull/8241)) [@vuule](https://github.com/vuule) +- Correct unused parameter warnings in dictionary algorithms ([#8239](https://github.com/NVIDIA/cudf/pull/8239)) [@robertmaynard](https://github.com/robertmaynard) +- Correct unused parameters in the copying algorithms ([#8232](https://github.com/NVIDIA/cudf/pull/8232)) [@robertmaynard](https://github.com/robertmaynard) +- IO statistics cleanup ([#8191](https://github.com/NVIDIA/cudf/pull/8191)) [@kaatish](https://github.com/kaatish) +- Refactor of rolling_window implementation. ([#8158](https://github.com/NVIDIA/cudf/pull/8158)) [@nvdbaranec](https://github.com/nvdbaranec) +- Add a flag for allowing single quotes in JSON strings. ([#8144](https://github.com/NVIDIA/cudf/pull/8144)) [@nvdbaranec](https://github.com/nvdbaranec) +- Column refactoring 2 ([#8130](https://github.com/NVIDIA/cudf/pull/8130)) [@vyasr](https://github.com/vyasr) +- support space in workspace ([#7956](https://github.com/NVIDIA/cudf/pull/7956)) [@jolorunyomi](https://github.com/jolorunyomi) +- Support collect_set on rolling window ([#7881](https://github.com/NVIDIA/cudf/pull/7881)) [@sperlingxx](https://github.com/sperlingxx) # cuDF 0.19.0 (21 Apr 2021) ## 🚨 Breaking Changes -- Allow hash_partition to take a seed value ([#7771](https://github.com/rapidsai/cudf/pull/7771)) [@magnatelee](https://github.com/magnatelee) -- Allow merging index column with data column using keyword "on" ([#7736](https://github.com/rapidsai/cudf/pull/7736)) [@skirui-source](https://github.com/skirui-source) -- Change JNI API to avoid loading native dependencies when creating sort order classes. ([#7729](https://github.com/rapidsai/cudf/pull/7729)) [@revans2](https://github.com/revans2) -- Replace device_vector with device_uvector in null_mask ([#7715](https://github.com/rapidsai/cudf/pull/7715)) [@harrism](https://github.com/harrism) -- Don't identify decimals as strings. ([#7710](https://github.com/rapidsai/cudf/pull/7710)) [@vyasr](https://github.com/vyasr) -- Fix Java Parquet write after writer API changes ([#7655](https://github.com/rapidsai/cudf/pull/7655)) [@revans2](https://github.com/revans2) -- Convert cudf::concatenate APIs to use spans and device_uvector ([#7621](https://github.com/rapidsai/cudf/pull/7621)) [@harrism](https://github.com/harrism) -- Update missing docstring examples in python public APIs ([#7546](https://github.com/rapidsai/cudf/pull/7546)) [@galipremsagar](https://github.com/galipremsagar) -- Remove unneeded step parameter from strings::detail::copy_slice ([#7525](https://github.com/rapidsai/cudf/pull/7525)) [@davidwendt](https://github.com/davidwendt) -- Rename ARROW_STATIC_LIB because it conflicts with one in FindArrow.cmake ([#7518](https://github.com/rapidsai/cudf/pull/7518)) [@trxcllnt](https://github.com/trxcllnt) -- Match Pandas logic for comparing two objects with nulls ([#7490](https://github.com/rapidsai/cudf/pull/7490)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add struct support to parquet writer ([#7461](https://github.com/rapidsai/cudf/pull/7461)) [@devavret](https://github.com/devavret) -- Join APIs that return gathermaps ([#7454](https://github.com/rapidsai/cudf/pull/7454)) [@shwina](https://github.com/shwina) -- `fixed_point` + `cudf::binary_operation` API Changes ([#7435](https://github.com/rapidsai/cudf/pull/7435)) [@codereport](https://github.com/codereport) -- Fix BUG: Exception when PYTHONOPTIMIZE=2 ([#7434](https://github.com/rapidsai/cudf/pull/7434)) [@skirui-source](https://github.com/skirui-source) -- Change nvtext::load_vocabulary_file to return a unique ptr ([#7424](https://github.com/rapidsai/cudf/pull/7424)) [@davidwendt](https://github.com/davidwendt) -- Refactor strings column factories ([#7397](https://github.com/rapidsai/cudf/pull/7397)) [@harrism](https://github.com/harrism) -- Use CMAKE_CUDA_ARCHITECTURES ([#7391](https://github.com/rapidsai/cudf/pull/7391)) [@robertmaynard](https://github.com/robertmaynard) -- Upgrade pandas to 1.2 ([#7375](https://github.com/rapidsai/cudf/pull/7375)) [@galipremsagar](https://github.com/galipremsagar) -- Rename `logical_cast` to `bit_cast` and allow additional conversions ([#7373](https://github.com/rapidsai/cudf/pull/7373)) [@ttnghia](https://github.com/ttnghia) -- Rework libcudf CMakeLists.txt to export targets for CPM ([#7107](https://github.com/rapidsai/cudf/pull/7107)) [@trxcllnt](https://github.com/trxcllnt) +- Allow hash_partition to take a seed value ([#7771](https://github.com/NVIDIA/cudf/pull/7771)) [@magnatelee](https://github.com/magnatelee) +- Allow merging index column with data column using keyword "on" ([#7736](https://github.com/NVIDIA/cudf/pull/7736)) [@skirui-source](https://github.com/skirui-source) +- Change JNI API to avoid loading native dependencies when creating sort order classes. ([#7729](https://github.com/NVIDIA/cudf/pull/7729)) [@revans2](https://github.com/revans2) +- Replace device_vector with device_uvector in null_mask ([#7715](https://github.com/NVIDIA/cudf/pull/7715)) [@harrism](https://github.com/harrism) +- Don't identify decimals as strings. ([#7710](https://github.com/NVIDIA/cudf/pull/7710)) [@vyasr](https://github.com/vyasr) +- Fix Java Parquet write after writer API changes ([#7655](https://github.com/NVIDIA/cudf/pull/7655)) [@revans2](https://github.com/revans2) +- Convert cudf::concatenate APIs to use spans and device_uvector ([#7621](https://github.com/NVIDIA/cudf/pull/7621)) [@harrism](https://github.com/harrism) +- Update missing docstring examples in python public APIs ([#7546](https://github.com/NVIDIA/cudf/pull/7546)) [@galipremsagar](https://github.com/galipremsagar) +- Remove unneeded step parameter from strings::detail::copy_slice ([#7525](https://github.com/NVIDIA/cudf/pull/7525)) [@davidwendt](https://github.com/davidwendt) +- Rename ARROW_STATIC_LIB because it conflicts with one in FindArrow.cmake ([#7518](https://github.com/NVIDIA/cudf/pull/7518)) [@trxcllnt](https://github.com/trxcllnt) +- Match Pandas logic for comparing two objects with nulls ([#7490](https://github.com/NVIDIA/cudf/pull/7490)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add struct support to parquet writer ([#7461](https://github.com/NVIDIA/cudf/pull/7461)) [@devavret](https://github.com/devavret) +- Join APIs that return gathermaps ([#7454](https://github.com/NVIDIA/cudf/pull/7454)) [@shwina](https://github.com/shwina) +- `fixed_point` + `cudf::binary_operation` API Changes ([#7435](https://github.com/NVIDIA/cudf/pull/7435)) [@codereport](https://github.com/codereport) +- Fix BUG: Exception when PYTHONOPTIMIZE=2 ([#7434](https://github.com/NVIDIA/cudf/pull/7434)) [@skirui-source](https://github.com/skirui-source) +- Change nvtext::load_vocabulary_file to return a unique ptr ([#7424](https://github.com/NVIDIA/cudf/pull/7424)) [@davidwendt](https://github.com/davidwendt) +- Refactor strings column factories ([#7397](https://github.com/NVIDIA/cudf/pull/7397)) [@harrism](https://github.com/harrism) +- Use CMAKE_CUDA_ARCHITECTURES ([#7391](https://github.com/NVIDIA/cudf/pull/7391)) [@robertmaynard](https://github.com/robertmaynard) +- Upgrade pandas to 1.2 ([#7375](https://github.com/NVIDIA/cudf/pull/7375)) [@galipremsagar](https://github.com/galipremsagar) +- Rename `logical_cast` to `bit_cast` and allow additional conversions ([#7373](https://github.com/NVIDIA/cudf/pull/7373)) [@ttnghia](https://github.com/ttnghia) +- Rework libcudf CMakeLists.txt to export targets for CPM ([#7107](https://github.com/NVIDIA/cudf/pull/7107)) [@trxcllnt](https://github.com/trxcllnt) ## 🐛 Bug Fixes -- Fix a `NameError` in meta dispatch API ([#7996](https://github.com/rapidsai/cudf/pull/7996)) [@galipremsagar](https://github.com/galipremsagar) -- Reindex in `DataFrame.__setitem__` ([#7957](https://github.com/rapidsai/cudf/pull/7957)) [@galipremsagar](https://github.com/galipremsagar) -- jitify direct-to-cubin compilation and caching. ([#7919](https://github.com/rapidsai/cudf/pull/7919)) [@cwharris](https://github.com/cwharris) -- Use dynamic cudart for nvcomp in java build ([#7896](https://github.com/rapidsai/cudf/pull/7896)) [@abellina](https://github.com/abellina) -- fix "incompatible redefinition" warnings ([#7894](https://github.com/rapidsai/cudf/pull/7894)) [@cwharris](https://github.com/cwharris) -- cudf consistently specifies the cuda runtime ([#7887](https://github.com/rapidsai/cudf/pull/7887)) [@robertmaynard](https://github.com/robertmaynard) -- disable verbose output for jitify_preprocess ([#7886](https://github.com/rapidsai/cudf/pull/7886)) [@cwharris](https://github.com/cwharris) -- CMake jit_preprocess_files function only runs when needed ([#7872](https://github.com/rapidsai/cudf/pull/7872)) [@robertmaynard](https://github.com/robertmaynard) -- Push DeviceScalar construction into cython for list.contains ([#7864](https://github.com/rapidsai/cudf/pull/7864)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- cudf now sets an install rpath of $ORIGIN ([#7863](https://github.com/rapidsai/cudf/pull/7863)) [@robertmaynard](https://github.com/robertmaynard) -- Don't install Thrust examples, tests, docs, and python files ([#7811](https://github.com/rapidsai/cudf/pull/7811)) [@robertmaynard](https://github.com/robertmaynard) -- Sort by index in groupby tests more consistently ([#7802](https://github.com/rapidsai/cudf/pull/7802)) [@shwina](https://github.com/shwina) -- Revert "Update conda recipes pinning of repo dependencies ([#7743)" (#7793](https://github.com/rapidsai/cudf/pull/7743)" (#7793)) [@raydouglass](https://github.com/raydouglass) -- Add decimal column handling in copy_type_metadata ([#7788](https://github.com/rapidsai/cudf/pull/7788)) [@shwina](https://github.com/shwina) -- Add column names validation in parquet writer ([#7786](https://github.com/rapidsai/cudf/pull/7786)) [@galipremsagar](https://github.com/galipremsagar) -- Fix Java explode outer unit tests ([#7782](https://github.com/rapidsai/cudf/pull/7782)) [@jlowe](https://github.com/jlowe) -- Fix compiler warning about non-POD types passed through ellipsis ([#7781](https://github.com/rapidsai/cudf/pull/7781)) [@jrhemstad](https://github.com/jrhemstad) -- User resource fix for replace_nulls ([#7769](https://github.com/rapidsai/cudf/pull/7769)) [@magnatelee](https://github.com/magnatelee) -- Fix type dispatch for columnar replace_nulls ([#7768](https://github.com/rapidsai/cudf/pull/7768)) [@jlowe](https://github.com/jlowe) -- Add `ignore_order` parameter to dask-cudf concat dispatch ([#7765](https://github.com/rapidsai/cudf/pull/7765)) [@galipremsagar](https://github.com/galipremsagar) -- Fix slicing and arrow representations of decimal columns ([#7755](https://github.com/rapidsai/cudf/pull/7755)) [@vyasr](https://github.com/vyasr) -- Fixing issue with explode_outer position not nulling position entries of null rows ([#7754](https://github.com/rapidsai/cudf/pull/7754)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Implement scatter for struct columns ([#7752](https://github.com/rapidsai/cudf/pull/7752)) [@ttnghia](https://github.com/ttnghia) -- Fix data corruption in string columns ([#7746](https://github.com/rapidsai/cudf/pull/7746)) [@galipremsagar](https://github.com/galipremsagar) -- Fix string length in stripe dictionary building ([#7744](https://github.com/rapidsai/cudf/pull/7744)) [@kaatish](https://github.com/kaatish) -- Update conda recipes pinning of repo dependencies ([#7743](https://github.com/rapidsai/cudf/pull/7743)) [@mike-wendt](https://github.com/mike-wendt) -- Enable dask dispatch to cuDF's `is_categorical_dtype` for cuDF objects ([#7740](https://github.com/rapidsai/cudf/pull/7740)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix dictionary size computation in ORC writer ([#7737](https://github.com/rapidsai/cudf/pull/7737)) [@vuule](https://github.com/vuule) -- Fix `cudf::cast` overflow for `decimal64` to `int32_t` or smaller in certain cases ([#7733](https://github.com/rapidsai/cudf/pull/7733)) [@codereport](https://github.com/codereport) -- Change JNI API to avoid loading native dependencies when creating sort order classes. ([#7729](https://github.com/rapidsai/cudf/pull/7729)) [@revans2](https://github.com/revans2) -- Disable column_view data accessors for unsupported types ([#7725](https://github.com/rapidsai/cudf/pull/7725)) [@jrhemstad](https://github.com/jrhemstad) -- Materialize `RangeIndex` when `index=True` in parquet writer ([#7711](https://github.com/rapidsai/cudf/pull/7711)) [@galipremsagar](https://github.com/galipremsagar) -- Don't identify decimals as strings. ([#7710](https://github.com/rapidsai/cudf/pull/7710)) [@vyasr](https://github.com/vyasr) -- Fix return type of `DataFrame.argsort` ([#7706](https://github.com/rapidsai/cudf/pull/7706)) [@galipremsagar](https://github.com/galipremsagar) -- Fix/correct cudf installed package requirements ([#7688](https://github.com/rapidsai/cudf/pull/7688)) [@robertmaynard](https://github.com/robertmaynard) -- Fix SparkMurmurHash3_32 hash inconsistencies with Apache Spark ([#7672](https://github.com/rapidsai/cudf/pull/7672)) [@jlowe](https://github.com/jlowe) -- Fix ORC reader issue with reading empty string columns ([#7656](https://github.com/rapidsai/cudf/pull/7656)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Fix Java Parquet write after writer API changes ([#7655](https://github.com/rapidsai/cudf/pull/7655)) [@revans2](https://github.com/revans2) -- Fixing empty null lists throwing explode_outer for a loop. ([#7649](https://github.com/rapidsai/cudf/pull/7649)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Fix internal compiler error during JNI Docker build ([#7645](https://github.com/rapidsai/cudf/pull/7645)) [@jlowe](https://github.com/jlowe) -- Fix Debug build break with device_uvectors in grouped_rolling.cu ([#7633](https://github.com/rapidsai/cudf/pull/7633)) [@mythrocks](https://github.com/mythrocks) -- Parquet reader: Fix issue when using skip_rows on non-nested columns containing nulls ([#7627](https://github.com/rapidsai/cudf/pull/7627)) [@nvdbaranec](https://github.com/nvdbaranec) -- Fix ORC reader for empty DataFrame/Table ([#7624](https://github.com/rapidsai/cudf/pull/7624)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Fix specifying GPU architecture in JNI build ([#7612](https://github.com/rapidsai/cudf/pull/7612)) [@jlowe](https://github.com/jlowe) -- Fix ORC writer OOM issue ([#7605](https://github.com/rapidsai/cudf/pull/7605)) [@vuule](https://github.com/vuule) -- Fix 0.18 --> 0.19 automerge ([#7589](https://github.com/rapidsai/cudf/pull/7589)) [@kkraus14](https://github.com/kkraus14) -- Fix ORC issue with incorrect timestamp nanosecond values ([#7581](https://github.com/rapidsai/cudf/pull/7581)) [@vuule](https://github.com/vuule) -- Fix missing Dask imports ([#7580](https://github.com/rapidsai/cudf/pull/7580)) [@kkraus14](https://github.com/kkraus14) -- CMAKE_CUDA_ARCHITECTURES doesn't change when build-system invokes cmake ([#7579](https://github.com/rapidsai/cudf/pull/7579)) [@robertmaynard](https://github.com/robertmaynard) -- Another fix for offsets_end() iterator in lists_column_view ([#7575](https://github.com/rapidsai/cudf/pull/7575)) [@ttnghia](https://github.com/ttnghia) -- Fix ORC writer output corruption with string columns ([#7565](https://github.com/rapidsai/cudf/pull/7565)) [@vuule](https://github.com/vuule) -- Fix cudf::lists::sort_lists failing for sliced column ([#7564](https://github.com/rapidsai/cudf/pull/7564)) [@ttnghia](https://github.com/ttnghia) -- FIX Fix Anaconda upload args ([#7558](https://github.com/rapidsai/cudf/pull/7558)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- Fix index mismatch issue in equality related APIs ([#7555](https://github.com/rapidsai/cudf/pull/7555)) [@galipremsagar](https://github.com/galipremsagar) -- FIX Revert gpuci_conda_retry on conda file output locations ([#7552](https://github.com/rapidsai/cudf/pull/7552)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- Fix offset_end iterator for lists_column_view, which was not correctl… ([#7551](https://github.com/rapidsai/cudf/pull/7551)) [@ttnghia](https://github.com/ttnghia) -- Fix no such file dlpack.h error when build libcudf ([#7549](https://github.com/rapidsai/cudf/pull/7549)) [@chenrui17](https://github.com/chenrui17) -- Update missing docstring examples in python public APIs ([#7546](https://github.com/rapidsai/cudf/pull/7546)) [@galipremsagar](https://github.com/galipremsagar) -- Decimal32 Build Fix ([#7544](https://github.com/rapidsai/cudf/pull/7544)) [@razajafri](https://github.com/razajafri) -- FIX Retry conda output location ([#7540](https://github.com/rapidsai/cudf/pull/7540)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- fix missing renames of dask git branches from master to main ([#7535](https://github.com/rapidsai/cudf/pull/7535)) [@kkraus14](https://github.com/kkraus14) -- Remove detail from device_span ([#7533](https://github.com/rapidsai/cudf/pull/7533)) [@rwlee](https://github.com/rwlee) -- Change dask and distributed branch to main ([#7532](https://github.com/rapidsai/cudf/pull/7532)) [@dantegd](https://github.com/dantegd) -- Update JNI build to use CUDF_USE_ARROW_STATIC ([#7526](https://github.com/rapidsai/cudf/pull/7526)) [@jlowe](https://github.com/jlowe) -- Make sure rmm::rmm CMake target is visible to cudf users ([#7524](https://github.com/rapidsai/cudf/pull/7524)) [@robertmaynard](https://github.com/robertmaynard) -- Fix contiguous_split not properly handling output partitions > 2 GB. ([#7515](https://github.com/rapidsai/cudf/pull/7515)) [@nvdbaranec](https://github.com/nvdbaranec) -- Change jit launch to safe_launch ([#7510](https://github.com/rapidsai/cudf/pull/7510)) [@devavret](https://github.com/devavret) -- Fix comparison between Datetime/Timedelta columns and NULL scalars ([#7504](https://github.com/rapidsai/cudf/pull/7504)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix off-by-one error in char-parallel string scalar replace ([#7502](https://github.com/rapidsai/cudf/pull/7502)) [@jlowe](https://github.com/jlowe) -- Fix JNI deprecation of all, put it on the wrong version before ([#7501](https://github.com/rapidsai/cudf/pull/7501)) [@revans2](https://github.com/revans2) -- Fix Series/Dataframe Mixed Arithmetic ([#7491](https://github.com/rapidsai/cudf/pull/7491)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Fix JNI build after removal of libcudf sub-libraries ([#7486](https://github.com/rapidsai/cudf/pull/7486)) [@jlowe](https://github.com/jlowe) -- Correctly compile benchmarks ([#7485](https://github.com/rapidsai/cudf/pull/7485)) [@robertmaynard](https://github.com/robertmaynard) -- Fix bool column corruption with ORC Reader ([#7483](https://github.com/rapidsai/cudf/pull/7483)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Fix `__repr__` for categorical dtype ([#7476](https://github.com/rapidsai/cudf/pull/7476)) [@galipremsagar](https://github.com/galipremsagar) -- Java cleaner synchronization ([#7474](https://github.com/rapidsai/cudf/pull/7474)) [@abellina](https://github.com/abellina) -- Fix java float/double parsing tests ([#7473](https://github.com/rapidsai/cudf/pull/7473)) [@revans2](https://github.com/revans2) -- Pass stream and user resource to make_default_constructed_scalar ([#7469](https://github.com/rapidsai/cudf/pull/7469)) [@magnatelee](https://github.com/magnatelee) -- Improve stability of dask_cudf.DataFrame.var and dask_cudf.DataFrame.std ([#7453](https://github.com/rapidsai/cudf/pull/7453)) [@rjzamora](https://github.com/rjzamora) -- Missing `device_storage_dispatch` change affecting `cudf::gather` ([#7449](https://github.com/rapidsai/cudf/pull/7449)) [@codereport](https://github.com/codereport) -- fix cuFile JNI compile errors ([#7445](https://github.com/rapidsai/cudf/pull/7445)) [@rongou](https://github.com/rongou) -- Support `Series.__setitem__` with key to a new row ([#7443](https://github.com/rapidsai/cudf/pull/7443)) [@isVoid](https://github.com/isVoid) -- Fix BUG: Exception when PYTHONOPTIMIZE=2 ([#7434](https://github.com/rapidsai/cudf/pull/7434)) [@skirui-source](https://github.com/skirui-source) -- Make inclusive scan safe for cases with leading nulls ([#7432](https://github.com/rapidsai/cudf/pull/7432)) [@magnatelee](https://github.com/magnatelee) -- Fix typo in list_device_view::pair_rep_end() ([#7423](https://github.com/rapidsai/cudf/pull/7423)) [@mythrocks](https://github.com/mythrocks) -- Fix string to double conversion and row equivalent comparison ([#7410](https://github.com/rapidsai/cudf/pull/7410)) [@ttnghia](https://github.com/ttnghia) -- Fix thrust failure when transferring data from device_vector to host_vector with vectors of size 1 ([#7382](https://github.com/rapidsai/cudf/pull/7382)) [@ttnghia](https://github.com/ttnghia) -- Fix std::exception catch-by-reference gcc9 compile error ([#7380](https://github.com/rapidsai/cudf/pull/7380)) [@davidwendt](https://github.com/davidwendt) -- Fix skiprows issue with ORC Reader ([#7359](https://github.com/rapidsai/cudf/pull/7359)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- fix Arrow CMake file ([#7358](https://github.com/rapidsai/cudf/pull/7358)) [@rongou](https://github.com/rongou) -- Fix lists::contains() for NaN and Decimals ([#7349](https://github.com/rapidsai/cudf/pull/7349)) [@mythrocks](https://github.com/mythrocks) -- Handle cupy array in `Dataframe.__setitem__` ([#7340](https://github.com/rapidsai/cudf/pull/7340)) [@galipremsagar](https://github.com/galipremsagar) -- Fix invalid-device-fn error in cudf::strings::replace_re with multiple regex's ([#7336](https://github.com/rapidsai/cudf/pull/7336)) [@davidwendt](https://github.com/davidwendt) -- FIX Add codecov upload block to gpu script ([#6860](https://github.com/rapidsai/cudf/pull/6860)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- Fix a `NameError` in meta dispatch API ([#7996](https://github.com/NVIDIA/cudf/pull/7996)) [@galipremsagar](https://github.com/galipremsagar) +- Reindex in `DataFrame.__setitem__` ([#7957](https://github.com/NVIDIA/cudf/pull/7957)) [@galipremsagar](https://github.com/galipremsagar) +- jitify direct-to-cubin compilation and caching. ([#7919](https://github.com/NVIDIA/cudf/pull/7919)) [@cwharris](https://github.com/cwharris) +- Use dynamic cudart for nvcomp in java build ([#7896](https://github.com/NVIDIA/cudf/pull/7896)) [@abellina](https://github.com/abellina) +- fix "incompatible redefinition" warnings ([#7894](https://github.com/NVIDIA/cudf/pull/7894)) [@cwharris](https://github.com/cwharris) +- cudf consistently specifies the cuda runtime ([#7887](https://github.com/NVIDIA/cudf/pull/7887)) [@robertmaynard](https://github.com/robertmaynard) +- disable verbose output for jitify_preprocess ([#7886](https://github.com/NVIDIA/cudf/pull/7886)) [@cwharris](https://github.com/cwharris) +- CMake jit_preprocess_files function only runs when needed ([#7872](https://github.com/NVIDIA/cudf/pull/7872)) [@robertmaynard](https://github.com/robertmaynard) +- Push DeviceScalar construction into cython for list.contains ([#7864](https://github.com/NVIDIA/cudf/pull/7864)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- cudf now sets an install rpath of $ORIGIN ([#7863](https://github.com/NVIDIA/cudf/pull/7863)) [@robertmaynard](https://github.com/robertmaynard) +- Don't install Thrust examples, tests, docs, and python files ([#7811](https://github.com/NVIDIA/cudf/pull/7811)) [@robertmaynard](https://github.com/robertmaynard) +- Sort by index in groupby tests more consistently ([#7802](https://github.com/NVIDIA/cudf/pull/7802)) [@shwina](https://github.com/shwina) +- Revert "Update conda recipes pinning of repo dependencies ([#7743)" (#7793](https://github.com/NVIDIA/cudf/pull/7743)" (#7793)) [@raydouglass](https://github.com/raydouglass) +- Add decimal column handling in copy_type_metadata ([#7788](https://github.com/NVIDIA/cudf/pull/7788)) [@shwina](https://github.com/shwina) +- Add column names validation in parquet writer ([#7786](https://github.com/NVIDIA/cudf/pull/7786)) [@galipremsagar](https://github.com/galipremsagar) +- Fix Java explode outer unit tests ([#7782](https://github.com/NVIDIA/cudf/pull/7782)) [@jlowe](https://github.com/jlowe) +- Fix compiler warning about non-POD types passed through ellipsis ([#7781](https://github.com/NVIDIA/cudf/pull/7781)) [@jrhemstad](https://github.com/jrhemstad) +- User resource fix for replace_nulls ([#7769](https://github.com/NVIDIA/cudf/pull/7769)) [@magnatelee](https://github.com/magnatelee) +- Fix type dispatch for columnar replace_nulls ([#7768](https://github.com/NVIDIA/cudf/pull/7768)) [@jlowe](https://github.com/jlowe) +- Add `ignore_order` parameter to dask-cudf concat dispatch ([#7765](https://github.com/NVIDIA/cudf/pull/7765)) [@galipremsagar](https://github.com/galipremsagar) +- Fix slicing and arrow representations of decimal columns ([#7755](https://github.com/NVIDIA/cudf/pull/7755)) [@vyasr](https://github.com/vyasr) +- Fixing issue with explode_outer position not nulling position entries of null rows ([#7754](https://github.com/NVIDIA/cudf/pull/7754)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Implement scatter for struct columns ([#7752](https://github.com/NVIDIA/cudf/pull/7752)) [@ttnghia](https://github.com/ttnghia) +- Fix data corruption in string columns ([#7746](https://github.com/NVIDIA/cudf/pull/7746)) [@galipremsagar](https://github.com/galipremsagar) +- Fix string length in stripe dictionary building ([#7744](https://github.com/NVIDIA/cudf/pull/7744)) [@kaatish](https://github.com/kaatish) +- Update conda recipes pinning of repo dependencies ([#7743](https://github.com/NVIDIA/cudf/pull/7743)) [@mike-wendt](https://github.com/mike-wendt) +- Enable dask dispatch to cuDF's `is_categorical_dtype` for cuDF objects ([#7740](https://github.com/NVIDIA/cudf/pull/7740)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix dictionary size computation in ORC writer ([#7737](https://github.com/NVIDIA/cudf/pull/7737)) [@vuule](https://github.com/vuule) +- Fix `cudf::cast` overflow for `decimal64` to `int32_t` or smaller in certain cases ([#7733](https://github.com/NVIDIA/cudf/pull/7733)) [@codereport](https://github.com/codereport) +- Change JNI API to avoid loading native dependencies when creating sort order classes. ([#7729](https://github.com/NVIDIA/cudf/pull/7729)) [@revans2](https://github.com/revans2) +- Disable column_view data accessors for unsupported types ([#7725](https://github.com/NVIDIA/cudf/pull/7725)) [@jrhemstad](https://github.com/jrhemstad) +- Materialize `RangeIndex` when `index=True` in parquet writer ([#7711](https://github.com/NVIDIA/cudf/pull/7711)) [@galipremsagar](https://github.com/galipremsagar) +- Don't identify decimals as strings. ([#7710](https://github.com/NVIDIA/cudf/pull/7710)) [@vyasr](https://github.com/vyasr) +- Fix return type of `DataFrame.argsort` ([#7706](https://github.com/NVIDIA/cudf/pull/7706)) [@galipremsagar](https://github.com/galipremsagar) +- Fix/correct cudf installed package requirements ([#7688](https://github.com/NVIDIA/cudf/pull/7688)) [@robertmaynard](https://github.com/robertmaynard) +- Fix SparkMurmurHash3_32 hash inconsistencies with Apache Spark ([#7672](https://github.com/NVIDIA/cudf/pull/7672)) [@jlowe](https://github.com/jlowe) +- Fix ORC reader issue with reading empty string columns ([#7656](https://github.com/NVIDIA/cudf/pull/7656)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Fix Java Parquet write after writer API changes ([#7655](https://github.com/NVIDIA/cudf/pull/7655)) [@revans2](https://github.com/revans2) +- Fixing empty null lists throwing explode_outer for a loop. ([#7649](https://github.com/NVIDIA/cudf/pull/7649)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Fix internal compiler error during JNI Docker build ([#7645](https://github.com/NVIDIA/cudf/pull/7645)) [@jlowe](https://github.com/jlowe) +- Fix Debug build break with device_uvectors in grouped_rolling.cu ([#7633](https://github.com/NVIDIA/cudf/pull/7633)) [@mythrocks](https://github.com/mythrocks) +- Parquet reader: Fix issue when using skip_rows on non-nested columns containing nulls ([#7627](https://github.com/NVIDIA/cudf/pull/7627)) [@nvdbaranec](https://github.com/nvdbaranec) +- Fix ORC reader for empty DataFrame/Table ([#7624](https://github.com/NVIDIA/cudf/pull/7624)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Fix specifying GPU architecture in JNI build ([#7612](https://github.com/NVIDIA/cudf/pull/7612)) [@jlowe](https://github.com/jlowe) +- Fix ORC writer OOM issue ([#7605](https://github.com/NVIDIA/cudf/pull/7605)) [@vuule](https://github.com/vuule) +- Fix 0.18 --> 0.19 automerge ([#7589](https://github.com/NVIDIA/cudf/pull/7589)) [@kkraus14](https://github.com/kkraus14) +- Fix ORC issue with incorrect timestamp nanosecond values ([#7581](https://github.com/NVIDIA/cudf/pull/7581)) [@vuule](https://github.com/vuule) +- Fix missing Dask imports ([#7580](https://github.com/NVIDIA/cudf/pull/7580)) [@kkraus14](https://github.com/kkraus14) +- CMAKE_CUDA_ARCHITECTURES doesn't change when build-system invokes cmake ([#7579](https://github.com/NVIDIA/cudf/pull/7579)) [@robertmaynard](https://github.com/robertmaynard) +- Another fix for offsets_end() iterator in lists_column_view ([#7575](https://github.com/NVIDIA/cudf/pull/7575)) [@ttnghia](https://github.com/ttnghia) +- Fix ORC writer output corruption with string columns ([#7565](https://github.com/NVIDIA/cudf/pull/7565)) [@vuule](https://github.com/vuule) +- Fix cudf::lists::sort_lists failing for sliced column ([#7564](https://github.com/NVIDIA/cudf/pull/7564)) [@ttnghia](https://github.com/ttnghia) +- FIX Fix Anaconda upload args ([#7558](https://github.com/NVIDIA/cudf/pull/7558)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- Fix index mismatch issue in equality related APIs ([#7555](https://github.com/NVIDIA/cudf/pull/7555)) [@galipremsagar](https://github.com/galipremsagar) +- FIX Revert gpuci_conda_retry on conda file output locations ([#7552](https://github.com/NVIDIA/cudf/pull/7552)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- Fix offset_end iterator for lists_column_view, which was not correctl… ([#7551](https://github.com/NVIDIA/cudf/pull/7551)) [@ttnghia](https://github.com/ttnghia) +- Fix no such file dlpack.h error when build libcudf ([#7549](https://github.com/NVIDIA/cudf/pull/7549)) [@chenrui17](https://github.com/chenrui17) +- Update missing docstring examples in python public APIs ([#7546](https://github.com/NVIDIA/cudf/pull/7546)) [@galipremsagar](https://github.com/galipremsagar) +- Decimal32 Build Fix ([#7544](https://github.com/NVIDIA/cudf/pull/7544)) [@razajafri](https://github.com/razajafri) +- FIX Retry conda output location ([#7540](https://github.com/NVIDIA/cudf/pull/7540)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- fix missing renames of dask git branches from master to main ([#7535](https://github.com/NVIDIA/cudf/pull/7535)) [@kkraus14](https://github.com/kkraus14) +- Remove detail from device_span ([#7533](https://github.com/NVIDIA/cudf/pull/7533)) [@rwlee](https://github.com/rwlee) +- Change dask and distributed branch to main ([#7532](https://github.com/NVIDIA/cudf/pull/7532)) [@dantegd](https://github.com/dantegd) +- Update JNI build to use CUDF_USE_ARROW_STATIC ([#7526](https://github.com/NVIDIA/cudf/pull/7526)) [@jlowe](https://github.com/jlowe) +- Make sure rmm::rmm CMake target is visible to cudf users ([#7524](https://github.com/NVIDIA/cudf/pull/7524)) [@robertmaynard](https://github.com/robertmaynard) +- Fix contiguous_split not properly handling output partitions > 2 GB. ([#7515](https://github.com/NVIDIA/cudf/pull/7515)) [@nvdbaranec](https://github.com/nvdbaranec) +- Change jit launch to safe_launch ([#7510](https://github.com/NVIDIA/cudf/pull/7510)) [@devavret](https://github.com/devavret) +- Fix comparison between Datetime/Timedelta columns and NULL scalars ([#7504](https://github.com/NVIDIA/cudf/pull/7504)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix off-by-one error in char-parallel string scalar replace ([#7502](https://github.com/NVIDIA/cudf/pull/7502)) [@jlowe](https://github.com/jlowe) +- Fix JNI deprecation of all, put it on the wrong version before ([#7501](https://github.com/NVIDIA/cudf/pull/7501)) [@revans2](https://github.com/revans2) +- Fix Series/Dataframe Mixed Arithmetic ([#7491](https://github.com/NVIDIA/cudf/pull/7491)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Fix JNI build after removal of libcudf sub-libraries ([#7486](https://github.com/NVIDIA/cudf/pull/7486)) [@jlowe](https://github.com/jlowe) +- Correctly compile benchmarks ([#7485](https://github.com/NVIDIA/cudf/pull/7485)) [@robertmaynard](https://github.com/robertmaynard) +- Fix bool column corruption with ORC Reader ([#7483](https://github.com/NVIDIA/cudf/pull/7483)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Fix `__repr__` for categorical dtype ([#7476](https://github.com/NVIDIA/cudf/pull/7476)) [@galipremsagar](https://github.com/galipremsagar) +- Java cleaner synchronization ([#7474](https://github.com/NVIDIA/cudf/pull/7474)) [@abellina](https://github.com/abellina) +- Fix java float/double parsing tests ([#7473](https://github.com/NVIDIA/cudf/pull/7473)) [@revans2](https://github.com/revans2) +- Pass stream and user resource to make_default_constructed_scalar ([#7469](https://github.com/NVIDIA/cudf/pull/7469)) [@magnatelee](https://github.com/magnatelee) +- Improve stability of dask_cudf.DataFrame.var and dask_cudf.DataFrame.std ([#7453](https://github.com/NVIDIA/cudf/pull/7453)) [@rjzamora](https://github.com/rjzamora) +- Missing `device_storage_dispatch` change affecting `cudf::gather` ([#7449](https://github.com/NVIDIA/cudf/pull/7449)) [@codereport](https://github.com/codereport) +- fix cuFile JNI compile errors ([#7445](https://github.com/NVIDIA/cudf/pull/7445)) [@rongou](https://github.com/rongou) +- Support `Series.__setitem__` with key to a new row ([#7443](https://github.com/NVIDIA/cudf/pull/7443)) [@isVoid](https://github.com/isVoid) +- Fix BUG: Exception when PYTHONOPTIMIZE=2 ([#7434](https://github.com/NVIDIA/cudf/pull/7434)) [@skirui-source](https://github.com/skirui-source) +- Make inclusive scan safe for cases with leading nulls ([#7432](https://github.com/NVIDIA/cudf/pull/7432)) [@magnatelee](https://github.com/magnatelee) +- Fix typo in list_device_view::pair_rep_end() ([#7423](https://github.com/NVIDIA/cudf/pull/7423)) [@mythrocks](https://github.com/mythrocks) +- Fix string to double conversion and row equivalent comparison ([#7410](https://github.com/NVIDIA/cudf/pull/7410)) [@ttnghia](https://github.com/ttnghia) +- Fix thrust failure when transferring data from device_vector to host_vector with vectors of size 1 ([#7382](https://github.com/NVIDIA/cudf/pull/7382)) [@ttnghia](https://github.com/ttnghia) +- Fix std::exception catch-by-reference gcc9 compile error ([#7380](https://github.com/NVIDIA/cudf/pull/7380)) [@davidwendt](https://github.com/davidwendt) +- Fix skiprows issue with ORC Reader ([#7359](https://github.com/NVIDIA/cudf/pull/7359)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- fix Arrow CMake file ([#7358](https://github.com/NVIDIA/cudf/pull/7358)) [@rongou](https://github.com/rongou) +- Fix lists::contains() for NaN and Decimals ([#7349](https://github.com/NVIDIA/cudf/pull/7349)) [@mythrocks](https://github.com/mythrocks) +- Handle cupy array in `Dataframe.__setitem__` ([#7340](https://github.com/NVIDIA/cudf/pull/7340)) [@galipremsagar](https://github.com/galipremsagar) +- Fix invalid-device-fn error in cudf::strings::replace_re with multiple regex's ([#7336](https://github.com/NVIDIA/cudf/pull/7336)) [@davidwendt](https://github.com/davidwendt) +- FIX Add codecov upload block to gpu script ([#6860](https://github.com/NVIDIA/cudf/pull/6860)) [@dillon-cullinan](https://github.com/dillon-cullinan) ## 📖 Documentation -- Fix join API doxygen ([#7890](https://github.com/rapidsai/cudf/pull/7890)) [@shwina](https://github.com/shwina) -- Add Resources to README. ([#7697](https://github.com/rapidsai/cudf/pull/7697)) [@bdice](https://github.com/bdice) -- Add `isin` examples in Docstring ([#7479](https://github.com/rapidsai/cudf/pull/7479)) [@galipremsagar](https://github.com/galipremsagar) -- Resolving unlinked type shorthands in cudf doc ([#7416](https://github.com/rapidsai/cudf/pull/7416)) [@isVoid](https://github.com/isVoid) -- Fix typo in regex.md doc page ([#7363](https://github.com/rapidsai/cudf/pull/7363)) [@davidwendt](https://github.com/davidwendt) -- Fix incorrect strings_column_view::chars_size documentation ([#7360](https://github.com/rapidsai/cudf/pull/7360)) [@jlowe](https://github.com/jlowe) +- Fix join API doxygen ([#7890](https://github.com/NVIDIA/cudf/pull/7890)) [@shwina](https://github.com/shwina) +- Add Resources to README. ([#7697](https://github.com/NVIDIA/cudf/pull/7697)) [@bdice](https://github.com/bdice) +- Add `isin` examples in Docstring ([#7479](https://github.com/NVIDIA/cudf/pull/7479)) [@galipremsagar](https://github.com/galipremsagar) +- Resolving unlinked type shorthands in cudf doc ([#7416](https://github.com/NVIDIA/cudf/pull/7416)) [@isVoid](https://github.com/isVoid) +- Fix typo in regex.md doc page ([#7363](https://github.com/NVIDIA/cudf/pull/7363)) [@davidwendt](https://github.com/davidwendt) +- Fix incorrect strings_column_view::chars_size documentation ([#7360](https://github.com/NVIDIA/cudf/pull/7360)) [@jlowe](https://github.com/jlowe) ## 🚀 New Features -- Enable basic reductions for decimal columns ([#7776](https://github.com/rapidsai/cudf/pull/7776)) [@ChrisJar](https://github.com/ChrisJar) -- Enable join on decimal columns ([#7764](https://github.com/rapidsai/cudf/pull/7764)) [@ChrisJar](https://github.com/ChrisJar) -- Allow merging index column with data column using keyword "on" ([#7736](https://github.com/rapidsai/cudf/pull/7736)) [@skirui-source](https://github.com/skirui-source) -- Implement DecimalColumn + Scalar and add cudf.Scalars of Decimal64Dtype ([#7732](https://github.com/rapidsai/cudf/pull/7732)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add support for `unique` groupby aggregation ([#7726](https://github.com/rapidsai/cudf/pull/7726)) [@shwina](https://github.com/shwina) -- Expose libcudf's label_bins function to cudf ([#7724](https://github.com/rapidsai/cudf/pull/7724)) [@vyasr](https://github.com/vyasr) -- Adding support for equi-join on struct ([#7720](https://github.com/rapidsai/cudf/pull/7720)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add decimal column comparison operations ([#7716](https://github.com/rapidsai/cudf/pull/7716)) [@isVoid](https://github.com/isVoid) -- Implement scan operations for decimal columns ([#7707](https://github.com/rapidsai/cudf/pull/7707)) [@ChrisJar](https://github.com/ChrisJar) -- Enable typecasting between decimal and int ([#7691](https://github.com/rapidsai/cudf/pull/7691)) [@ChrisJar](https://github.com/ChrisJar) -- Enable decimal support in parquet writer ([#7673](https://github.com/rapidsai/cudf/pull/7673)) [@devavret](https://github.com/devavret) -- Adds `list.unique` API ([#7664](https://github.com/rapidsai/cudf/pull/7664)) [@isVoid](https://github.com/isVoid) -- Fix NaN handling in drop_list_duplicates ([#7662](https://github.com/rapidsai/cudf/pull/7662)) [@ttnghia](https://github.com/ttnghia) -- Add `lists.sort_values` API ([#7657](https://github.com/rapidsai/cudf/pull/7657)) [@isVoid](https://github.com/isVoid) -- Add is_integer API that can check for the validity of a string-to-integer conversion ([#7642](https://github.com/rapidsai/cudf/pull/7642)) [@ttnghia](https://github.com/ttnghia) -- Adds `explode` API ([#7607](https://github.com/rapidsai/cudf/pull/7607)) [@isVoid](https://github.com/isVoid) -- Adds `list.take`, python binding for `cudf::lists::segmented_gather` ([#7591](https://github.com/rapidsai/cudf/pull/7591)) [@isVoid](https://github.com/isVoid) -- Implement cudf::label_bins() ([#7554](https://github.com/rapidsai/cudf/pull/7554)) [@vyasr](https://github.com/vyasr) -- Add Python bindings for `lists::contains` ([#7547](https://github.com/rapidsai/cudf/pull/7547)) [@skirui-source](https://github.com/skirui-source) -- cudf::row_bit_count() support. ([#7534](https://github.com/rapidsai/cudf/pull/7534)) [@nvdbaranec](https://github.com/nvdbaranec) -- Implement drop_list_duplicates ([#7528](https://github.com/rapidsai/cudf/pull/7528)) [@ttnghia](https://github.com/ttnghia) -- Add Python bindings for `lists::extract_lists_element` ([#7505](https://github.com/rapidsai/cudf/pull/7505)) [@skirui-source](https://github.com/skirui-source) -- Add explode_outer and explode_outer_position ([#7499](https://github.com/rapidsai/cudf/pull/7499)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Match Pandas logic for comparing two objects with nulls ([#7490](https://github.com/rapidsai/cudf/pull/7490)) [@brandon-b-miller](https://github.com/brandon-b-miller) -- Add struct support to parquet writer ([#7461](https://github.com/rapidsai/cudf/pull/7461)) [@devavret](https://github.com/devavret) -- Enable type conversion from float to decimal type ([#7450](https://github.com/rapidsai/cudf/pull/7450)) [@ChrisJar](https://github.com/ChrisJar) -- Add cython for converting strings/fixed-point functions ([#7429](https://github.com/rapidsai/cudf/pull/7429)) [@davidwendt](https://github.com/davidwendt) -- Add struct column support to cudf::sort and cudf::sorted_order ([#7422](https://github.com/rapidsai/cudf/pull/7422)) [@karthikeyann](https://github.com/karthikeyann) -- Implement groupby collect_set ([#7420](https://github.com/rapidsai/cudf/pull/7420)) [@ttnghia](https://github.com/ttnghia) -- Merge branch-0.18 into branch-0.19 ([#7411](https://github.com/rapidsai/cudf/pull/7411)) [@raydouglass](https://github.com/raydouglass) -- Refactor strings column factories ([#7397](https://github.com/rapidsai/cudf/pull/7397)) [@harrism](https://github.com/harrism) -- Add groupby scan operations (sort groupby) ([#7387](https://github.com/rapidsai/cudf/pull/7387)) [@karthikeyann](https://github.com/karthikeyann) -- Add cudf::explode_position ([#7376](https://github.com/rapidsai/cudf/pull/7376)) [@hyperbolic2346](https://github.com/hyperbolic2346) -- Add string conversion to/from decimal values libcudf APIs ([#7364](https://github.com/rapidsai/cudf/pull/7364)) [@davidwendt](https://github.com/davidwendt) -- Add groupby SUM_OF_SQUARES support ([#7362](https://github.com/rapidsai/cudf/pull/7362)) [@karthikeyann](https://github.com/karthikeyann) -- Add `Series.drop` api ([#7304](https://github.com/rapidsai/cudf/pull/7304)) [@isVoid](https://github.com/isVoid) -- get_json_object() implementation ([#7286](https://github.com/rapidsai/cudf/pull/7286)) [@nvdbaranec](https://github.com/nvdbaranec) -- Python API for `LIstMethods.len()` ([#7283](https://github.com/rapidsai/cudf/pull/7283)) [@isVoid](https://github.com/isVoid) -- Support null_policy::EXCLUDE for COLLECT rolling aggregation ([#7264](https://github.com/rapidsai/cudf/pull/7264)) [@mythrocks](https://github.com/mythrocks) -- Add support for special tokens in nvtext::subword_tokenizer ([#7254](https://github.com/rapidsai/cudf/pull/7254)) [@davidwendt](https://github.com/davidwendt) -- Fix inplace update of data and add Series.update ([#7201](https://github.com/rapidsai/cudf/pull/7201)) [@galipremsagar](https://github.com/galipremsagar) -- Implement `cudf::group_by` (hash) for `decimal32` and `decimal64` ([#7190](https://github.com/rapidsai/cudf/pull/7190)) [@codereport](https://github.com/codereport) -- Adding support to specify "level" parameter for `Dataframe.rename` ([#7135](https://github.com/rapidsai/cudf/pull/7135)) [@skirui-source](https://github.com/skirui-source) +- Enable basic reductions for decimal columns ([#7776](https://github.com/NVIDIA/cudf/pull/7776)) [@ChrisJar](https://github.com/ChrisJar) +- Enable join on decimal columns ([#7764](https://github.com/NVIDIA/cudf/pull/7764)) [@ChrisJar](https://github.com/ChrisJar) +- Allow merging index column with data column using keyword "on" ([#7736](https://github.com/NVIDIA/cudf/pull/7736)) [@skirui-source](https://github.com/skirui-source) +- Implement DecimalColumn + Scalar and add cudf.Scalars of Decimal64Dtype ([#7732](https://github.com/NVIDIA/cudf/pull/7732)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add support for `unique` groupby aggregation ([#7726](https://github.com/NVIDIA/cudf/pull/7726)) [@shwina](https://github.com/shwina) +- Expose libcudf's label_bins function to cudf ([#7724](https://github.com/NVIDIA/cudf/pull/7724)) [@vyasr](https://github.com/vyasr) +- Adding support for equi-join on struct ([#7720](https://github.com/NVIDIA/cudf/pull/7720)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add decimal column comparison operations ([#7716](https://github.com/NVIDIA/cudf/pull/7716)) [@isVoid](https://github.com/isVoid) +- Implement scan operations for decimal columns ([#7707](https://github.com/NVIDIA/cudf/pull/7707)) [@ChrisJar](https://github.com/ChrisJar) +- Enable typecasting between decimal and int ([#7691](https://github.com/NVIDIA/cudf/pull/7691)) [@ChrisJar](https://github.com/ChrisJar) +- Enable decimal support in parquet writer ([#7673](https://github.com/NVIDIA/cudf/pull/7673)) [@devavret](https://github.com/devavret) +- Adds `list.unique` API ([#7664](https://github.com/NVIDIA/cudf/pull/7664)) [@isVoid](https://github.com/isVoid) +- Fix NaN handling in drop_list_duplicates ([#7662](https://github.com/NVIDIA/cudf/pull/7662)) [@ttnghia](https://github.com/ttnghia) +- Add `lists.sort_values` API ([#7657](https://github.com/NVIDIA/cudf/pull/7657)) [@isVoid](https://github.com/isVoid) +- Add is_integer API that can check for the validity of a string-to-integer conversion ([#7642](https://github.com/NVIDIA/cudf/pull/7642)) [@ttnghia](https://github.com/ttnghia) +- Adds `explode` API ([#7607](https://github.com/NVIDIA/cudf/pull/7607)) [@isVoid](https://github.com/isVoid) +- Adds `list.take`, python binding for `cudf::lists::segmented_gather` ([#7591](https://github.com/NVIDIA/cudf/pull/7591)) [@isVoid](https://github.com/isVoid) +- Implement cudf::label_bins() ([#7554](https://github.com/NVIDIA/cudf/pull/7554)) [@vyasr](https://github.com/vyasr) +- Add Python bindings for `lists::contains` ([#7547](https://github.com/NVIDIA/cudf/pull/7547)) [@skirui-source](https://github.com/skirui-source) +- cudf::row_bit_count() support. ([#7534](https://github.com/NVIDIA/cudf/pull/7534)) [@nvdbaranec](https://github.com/nvdbaranec) +- Implement drop_list_duplicates ([#7528](https://github.com/NVIDIA/cudf/pull/7528)) [@ttnghia](https://github.com/ttnghia) +- Add Python bindings for `lists::extract_lists_element` ([#7505](https://github.com/NVIDIA/cudf/pull/7505)) [@skirui-source](https://github.com/skirui-source) +- Add explode_outer and explode_outer_position ([#7499](https://github.com/NVIDIA/cudf/pull/7499)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Match Pandas logic for comparing two objects with nulls ([#7490](https://github.com/NVIDIA/cudf/pull/7490)) [@brandon-b-miller](https://github.com/brandon-b-miller) +- Add struct support to parquet writer ([#7461](https://github.com/NVIDIA/cudf/pull/7461)) [@devavret](https://github.com/devavret) +- Enable type conversion from float to decimal type ([#7450](https://github.com/NVIDIA/cudf/pull/7450)) [@ChrisJar](https://github.com/ChrisJar) +- Add cython for converting strings/fixed-point functions ([#7429](https://github.com/NVIDIA/cudf/pull/7429)) [@davidwendt](https://github.com/davidwendt) +- Add struct column support to cudf::sort and cudf::sorted_order ([#7422](https://github.com/NVIDIA/cudf/pull/7422)) [@karthikeyann](https://github.com/karthikeyann) +- Implement groupby collect_set ([#7420](https://github.com/NVIDIA/cudf/pull/7420)) [@ttnghia](https://github.com/ttnghia) +- Merge branch-0.18 into branch-0.19 ([#7411](https://github.com/NVIDIA/cudf/pull/7411)) [@raydouglass](https://github.com/raydouglass) +- Refactor strings column factories ([#7397](https://github.com/NVIDIA/cudf/pull/7397)) [@harrism](https://github.com/harrism) +- Add groupby scan operations (sort groupby) ([#7387](https://github.com/NVIDIA/cudf/pull/7387)) [@karthikeyann](https://github.com/karthikeyann) +- Add cudf::explode_position ([#7376](https://github.com/NVIDIA/cudf/pull/7376)) [@hyperbolic2346](https://github.com/hyperbolic2346) +- Add string conversion to/from decimal values libcudf APIs ([#7364](https://github.com/NVIDIA/cudf/pull/7364)) [@davidwendt](https://github.com/davidwendt) +- Add groupby SUM_OF_SQUARES support ([#7362](https://github.com/NVIDIA/cudf/pull/7362)) [@karthikeyann](https://github.com/karthikeyann) +- Add `Series.drop` api ([#7304](https://github.com/NVIDIA/cudf/pull/7304)) [@isVoid](https://github.com/isVoid) +- get_json_object() implementation ([#7286](https://github.com/NVIDIA/cudf/pull/7286)) [@nvdbaranec](https://github.com/nvdbaranec) +- Python API for `LIstMethods.len()` ([#7283](https://github.com/NVIDIA/cudf/pull/7283)) [@isVoid](https://github.com/isVoid) +- Support null_policy::EXCLUDE for COLLECT rolling aggregation ([#7264](https://github.com/NVIDIA/cudf/pull/7264)) [@mythrocks](https://github.com/mythrocks) +- Add support for special tokens in nvtext::subword_tokenizer ([#7254](https://github.com/NVIDIA/cudf/pull/7254)) [@davidwendt](https://github.com/davidwendt) +- Fix inplace update of data and add Series.update ([#7201](https://github.com/NVIDIA/cudf/pull/7201)) [@galipremsagar](https://github.com/galipremsagar) +- Implement `cudf::group_by` (hash) for `decimal32` and `decimal64` ([#7190](https://github.com/NVIDIA/cudf/pull/7190)) [@codereport](https://github.com/codereport) +- Adding support to specify "level" parameter for `Dataframe.rename` ([#7135](https://github.com/NVIDIA/cudf/pull/7135)) [@skirui-source](https://github.com/skirui-source) ## 🛠️ Improvements -- fix GDS include path for version 0.95 ([#7877](https://github.com/rapidsai/cudf/pull/7877)) [@rongou](https://github.com/rongou) -- Update `dask` + `distributed` to `2021.4.0` ([#7858](https://github.com/rapidsai/cudf/pull/7858)) [@jakirkham](https://github.com/jakirkham) -- Add ability to extract include dirs from `CUDF_HOME` ([#7848](https://github.com/rapidsai/cudf/pull/7848)) [@galipremsagar](https://github.com/galipremsagar) -- Add USE_GDS as an option in build script ([#7833](https://github.com/rapidsai/cudf/pull/7833)) [@pxLi](https://github.com/pxLi) -- add an allocate method with stream in java DeviceMemoryBuffer ([#7826](https://github.com/rapidsai/cudf/pull/7826)) [@rongou](https://github.com/rongou) -- Constrain dask and distributed versions to 2021.3.1 ([#7825](https://github.com/rapidsai/cudf/pull/7825)) [@shwina](https://github.com/shwina) -- Revert dask versioning of concat dispatch ([#7823](https://github.com/rapidsai/cudf/pull/7823)) [@galipremsagar](https://github.com/galipremsagar) -- add copy methods in Java memory buffer ([#7791](https://github.com/rapidsai/cudf/pull/7791)) [@rongou](https://github.com/rongou) -- Update README and CONTRIBUTING for 0.19 ([#7778](https://github.com/rapidsai/cudf/pull/7778)) [@robertmaynard](https://github.com/robertmaynard) -- Allow hash_partition to take a seed value ([#7771](https://github.com/rapidsai/cudf/pull/7771)) [@magnatelee](https://github.com/magnatelee) -- Turn on NVTX by default in java build ([#7761](https://github.com/rapidsai/cudf/pull/7761)) [@tgravescs](https://github.com/tgravescs) -- Add Java bindings to join gather map APIs ([#7751](https://github.com/rapidsai/cudf/pull/7751)) [@jlowe](https://github.com/jlowe) -- Add replacements column support for Java replaceNulls ([#7750](https://github.com/rapidsai/cudf/pull/7750)) [@jlowe](https://github.com/jlowe) -- Add Java bindings for row_bit_count ([#7749](https://github.com/rapidsai/cudf/pull/7749)) [@jlowe](https://github.com/jlowe) -- Remove unused JVM array creation ([#7748](https://github.com/rapidsai/cudf/pull/7748)) [@jlowe](https://github.com/jlowe) -- Added JNI support for new is_integer ([#7739](https://github.com/rapidsai/cudf/pull/7739)) [@revans2](https://github.com/revans2) -- Create and promote library aliases in libcudf installations ([#7734](https://github.com/rapidsai/cudf/pull/7734)) [@trxcllnt](https://github.com/trxcllnt) -- Support groupby operations for decimal dtypes ([#7731](https://github.com/rapidsai/cudf/pull/7731)) [@vyasr](https://github.com/vyasr) -- Memory map the input file only when GDS compatibility mode is not used ([#7717](https://github.com/rapidsai/cudf/pull/7717)) [@vuule](https://github.com/vuule) -- Replace device_vector with device_uvector in null_mask ([#7715](https://github.com/rapidsai/cudf/pull/7715)) [@harrism](https://github.com/harrism) -- Struct hashing support for SerialMurmur3 and SparkMurmur3 ([#7714](https://github.com/rapidsai/cudf/pull/7714)) [@jlowe](https://github.com/jlowe) -- Add gbenchmark for nvtext replace-tokens function ([#7708](https://github.com/rapidsai/cudf/pull/7708)) [@davidwendt](https://github.com/davidwendt) -- Use stream in groupby calls ([#7705](https://github.com/rapidsai/cudf/pull/7705)) [@karthikeyann](https://github.com/karthikeyann) -- Update codeowners file ([#7701](https://github.com/rapidsai/cudf/pull/7701)) [@ajschmidt8](https://github.com/ajschmidt8) -- Cleanup groupby to use host_span, device_span, device_uvector ([#7698](https://github.com/rapidsai/cudf/pull/7698)) [@karthikeyann](https://github.com/karthikeyann) -- Add gbenchmark for nvtext ngrams functions ([#7693](https://github.com/rapidsai/cudf/pull/7693)) [@davidwendt](https://github.com/davidwendt) -- Misc Python/Cython optimizations ([#7686](https://github.com/rapidsai/cudf/pull/7686)) [@shwina](https://github.com/shwina) -- Add gbenchmark for nvtext tokenize functions ([#7684](https://github.com/rapidsai/cudf/pull/7684)) [@davidwendt](https://github.com/davidwendt) -- Add column_device_view to orc writer ([#7676](https://github.com/rapidsai/cudf/pull/7676)) [@kaatish](https://github.com/kaatish) -- cudf_kafka now uses cuDF CMake export targets (CPM) ([#7674](https://github.com/rapidsai/cudf/pull/7674)) [@robertmaynard](https://github.com/robertmaynard) -- Add gbenchmark for nvtext normalize functions ([#7668](https://github.com/rapidsai/cudf/pull/7668)) [@davidwendt](https://github.com/davidwendt) -- Resolve unnecessary import of thrust/optional.hpp in types.hpp ([#7667](https://github.com/rapidsai/cudf/pull/7667)) [@vyasr](https://github.com/vyasr) -- Feature/optimize accessor copy ([#7660](https://github.com/rapidsai/cudf/pull/7660)) [@vyasr](https://github.com/vyasr) -- Fix `find_package(cudf)` ([#7658](https://github.com/rapidsai/cudf/pull/7658)) [@trxcllnt](https://github.com/trxcllnt) -- Work-around for gcc7 compile error on Centos7 ([#7652](https://github.com/rapidsai/cudf/pull/7652)) [@davidwendt](https://github.com/davidwendt) -- Add in JNI support for count_elements ([#7651](https://github.com/rapidsai/cudf/pull/7651)) [@revans2](https://github.com/revans2) -- Fix issues with building cudf in a non-conda environment ([#7647](https://github.com/rapidsai/cudf/pull/7647)) [@galipremsagar](https://github.com/galipremsagar) -- Refactor ConfigureCUDA to not conditionally insert compiler flags ([#7643](https://github.com/rapidsai/cudf/pull/7643)) [@robertmaynard](https://github.com/robertmaynard) -- Add gbenchmark for converting strings to/from timestamps ([#7641](https://github.com/rapidsai/cudf/pull/7641)) [@davidwendt](https://github.com/davidwendt) -- Handle constructing a `cudf.Scalar` from a `cudf.Scalar` ([#7639](https://github.com/rapidsai/cudf/pull/7639)) [@shwina](https://github.com/shwina) -- Add in JNI support for table partition ([#7637](https://github.com/rapidsai/cudf/pull/7637)) [@revans2](https://github.com/revans2) -- Add explicit fixed_point merge test ([#7635](https://github.com/rapidsai/cudf/pull/7635)) [@codereport](https://github.com/codereport) -- Add JNI support for IDENTITY hash partitioning ([#7626](https://github.com/rapidsai/cudf/pull/7626)) [@revans2](https://github.com/revans2) -- Java support on explode_outer ([#7625](https://github.com/rapidsai/cudf/pull/7625)) [@sperlingxx](https://github.com/sperlingxx) -- Java support of casting string from/to decimal ([#7623](https://github.com/rapidsai/cudf/pull/7623)) [@sperlingxx](https://github.com/sperlingxx) -- Convert cudf::concatenate APIs to use spans and device_uvector ([#7621](https://github.com/rapidsai/cudf/pull/7621)) [@harrism](https://github.com/harrism) -- Add gbenchmark for cudf::strings::translate function ([#7617](https://github.com/rapidsai/cudf/pull/7617)) [@davidwendt](https://github.com/davidwendt) -- Use file(COPY ) over file(INSTALL ) so cmake output is reduced ([#7616](https://github.com/rapidsai/cudf/pull/7616)) [@robertmaynard](https://github.com/robertmaynard) -- Use rmm::device_uvector in place of rmm::device_vector for ORC reader/writer and cudf::io::column_buffer ([#7614](https://github.com/rapidsai/cudf/pull/7614)) [@vuule](https://github.com/vuule) -- Refactor Java host-side buffer concatenation to expose separate steps ([#7610](https://github.com/rapidsai/cudf/pull/7610)) [@jlowe](https://github.com/jlowe) -- Add gbenchmarks for string substrings functions ([#7603](https://github.com/rapidsai/cudf/pull/7603)) [@davidwendt](https://github.com/davidwendt) -- Refactor string conversion check ([#7599](https://github.com/rapidsai/cudf/pull/7599)) [@ttnghia](https://github.com/ttnghia) -- JNI: Pass names of children struct columns to native Arrow IPC writer ([#7598](https://github.com/rapidsai/cudf/pull/7598)) [@firestarman](https://github.com/firestarman) -- Revert "ENH Fix stale GHA and prevent duplicates " ([#7595](https://github.com/rapidsai/cudf/pull/7595)) [@mike-wendt](https://github.com/mike-wendt) -- ENH Fix stale GHA and prevent duplicates ([#7594](https://github.com/rapidsai/cudf/pull/7594)) [@mike-wendt](https://github.com/mike-wendt) -- Fix auto-detecting GPU architectures ([#7593](https://github.com/rapidsai/cudf/pull/7593)) [@trxcllnt](https://github.com/trxcllnt) -- Reduce cudf library size ([#7583](https://github.com/rapidsai/cudf/pull/7583)) [@robertmaynard](https://github.com/robertmaynard) -- Optimize cudf::make_strings_column for long strings ([#7576](https://github.com/rapidsai/cudf/pull/7576)) [@davidwendt](https://github.com/davidwendt) -- Always build and export the cudf::cudftestutil target ([#7574](https://github.com/rapidsai/cudf/pull/7574)) [@trxcllnt](https://github.com/trxcllnt) -- Eliminate literal parameters to uvector::set_element_async and device_scalar::set_value ([#7563](https://github.com/rapidsai/cudf/pull/7563)) [@harrism](https://github.com/harrism) -- Add gbenchmark for strings::concatenate ([#7560](https://github.com/rapidsai/cudf/pull/7560)) [@davidwendt](https://github.com/davidwendt) -- Update Changelog Link ([#7550](https://github.com/rapidsai/cudf/pull/7550)) [@ajschmidt8](https://github.com/ajschmidt8) -- Add gbenchmarks for strings replace regex functions ([#7541](https://github.com/rapidsai/cudf/pull/7541)) [@davidwendt](https://github.com/davidwendt) -- Add `__repr__` for Column and ColumnAccessor ([#7531](https://github.com/rapidsai/cudf/pull/7531)) [@shwina](https://github.com/shwina) -- Support Decimal DIV changes in cudf ([#7527](https://github.com/rapidsai/cudf/pull/7527)) [@razajafri](https://github.com/razajafri) -- Remove unneeded step parameter from strings::detail::copy_slice ([#7525](https://github.com/rapidsai/cudf/pull/7525)) [@davidwendt](https://github.com/davidwendt) -- Use device_uvector, device_span in sort groupby ([#7523](https://github.com/rapidsai/cudf/pull/7523)) [@karthikeyann](https://github.com/karthikeyann) -- Add gbenchmarks for strings extract function ([#7522](https://github.com/rapidsai/cudf/pull/7522)) [@davidwendt](https://github.com/davidwendt) -- Rename ARROW_STATIC_LIB because it conflicts with one in FindArrow.cmake ([#7518](https://github.com/rapidsai/cudf/pull/7518)) [@trxcllnt](https://github.com/trxcllnt) -- Reduce compile time/size for scan.cu ([#7516](https://github.com/rapidsai/cudf/pull/7516)) [@davidwendt](https://github.com/davidwendt) -- Change device_vector to device_uvector in nvtext source files ([#7512](https://github.com/rapidsai/cudf/pull/7512)) [@davidwendt](https://github.com/davidwendt) -- Removed unneeded includes from traits.hpp ([#7509](https://github.com/rapidsai/cudf/pull/7509)) [@davidwendt](https://github.com/davidwendt) -- FIX Remove random build directory generation for ccache ([#7508](https://github.com/rapidsai/cudf/pull/7508)) [@dillon-cullinan](https://github.com/dillon-cullinan) -- xfail failing pytest in pandas 1.2.3 ([#7507](https://github.com/rapidsai/cudf/pull/7507)) [@galipremsagar](https://github.com/galipremsagar) -- JNI bit cast ([#7493](https://github.com/rapidsai/cudf/pull/7493)) [@revans2](https://github.com/revans2) -- Combine rolling window function tests ([#7480](https://github.com/rapidsai/cudf/pull/7480)) [@mythrocks](https://github.com/mythrocks) -- Prepare Changelog for Automation ([#7477](https://github.com/rapidsai/cudf/pull/7477)) [@ajschmidt8](https://github.com/ajschmidt8) -- Java support for explode position ([#7471](https://github.com/rapidsai/cudf/pull/7471)) [@sperlingxx](https://github.com/sperlingxx) -- Update 0.18 changelog entry ([#7463](https://github.com/rapidsai/cudf/pull/7463)) [@ajschmidt8](https://github.com/ajschmidt8) -- JNI: Support skipping nulls for collect aggregation ([#7457](https://github.com/rapidsai/cudf/pull/7457)) [@firestarman](https://github.com/firestarman) -- Join APIs that return gathermaps ([#7454](https://github.com/rapidsai/cudf/pull/7454)) [@shwina](https://github.com/shwina) -- Remove dependence on managed memory for multimap test ([#7451](https://github.com/rapidsai/cudf/pull/7451)) [@jrhemstad](https://github.com/jrhemstad) -- Use cuFile for Parquet IO when available ([#7444](https://github.com/rapidsai/cudf/pull/7444)) [@vuule](https://github.com/vuule) -- Statistics cleanup ([#7439](https://github.com/rapidsai/cudf/pull/7439)) [@kaatish](https://github.com/kaatish) -- Add gbenchmarks for strings filter functions ([#7438](https://github.com/rapidsai/cudf/pull/7438)) [@davidwendt](https://github.com/davidwendt) -- `fixed_point` + `cudf::binary_operation` API Changes ([#7435](https://github.com/rapidsai/cudf/pull/7435)) [@codereport](https://github.com/codereport) -- Improve string gather performance ([#7433](https://github.com/rapidsai/cudf/pull/7433)) [@jlowe](https://github.com/jlowe) -- Don't use user resource for a temporary allocation in sort_by_key ([#7431](https://github.com/rapidsai/cudf/pull/7431)) [@magnatelee](https://github.com/magnatelee) -- Detail APIs for datetime functions ([#7430](https://github.com/rapidsai/cudf/pull/7430)) [@magnatelee](https://github.com/magnatelee) -- Replace thrust::max_element with thrust::reduce in strings findall_re ([#7428](https://github.com/rapidsai/cudf/pull/7428)) [@davidwendt](https://github.com/davidwendt) -- Add gbenchmark for strings split/split_record functions ([#7427](https://github.com/rapidsai/cudf/pull/7427)) [@davidwendt](https://github.com/davidwendt) -- Update JNI build to use CMAKE_CUDA_ARCHITECTURES ([#7425](https://github.com/rapidsai/cudf/pull/7425)) [@jlowe](https://github.com/jlowe) -- Change nvtext::load_vocabulary_file to return a unique ptr ([#7424](https://github.com/rapidsai/cudf/pull/7424)) [@davidwendt](https://github.com/davidwendt) -- Simplify type dispatch with `device_storage_dispatch` ([#7419](https://github.com/rapidsai/cudf/pull/7419)) [@codereport](https://github.com/codereport) -- Java support for casting of nested child columns ([#7417](https://github.com/rapidsai/cudf/pull/7417)) [@razajafri](https://github.com/razajafri) -- Improve scalar string replace performance for long strings ([#7415](https://github.com/rapidsai/cudf/pull/7415)) [@jlowe](https://github.com/jlowe) -- Remove unneeded temporary device vector for strings scatter specialization ([#7409](https://github.com/rapidsai/cudf/pull/7409)) [@davidwendt](https://github.com/davidwendt) -- bitmask_or implementation with bitmask refactor ([#7406](https://github.com/rapidsai/cudf/pull/7406)) [@rwlee](https://github.com/rwlee) -- Add other cudf::strings::replace functions to current strings replace gbenchmark ([#7403](https://github.com/rapidsai/cudf/pull/7403)) [@davidwendt](https://github.com/davidwendt) -- Clean up included headers in `device_operators.cuh` ([#7401](https://github.com/rapidsai/cudf/pull/7401)) [@codereport](https://github.com/codereport) -- Move nullable index iterator to indexalator factory ([#7399](https://github.com/rapidsai/cudf/pull/7399)) [@davidwendt](https://github.com/davidwendt) -- ENH Pass ccache variables to conda recipe & use Ninja in CI ([#7398](https://github.com/rapidsai/cudf/pull/7398)) [@Ethyling](https://github.com/Ethyling) -- upgrade maven-antrun-plugin to support maven parallel builds ([#7393](https://github.com/rapidsai/cudf/pull/7393)) [@rongou](https://github.com/rongou) -- Add gbenchmark for strings find/contains functions ([#7392](https://github.com/rapidsai/cudf/pull/7392)) [@davidwendt](https://github.com/davidwendt) -- Use CMAKE_CUDA_ARCHITECTURES ([#7391](https://github.com/rapidsai/cudf/pull/7391)) [@robertmaynard](https://github.com/robertmaynard) -- Refactor libcudf strings::replace to use make_strings_children utility ([#7384](https://github.com/rapidsai/cudf/pull/7384)) [@davidwendt](https://github.com/davidwendt) -- Added in JNI support for out of core sort algorithm ([#7381](https://github.com/rapidsai/cudf/pull/7381)) [@revans2](https://github.com/revans2) -- Upgrade pandas to 1.2 ([#7375](https://github.com/rapidsai/cudf/pull/7375)) [@galipremsagar](https://github.com/galipremsagar) -- Rename `logical_cast` to `bit_cast` and allow additional conversions ([#7373](https://github.com/rapidsai/cudf/pull/7373)) [@ttnghia](https://github.com/ttnghia) -- jitify 2 support ([#7372](https://github.com/rapidsai/cudf/pull/7372)) [@cwharris](https://github.com/cwharris) -- compile_udf: Cache PTX for similar functions ([#7371](https://github.com/rapidsai/cudf/pull/7371)) [@gmarkall](https://github.com/gmarkall) -- Add string scalar replace benchmark ([#7369](https://github.com/rapidsai/cudf/pull/7369)) [@jlowe](https://github.com/jlowe) -- Add gbenchmark for strings contains_re/count_re functions ([#7366](https://github.com/rapidsai/cudf/pull/7366)) [@davidwendt](https://github.com/davidwendt) -- Update orc reader and writer fuzz tests ([#7357](https://github.com/rapidsai/cudf/pull/7357)) [@galipremsagar](https://github.com/galipremsagar) -- Improve url_decode performance for long strings ([#7353](https://github.com/rapidsai/cudf/pull/7353)) [@jlowe](https://github.com/jlowe) -- `cudf::ast` Small Refactorings ([#7352](https://github.com/rapidsai/cudf/pull/7352)) [@codereport](https://github.com/codereport) -- Remove std::cout and print in the scatter test function EmptyListsOfNullableStrings. ([#7342](https://github.com/rapidsai/cudf/pull/7342)) [@ttnghia](https://github.com/ttnghia) -- Use `cudf::detail::make_counting_transform_iterator` ([#7338](https://github.com/rapidsai/cudf/pull/7338)) [@codereport](https://github.com/codereport) -- Change block size parameter from a global to a template param. ([#7333](https://github.com/rapidsai/cudf/pull/7333)) [@nvdbaranec](https://github.com/nvdbaranec) -- Partial clean up of ORC writer ([#7324](https://github.com/rapidsai/cudf/pull/7324)) [@vuule](https://github.com/vuule) -- Add gbenchmark for cudf::strings::to_lower ([#7316](https://github.com/rapidsai/cudf/pull/7316)) [@davidwendt](https://github.com/davidwendt) -- Update Java bindings version to 0.19-SNAPSHOT ([#7307](https://github.com/rapidsai/cudf/pull/7307)) [@pxLi](https://github.com/pxLi) -- Move `cudf::test::make_counting_transform_iterator` to `cudf/detail/iterator.cuh` ([#7306](https://github.com/rapidsai/cudf/pull/7306)) [@codereport](https://github.com/codereport) -- Use string literals in `fixed_point` `release_assert`s ([#7303](https://github.com/rapidsai/cudf/pull/7303)) [@codereport](https://github.com/codereport) -- Fix merge conflicts for #7295 ([#7297](https://github.com/rapidsai/cudf/pull/7297)) [@ajschmidt8](https://github.com/ajschmidt8) -- Add UTF-8 chars to create_random_column<string_view> benchmark utility ([#7292](https://github.com/rapidsai/cudf/pull/7292)) [@davidwendt](https://github.com/davidwendt) -- Abstracting block reduce and block scan from cuIO kernels with `cub` apis ([#7278](https://github.com/rapidsai/cudf/pull/7278)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) -- Build.sh use cmake --build to drive build system invocation ([#7270](https://github.com/rapidsai/cudf/pull/7270)) [@robertmaynard](https://github.com/robertmaynard) -- Refactor dictionary support for reductions any/all ([#7242](https://github.com/rapidsai/cudf/pull/7242)) [@davidwendt](https://github.com/davidwendt) -- Replace stream.value() with stream for stream_view args ([#7236](https://github.com/rapidsai/cudf/pull/7236)) [@karthikeyann](https://github.com/karthikeyann) -- Interval index and interval_range ([#7182](https://github.com/rapidsai/cudf/pull/7182)) [@marlenezw](https://github.com/marlenezw) -- avro reader integration tests ([#7156](https://github.com/rapidsai/cudf/pull/7156)) [@cwharris](https://github.com/cwharris) -- Rework libcudf CMakeLists.txt to export targets for CPM ([#7107](https://github.com/rapidsai/cudf/pull/7107)) [@trxcllnt](https://github.com/trxcllnt) -- Adding Interval Dtype ([#6984](https://github.com/rapidsai/cudf/pull/6984)) [@marlenezw](https://github.com/marlenezw) -- Cleaning up `for` loops with `make_(counting_)transform_iterator` ([#6546](https://github.com/rapidsai/cudf/pull/6546)) [@codereport](https://github.com/codereport) +- fix GDS include path for version 0.95 ([#7877](https://github.com/NVIDIA/cudf/pull/7877)) [@rongou](https://github.com/rongou) +- Update `dask` + `distributed` to `2021.4.0` ([#7858](https://github.com/NVIDIA/cudf/pull/7858)) [@jakirkham](https://github.com/jakirkham) +- Add ability to extract include dirs from `CUDF_HOME` ([#7848](https://github.com/NVIDIA/cudf/pull/7848)) [@galipremsagar](https://github.com/galipremsagar) +- Add USE_GDS as an option in build script ([#7833](https://github.com/NVIDIA/cudf/pull/7833)) [@pxLi](https://github.com/pxLi) +- add an allocate method with stream in java DeviceMemoryBuffer ([#7826](https://github.com/NVIDIA/cudf/pull/7826)) [@rongou](https://github.com/rongou) +- Constrain dask and distributed versions to 2021.3.1 ([#7825](https://github.com/NVIDIA/cudf/pull/7825)) [@shwina](https://github.com/shwina) +- Revert dask versioning of concat dispatch ([#7823](https://github.com/NVIDIA/cudf/pull/7823)) [@galipremsagar](https://github.com/galipremsagar) +- add copy methods in Java memory buffer ([#7791](https://github.com/NVIDIA/cudf/pull/7791)) [@rongou](https://github.com/rongou) +- Update README and CONTRIBUTING for 0.19 ([#7778](https://github.com/NVIDIA/cudf/pull/7778)) [@robertmaynard](https://github.com/robertmaynard) +- Allow hash_partition to take a seed value ([#7771](https://github.com/NVIDIA/cudf/pull/7771)) [@magnatelee](https://github.com/magnatelee) +- Turn on NVTX by default in java build ([#7761](https://github.com/NVIDIA/cudf/pull/7761)) [@tgravescs](https://github.com/tgravescs) +- Add Java bindings to join gather map APIs ([#7751](https://github.com/NVIDIA/cudf/pull/7751)) [@jlowe](https://github.com/jlowe) +- Add replacements column support for Java replaceNulls ([#7750](https://github.com/NVIDIA/cudf/pull/7750)) [@jlowe](https://github.com/jlowe) +- Add Java bindings for row_bit_count ([#7749](https://github.com/NVIDIA/cudf/pull/7749)) [@jlowe](https://github.com/jlowe) +- Remove unused JVM array creation ([#7748](https://github.com/NVIDIA/cudf/pull/7748)) [@jlowe](https://github.com/jlowe) +- Added JNI support for new is_integer ([#7739](https://github.com/NVIDIA/cudf/pull/7739)) [@revans2](https://github.com/revans2) +- Create and promote library aliases in libcudf installations ([#7734](https://github.com/NVIDIA/cudf/pull/7734)) [@trxcllnt](https://github.com/trxcllnt) +- Support groupby operations for decimal dtypes ([#7731](https://github.com/NVIDIA/cudf/pull/7731)) [@vyasr](https://github.com/vyasr) +- Memory map the input file only when GDS compatibility mode is not used ([#7717](https://github.com/NVIDIA/cudf/pull/7717)) [@vuule](https://github.com/vuule) +- Replace device_vector with device_uvector in null_mask ([#7715](https://github.com/NVIDIA/cudf/pull/7715)) [@harrism](https://github.com/harrism) +- Struct hashing support for SerialMurmur3 and SparkMurmur3 ([#7714](https://github.com/NVIDIA/cudf/pull/7714)) [@jlowe](https://github.com/jlowe) +- Add gbenchmark for nvtext replace-tokens function ([#7708](https://github.com/NVIDIA/cudf/pull/7708)) [@davidwendt](https://github.com/davidwendt) +- Use stream in groupby calls ([#7705](https://github.com/NVIDIA/cudf/pull/7705)) [@karthikeyann](https://github.com/karthikeyann) +- Update codeowners file ([#7701](https://github.com/NVIDIA/cudf/pull/7701)) [@ajschmidt8](https://github.com/ajschmidt8) +- Cleanup groupby to use host_span, device_span, device_uvector ([#7698](https://github.com/NVIDIA/cudf/pull/7698)) [@karthikeyann](https://github.com/karthikeyann) +- Add gbenchmark for nvtext ngrams functions ([#7693](https://github.com/NVIDIA/cudf/pull/7693)) [@davidwendt](https://github.com/davidwendt) +- Misc Python/Cython optimizations ([#7686](https://github.com/NVIDIA/cudf/pull/7686)) [@shwina](https://github.com/shwina) +- Add gbenchmark for nvtext tokenize functions ([#7684](https://github.com/NVIDIA/cudf/pull/7684)) [@davidwendt](https://github.com/davidwendt) +- Add column_device_view to orc writer ([#7676](https://github.com/NVIDIA/cudf/pull/7676)) [@kaatish](https://github.com/kaatish) +- cudf_kafka now uses cuDF CMake export targets (CPM) ([#7674](https://github.com/NVIDIA/cudf/pull/7674)) [@robertmaynard](https://github.com/robertmaynard) +- Add gbenchmark for nvtext normalize functions ([#7668](https://github.com/NVIDIA/cudf/pull/7668)) [@davidwendt](https://github.com/davidwendt) +- Resolve unnecessary import of thrust/optional.hpp in types.hpp ([#7667](https://github.com/NVIDIA/cudf/pull/7667)) [@vyasr](https://github.com/vyasr) +- Feature/optimize accessor copy ([#7660](https://github.com/NVIDIA/cudf/pull/7660)) [@vyasr](https://github.com/vyasr) +- Fix `find_package(cudf)` ([#7658](https://github.com/NVIDIA/cudf/pull/7658)) [@trxcllnt](https://github.com/trxcllnt) +- Work-around for gcc7 compile error on Centos7 ([#7652](https://github.com/NVIDIA/cudf/pull/7652)) [@davidwendt](https://github.com/davidwendt) +- Add in JNI support for count_elements ([#7651](https://github.com/NVIDIA/cudf/pull/7651)) [@revans2](https://github.com/revans2) +- Fix issues with building cudf in a non-conda environment ([#7647](https://github.com/NVIDIA/cudf/pull/7647)) [@galipremsagar](https://github.com/galipremsagar) +- Refactor ConfigureCUDA to not conditionally insert compiler flags ([#7643](https://github.com/NVIDIA/cudf/pull/7643)) [@robertmaynard](https://github.com/robertmaynard) +- Add gbenchmark for converting strings to/from timestamps ([#7641](https://github.com/NVIDIA/cudf/pull/7641)) [@davidwendt](https://github.com/davidwendt) +- Handle constructing a `cudf.Scalar` from a `cudf.Scalar` ([#7639](https://github.com/NVIDIA/cudf/pull/7639)) [@shwina](https://github.com/shwina) +- Add in JNI support for table partition ([#7637](https://github.com/NVIDIA/cudf/pull/7637)) [@revans2](https://github.com/revans2) +- Add explicit fixed_point merge test ([#7635](https://github.com/NVIDIA/cudf/pull/7635)) [@codereport](https://github.com/codereport) +- Add JNI support for IDENTITY hash partitioning ([#7626](https://github.com/NVIDIA/cudf/pull/7626)) [@revans2](https://github.com/revans2) +- Java support on explode_outer ([#7625](https://github.com/NVIDIA/cudf/pull/7625)) [@sperlingxx](https://github.com/sperlingxx) +- Java support of casting string from/to decimal ([#7623](https://github.com/NVIDIA/cudf/pull/7623)) [@sperlingxx](https://github.com/sperlingxx) +- Convert cudf::concatenate APIs to use spans and device_uvector ([#7621](https://github.com/NVIDIA/cudf/pull/7621)) [@harrism](https://github.com/harrism) +- Add gbenchmark for cudf::strings::translate function ([#7617](https://github.com/NVIDIA/cudf/pull/7617)) [@davidwendt](https://github.com/davidwendt) +- Use file(COPY ) over file(INSTALL ) so cmake output is reduced ([#7616](https://github.com/NVIDIA/cudf/pull/7616)) [@robertmaynard](https://github.com/robertmaynard) +- Use rmm::device_uvector in place of rmm::device_vector for ORC reader/writer and cudf::io::column_buffer ([#7614](https://github.com/NVIDIA/cudf/pull/7614)) [@vuule](https://github.com/vuule) +- Refactor Java host-side buffer concatenation to expose separate steps ([#7610](https://github.com/NVIDIA/cudf/pull/7610)) [@jlowe](https://github.com/jlowe) +- Add gbenchmarks for string substrings functions ([#7603](https://github.com/NVIDIA/cudf/pull/7603)) [@davidwendt](https://github.com/davidwendt) +- Refactor string conversion check ([#7599](https://github.com/NVIDIA/cudf/pull/7599)) [@ttnghia](https://github.com/ttnghia) +- JNI: Pass names of children struct columns to native Arrow IPC writer ([#7598](https://github.com/NVIDIA/cudf/pull/7598)) [@firestarman](https://github.com/firestarman) +- Revert "ENH Fix stale GHA and prevent duplicates " ([#7595](https://github.com/NVIDIA/cudf/pull/7595)) [@mike-wendt](https://github.com/mike-wendt) +- ENH Fix stale GHA and prevent duplicates ([#7594](https://github.com/NVIDIA/cudf/pull/7594)) [@mike-wendt](https://github.com/mike-wendt) +- Fix auto-detecting GPU architectures ([#7593](https://github.com/NVIDIA/cudf/pull/7593)) [@trxcllnt](https://github.com/trxcllnt) +- Reduce cudf library size ([#7583](https://github.com/NVIDIA/cudf/pull/7583)) [@robertmaynard](https://github.com/robertmaynard) +- Optimize cudf::make_strings_column for long strings ([#7576](https://github.com/NVIDIA/cudf/pull/7576)) [@davidwendt](https://github.com/davidwendt) +- Always build and export the cudf::cudftestutil target ([#7574](https://github.com/NVIDIA/cudf/pull/7574)) [@trxcllnt](https://github.com/trxcllnt) +- Eliminate literal parameters to uvector::set_element_async and device_scalar::set_value ([#7563](https://github.com/NVIDIA/cudf/pull/7563)) [@harrism](https://github.com/harrism) +- Add gbenchmark for strings::concatenate ([#7560](https://github.com/NVIDIA/cudf/pull/7560)) [@davidwendt](https://github.com/davidwendt) +- Update Changelog Link ([#7550](https://github.com/NVIDIA/cudf/pull/7550)) [@ajschmidt8](https://github.com/ajschmidt8) +- Add gbenchmarks for strings replace regex functions ([#7541](https://github.com/NVIDIA/cudf/pull/7541)) [@davidwendt](https://github.com/davidwendt) +- Add `__repr__` for Column and ColumnAccessor ([#7531](https://github.com/NVIDIA/cudf/pull/7531)) [@shwina](https://github.com/shwina) +- Support Decimal DIV changes in cudf ([#7527](https://github.com/NVIDIA/cudf/pull/7527)) [@razajafri](https://github.com/razajafri) +- Remove unneeded step parameter from strings::detail::copy_slice ([#7525](https://github.com/NVIDIA/cudf/pull/7525)) [@davidwendt](https://github.com/davidwendt) +- Use device_uvector, device_span in sort groupby ([#7523](https://github.com/NVIDIA/cudf/pull/7523)) [@karthikeyann](https://github.com/karthikeyann) +- Add gbenchmarks for strings extract function ([#7522](https://github.com/NVIDIA/cudf/pull/7522)) [@davidwendt](https://github.com/davidwendt) +- Rename ARROW_STATIC_LIB because it conflicts with one in FindArrow.cmake ([#7518](https://github.com/NVIDIA/cudf/pull/7518)) [@trxcllnt](https://github.com/trxcllnt) +- Reduce compile time/size for scan.cu ([#7516](https://github.com/NVIDIA/cudf/pull/7516)) [@davidwendt](https://github.com/davidwendt) +- Change device_vector to device_uvector in nvtext source files ([#7512](https://github.com/NVIDIA/cudf/pull/7512)) [@davidwendt](https://github.com/davidwendt) +- Removed unneeded includes from traits.hpp ([#7509](https://github.com/NVIDIA/cudf/pull/7509)) [@davidwendt](https://github.com/davidwendt) +- FIX Remove random build directory generation for ccache ([#7508](https://github.com/NVIDIA/cudf/pull/7508)) [@dillon-cullinan](https://github.com/dillon-cullinan) +- xfail failing pytest in pandas 1.2.3 ([#7507](https://github.com/NVIDIA/cudf/pull/7507)) [@galipremsagar](https://github.com/galipremsagar) +- JNI bit cast ([#7493](https://github.com/NVIDIA/cudf/pull/7493)) [@revans2](https://github.com/revans2) +- Combine rolling window function tests ([#7480](https://github.com/NVIDIA/cudf/pull/7480)) [@mythrocks](https://github.com/mythrocks) +- Prepare Changelog for Automation ([#7477](https://github.com/NVIDIA/cudf/pull/7477)) [@ajschmidt8](https://github.com/ajschmidt8) +- Java support for explode position ([#7471](https://github.com/NVIDIA/cudf/pull/7471)) [@sperlingxx](https://github.com/sperlingxx) +- Update 0.18 changelog entry ([#7463](https://github.com/NVIDIA/cudf/pull/7463)) [@ajschmidt8](https://github.com/ajschmidt8) +- JNI: Support skipping nulls for collect aggregation ([#7457](https://github.com/NVIDIA/cudf/pull/7457)) [@firestarman](https://github.com/firestarman) +- Join APIs that return gathermaps ([#7454](https://github.com/NVIDIA/cudf/pull/7454)) [@shwina](https://github.com/shwina) +- Remove dependence on managed memory for multimap test ([#7451](https://github.com/NVIDIA/cudf/pull/7451)) [@jrhemstad](https://github.com/jrhemstad) +- Use cuFile for Parquet IO when available ([#7444](https://github.com/NVIDIA/cudf/pull/7444)) [@vuule](https://github.com/vuule) +- Statistics cleanup ([#7439](https://github.com/NVIDIA/cudf/pull/7439)) [@kaatish](https://github.com/kaatish) +- Add gbenchmarks for strings filter functions ([#7438](https://github.com/NVIDIA/cudf/pull/7438)) [@davidwendt](https://github.com/davidwendt) +- `fixed_point` + `cudf::binary_operation` API Changes ([#7435](https://github.com/NVIDIA/cudf/pull/7435)) [@codereport](https://github.com/codereport) +- Improve string gather performance ([#7433](https://github.com/NVIDIA/cudf/pull/7433)) [@jlowe](https://github.com/jlowe) +- Don't use user resource for a temporary allocation in sort_by_key ([#7431](https://github.com/NVIDIA/cudf/pull/7431)) [@magnatelee](https://github.com/magnatelee) +- Detail APIs for datetime functions ([#7430](https://github.com/NVIDIA/cudf/pull/7430)) [@magnatelee](https://github.com/magnatelee) +- Replace thrust::max_element with thrust::reduce in strings findall_re ([#7428](https://github.com/NVIDIA/cudf/pull/7428)) [@davidwendt](https://github.com/davidwendt) +- Add gbenchmark for strings split/split_record functions ([#7427](https://github.com/NVIDIA/cudf/pull/7427)) [@davidwendt](https://github.com/davidwendt) +- Update JNI build to use CMAKE_CUDA_ARCHITECTURES ([#7425](https://github.com/NVIDIA/cudf/pull/7425)) [@jlowe](https://github.com/jlowe) +- Change nvtext::load_vocabulary_file to return a unique ptr ([#7424](https://github.com/NVIDIA/cudf/pull/7424)) [@davidwendt](https://github.com/davidwendt) +- Simplify type dispatch with `device_storage_dispatch` ([#7419](https://github.com/NVIDIA/cudf/pull/7419)) [@codereport](https://github.com/codereport) +- Java support for casting of nested child columns ([#7417](https://github.com/NVIDIA/cudf/pull/7417)) [@razajafri](https://github.com/razajafri) +- Improve scalar string replace performance for long strings ([#7415](https://github.com/NVIDIA/cudf/pull/7415)) [@jlowe](https://github.com/jlowe) +- Remove unneeded temporary device vector for strings scatter specialization ([#7409](https://github.com/NVIDIA/cudf/pull/7409)) [@davidwendt](https://github.com/davidwendt) +- bitmask_or implementation with bitmask refactor ([#7406](https://github.com/NVIDIA/cudf/pull/7406)) [@rwlee](https://github.com/rwlee) +- Add other cudf::strings::replace functions to current strings replace gbenchmark ([#7403](https://github.com/NVIDIA/cudf/pull/7403)) [@davidwendt](https://github.com/davidwendt) +- Clean up included headers in `device_operators.cuh` ([#7401](https://github.com/NVIDIA/cudf/pull/7401)) [@codereport](https://github.com/codereport) +- Move nullable index iterator to indexalator factory ([#7399](https://github.com/NVIDIA/cudf/pull/7399)) [@davidwendt](https://github.com/davidwendt) +- ENH Pass ccache variables to conda recipe & use Ninja in CI ([#7398](https://github.com/NVIDIA/cudf/pull/7398)) [@Ethyling](https://github.com/Ethyling) +- upgrade maven-antrun-plugin to support maven parallel builds ([#7393](https://github.com/NVIDIA/cudf/pull/7393)) [@rongou](https://github.com/rongou) +- Add gbenchmark for strings find/contains functions ([#7392](https://github.com/NVIDIA/cudf/pull/7392)) [@davidwendt](https://github.com/davidwendt) +- Use CMAKE_CUDA_ARCHITECTURES ([#7391](https://github.com/NVIDIA/cudf/pull/7391)) [@robertmaynard](https://github.com/robertmaynard) +- Refactor libcudf strings::replace to use make_strings_children utility ([#7384](https://github.com/NVIDIA/cudf/pull/7384)) [@davidwendt](https://github.com/davidwendt) +- Added in JNI support for out of core sort algorithm ([#7381](https://github.com/NVIDIA/cudf/pull/7381)) [@revans2](https://github.com/revans2) +- Upgrade pandas to 1.2 ([#7375](https://github.com/NVIDIA/cudf/pull/7375)) [@galipremsagar](https://github.com/galipremsagar) +- Rename `logical_cast` to `bit_cast` and allow additional conversions ([#7373](https://github.com/NVIDIA/cudf/pull/7373)) [@ttnghia](https://github.com/ttnghia) +- jitify 2 support ([#7372](https://github.com/NVIDIA/cudf/pull/7372)) [@cwharris](https://github.com/cwharris) +- compile_udf: Cache PTX for similar functions ([#7371](https://github.com/NVIDIA/cudf/pull/7371)) [@gmarkall](https://github.com/gmarkall) +- Add string scalar replace benchmark ([#7369](https://github.com/NVIDIA/cudf/pull/7369)) [@jlowe](https://github.com/jlowe) +- Add gbenchmark for strings contains_re/count_re functions ([#7366](https://github.com/NVIDIA/cudf/pull/7366)) [@davidwendt](https://github.com/davidwendt) +- Update orc reader and writer fuzz tests ([#7357](https://github.com/NVIDIA/cudf/pull/7357)) [@galipremsagar](https://github.com/galipremsagar) +- Improve url_decode performance for long strings ([#7353](https://github.com/NVIDIA/cudf/pull/7353)) [@jlowe](https://github.com/jlowe) +- `cudf::ast` Small Refactorings ([#7352](https://github.com/NVIDIA/cudf/pull/7352)) [@codereport](https://github.com/codereport) +- Remove std::cout and print in the scatter test function EmptyListsOfNullableStrings. ([#7342](https://github.com/NVIDIA/cudf/pull/7342)) [@ttnghia](https://github.com/ttnghia) +- Use `cudf::detail::make_counting_transform_iterator` ([#7338](https://github.com/NVIDIA/cudf/pull/7338)) [@codereport](https://github.com/codereport) +- Change block size parameter from a global to a template param. ([#7333](https://github.com/NVIDIA/cudf/pull/7333)) [@nvdbaranec](https://github.com/nvdbaranec) +- Partial clean up of ORC writer ([#7324](https://github.com/NVIDIA/cudf/pull/7324)) [@vuule](https://github.com/vuule) +- Add gbenchmark for cudf::strings::to_lower ([#7316](https://github.com/NVIDIA/cudf/pull/7316)) [@davidwendt](https://github.com/davidwendt) +- Update Java bindings version to 0.19-SNAPSHOT ([#7307](https://github.com/NVIDIA/cudf/pull/7307)) [@pxLi](https://github.com/pxLi) +- Move `cudf::test::make_counting_transform_iterator` to `cudf/detail/iterator.cuh` ([#7306](https://github.com/NVIDIA/cudf/pull/7306)) [@codereport](https://github.com/codereport) +- Use string literals in `fixed_point` `release_assert`s ([#7303](https://github.com/NVIDIA/cudf/pull/7303)) [@codereport](https://github.com/codereport) +- Fix merge conflicts for #7295 ([#7297](https://github.com/NVIDIA/cudf/pull/7297)) [@ajschmidt8](https://github.com/ajschmidt8) +- Add UTF-8 chars to create_random_column<string_view> benchmark utility ([#7292](https://github.com/NVIDIA/cudf/pull/7292)) [@davidwendt](https://github.com/davidwendt) +- Abstracting block reduce and block scan from cuIO kernels with `cub` apis ([#7278](https://github.com/NVIDIA/cudf/pull/7278)) [@rgsl888prabhu](https://github.com/rgsl888prabhu) +- Build.sh use cmake --build to drive build system invocation ([#7270](https://github.com/NVIDIA/cudf/pull/7270)) [@robertmaynard](https://github.com/robertmaynard) +- Refactor dictionary support for reductions any/all ([#7242](https://github.com/NVIDIA/cudf/pull/7242)) [@davidwendt](https://github.com/davidwendt) +- Replace stream.value() with stream for stream_view args ([#7236](https://github.com/NVIDIA/cudf/pull/7236)) [@karthikeyann](https://github.com/karthikeyann) +- Interval index and interval_range ([#7182](https://github.com/NVIDIA/cudf/pull/7182)) [@marlenezw](https://github.com/marlenezw) +- avro reader integration tests ([#7156](https://github.com/NVIDIA/cudf/pull/7156)) [@cwharris](https://github.com/cwharris) +- Rework libcudf CMakeLists.txt to export targets for CPM ([#7107](https://github.com/NVIDIA/cudf/pull/7107)) [@trxcllnt](https://github.com/trxcllnt) +- Adding Interval Dtype ([#6984](https://github.com/NVIDIA/cudf/pull/6984)) [@marlenezw](https://github.com/marlenezw) +- Cleaning up `for` loops with `make_(counting_)transform_iterator` ([#6546](https://github.com/NVIDIA/cudf/pull/6546)) [@codereport](https://github.com/codereport) # cuDF 0.18.0 (24 Feb 2021) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 92313faba679..5f4d7877582d 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -3,11 +3,11 @@ Contributions to cuDF fall into the following categories: 1. To report a bug, request a new feature, or report a problem with documentation, please file an - [issue](https://github.com/rapidsai/cudf/issues/new/choose) describing the problem or new feature + [issue](https://github.com/NVIDIA/cudf/issues/new/choose) describing the problem or new feature in detail. The RAPIDS team evaluates and triages issues, and schedules them for a release. If you believe the issue needs priority attention, please comment on the issue to notify the team. 2. To propose and implement a new feature, please file a new feature request - [issue](https://github.com/rapidsai/cudf/issues/new/choose). Describe the intended feature and + [issue](https://github.com/NVIDIA/cudf/issues/new/choose). Describe the intended feature and discuss the design and implementation with the team and community. Once the team agrees that the plan looks good, go ahead and implement it, using the [code contributions](#code-contributions) guide below. @@ -42,15 +42,15 @@ documentation docs](https://docs.rapids.ai/api/cudf/stable/developer_guide/docum 1. Follow the guide at the bottom of this page for [Setting up your build environment](#setting-up-your-build-environment). 2. Find an issue to work on. The best way is to look for the - [good first issue](https://github.com/rapidsai/cudf/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) - or [help wanted](https://github.com/rapidsai/cudf/issues?q=is%3Aissue+is%3Aopen+label%3A%22help+wanted%22) + [good first issue](https://github.com/NVIDIA/cudf/issues?q=is%3Aissue+is%3Aopen+label%3A%22good+first+issue%22) + or [help wanted](https://github.com/NVIDIA/cudf/issues?q=is%3Aissue+is%3Aopen+label%3A%22help+wanted%22) labels. 3. Comment on the issue stating that you are going to work on it. 4. Create a fork of the cudf repository and check out a branch with a name that describes your planned work. For example, `fix-documentation`. 5. Write code to address the issue or implement the feature. 6. Add unit tests and unit benchmarks. -7. [Create your pull request](https://github.com/rapidsai/cudf/compare). To run continuous integration (CI) tests without requesting review, open a draft pull request. +7. [Create your pull request](https://github.com/NVIDIA/cudf/compare). To run continuous integration (CI) tests without requesting review, open a draft pull request. 8. Verify that CI passes all [status checks](https://docs.github.com/en/pull-requests/collaborating-with-pull-requests/collaborating-on-repositories-with-code-quality-features/about-status-checks). Fix if needed. 9. Wait for other developers to review your code and update code as needed. @@ -66,7 +66,7 @@ If you are unsure about anything, don't hesitate to comment on issues and ask fo Once you have gotten your feet wet and are more comfortable with the code, you can look at the prioritized issues for our next release in our -[project boards](https://github.com/rapidsai/cudf/projects). +[project boards](https://github.com/NVIDIA/cudf/projects). **Note:** Always look at the release board that is [currently under development](https://docs.rapids.ai/maintainers/) for issues to work on. This is @@ -109,7 +109,7 @@ You can obtain CUDA from ```bash CUDF_HOME=$(pwd)/cudf -git clone https://github.com/rapidsai/cudf.git $CUDF_HOME +git clone https://github.com/NVIDIA/cudf.git $CUDF_HOME cd $CUDF_HOME ``` diff --git a/README.md b/README.md index 529e50dabb35..60ef0be89fa9 100644 --- a/README.md +++ b/README.md @@ -131,7 +131,7 @@ lf.drop_nulls().group_by(["A", "B"]).mean().collect(engine="gpu") ## Questions and Discussion -For bug reports or feature requests, please [file an issue](https://github.com/rapidsai/cudf/issues/new/choose) on the GitHub issue tracker. +For bug reports or feature requests, please [file an issue](https://github.com/NVIDIA/cudf/issues/new/choose) on the GitHub issue tracker. For questions or discussion about cuDF and GPU data processing, feel free to post in the [RAPIDS Slack](https://rapids.ai/slack-invite) workspace. diff --git a/ci/cudf_pandas_scripts/pandas-tests/run.sh b/ci/cudf_pandas_scripts/pandas-tests/run.sh index 1283aa24f349..eae1b272d367 100755 --- a/ci/cudf_pandas_scripts/pandas-tests/run.sh +++ b/ci/cudf_pandas_scripts/pandas-tests/run.sh @@ -1,5 +1,5 @@ #!/usr/bin/env bash -# SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. +# SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 set -euo pipefail @@ -59,7 +59,7 @@ MAIN_RUN_ID=$( gh run list \ -w "Pandas Test Job" \ -b "$(<./RAPIDS_BRANCH)" \ - --repo 'rapidsai/cudf' \ + --repo 'NVIDIA/cudf' \ --status success \ --limit 7 \ --json 'createdAt,databaseId' \ @@ -73,7 +73,7 @@ fi rapids-logger "Fetching latest available results from nightly: ${MAIN_RUN_ID}" gh run download \ - --repo 'rapidsai/cudf' \ + --repo 'NVIDIA/cudf' \ --name main-results.json \ $MAIN_RUN_ID diff --git a/ci/cudf_pandas_scripts/run_tests.sh b/ci/cudf_pandas_scripts/run_tests.sh index baaf9962b1b1..bdd14aa502ad 100755 --- a/ci/cudf_pandas_scripts/run_tests.sh +++ b/ci/cudf_pandas_scripts/run_tests.sh @@ -73,7 +73,7 @@ python -m pytest -p cudf.pandas \ # even with pytest.mark.no_cover. This likely stems from specialized logic in # coveragepy and pytest-cov for distributed testing (pytest-dev/pytest-cov#246). # As a workaround, we run profiler tests separately without parallelism or `--cov`. -# More details: https://github.com/rapidsai/cudf/pull/16930#issuecomment-2707873968 +# More details: https://github.com/NVIDIA/cudf/pull/16930#issuecomment-2707873968 python -m pytest -p cudf.pandas \ --ignore=./python/cudf/cudf_pandas_tests/third_party_integration_tests/ \ --config-file=./python/cudf/pyproject.toml \ diff --git a/ci/narwhals_cudf_pandas_test_plugin.py b/ci/narwhals_cudf_pandas_test_plugin.py index b05432e2ff41..fac4266150f5 100644 --- a/ci/narwhals_cudf_pandas_test_plugin.py +++ b/ci/narwhals_cudf_pandas_test_plugin.py @@ -14,19 +14,19 @@ TESTS_TO_SKIP: Mapping[str, str] = { - "tests/expr_and_series/list/len_test.py::test_pandas_object_series": "cudf.pandas does not raise on object-dtype Series construction (rapidsai/cudf#18248).", - "tests/expr_and_series/struct_/field_test.py::test_pandas_object_series": "cudf.pandas does not raise on object-dtype Series construction (rapidsai/cudf#18248).", - "tests/frame/to_arrow_test.py::test_to_arrow[pandas]": "cudf.pandas cannot convert to a host Arrow object (rapidsai/cudf#18248).", + "tests/expr_and_series/list/len_test.py::test_pandas_object_series": "cudf.pandas does not raise on object-dtype Series construction (NVIDIA/cudf#18248).", + "tests/expr_and_series/struct_/field_test.py::test_pandas_object_series": "cudf.pandas does not raise on object-dtype Series construction (NVIDIA/cudf#18248).", + "tests/frame/to_arrow_test.py::test_to_arrow[pandas]": "cudf.pandas cannot convert to a host Arrow object (NVIDIA/cudf#18248).", "tests/series_only/from_iterable_test.py::test_series_from_iterable[pandas-polars.series.series.Series-String]": "Flaky under cudf.pandas: a pandas DeprecationWarning (turned into an error by Narwhals' filterwarnings) fires non-deterministically, and Narwhals otherwise marks this param xfail(strict) so an XPASS also fails.", } EXPECTED_FAILURES: Mapping[str, str] = { - "tests/expr_and_series/all_horizontal_test.py::test_all_ignore_nulls[pandas]": "rapidsai/cudf#19417: Kleene any/all horizontal with nulls.", - "tests/expr_and_series/all_horizontal_test.py::test_allh_kleene[pandas]": "rapidsai/cudf#19417: Kleene any/all horizontal with nulls.", - "tests/expr_and_series/any_horizontal_test.py::test_anyh_kleene[pandas]": "rapidsai/cudf#19417: Kleene any/all horizontal with nulls.", + "tests/expr_and_series/all_horizontal_test.py::test_all_ignore_nulls[pandas]": "NVIDIA/cudf#19417: Kleene any/all horizontal with nulls.", + "tests/expr_and_series/all_horizontal_test.py::test_allh_kleene[pandas]": "NVIDIA/cudf#19417: Kleene any/all horizontal with nulls.", + "tests/expr_and_series/any_horizontal_test.py::test_anyh_kleene[pandas]": "NVIDIA/cudf#19417: Kleene any/all horizontal with nulls.", "tests/expr_and_series/cast_test.py::test_pandas_pyarrow_dtypes": "cudf.pandas preserves pyarrow extension dtypes, causing PyArrow dtype handling differences.", - "tests/expr_and_series/dt/offset_by_test.py::test_offset_by_date_pandas": "rapidsai/cudf#19418: dt.offset_by on date columns.", + "tests/expr_and_series/dt/offset_by_test.py::test_offset_by_date_pandas": "NVIDIA/cudf#19418: dt.offset_by on date columns.", "tests/expr_and_series/fill_null_test.py::test_fill_null_pandas_downcast": "cudf.pandas represents nullable bool natively, so fill_null keeps a bool (not object) dtype.", "tests/expr_and_series/list/get_test.py::test_get_series[pandas-0-expected0]": "cudf.pandas List.get does not raise on out-of-bounds indices.", "tests/expr_and_series/log_test.py::test_log_dtype_pandas": "cudf.pandas promotes the result of log to float64.", @@ -45,8 +45,8 @@ "tests/expr_and_series/pandas_str_dtypes_test.py::test_pandas_str_types[left_dtype29-right_dtype29-result_dtype29]": "cudf.pandas returns a different pandas string dtype than the test expects.", "tests/expr_and_series/pandas_str_dtypes_test.py::test_pandas_str_types[left_dtype30-right_dtype30-result_dtype30]": "cudf.pandas returns a different pandas string dtype than the test expects.", "tests/expr_and_series/pandas_str_dtypes_test.py::test_pandas_str_types[left_dtype31-right_dtype31-result_dtype31]": "cudf.pandas returns a different pandas string dtype than the test expects.", - "tests/frame/select_test.py::test_select_boolean_cols": "rapidsai/cudf#19421: selecting boolean columns raises a length-mismatch error.", - "tests/frame/select_test.py::test_select_boolean_cols_multi_group_by": "rapidsai/cudf#19421: selecting boolean columns raises a length-mismatch error.", + "tests/frame/select_test.py::test_select_boolean_cols": "NVIDIA/cudf#19421: selecting boolean columns raises a length-mismatch error.", + "tests/frame/select_test.py::test_select_boolean_cols_multi_group_by": "NVIDIA/cudf#19421: selecting boolean columns raises a length-mismatch error.", "tests/frame/top_k_test.py::test_top_k[pandas]": "cudf.pandas top_k returns rows in a different order.", "tests/series_only/from_iterable_test.py::test_series_from_iterable[polars-cudf.pandas.fast_slow_proxy._FunctionProxy-Float64]": "cudf.pandas Series constructor rejects nullable/pyarrow-backed arrays produced via the fast-slow proxy.", "tests/series_only/from_iterable_test.py::test_series_from_iterable[polars-cudf.pandas.fast_slow_proxy._FunctionProxy-Int32]": "cudf.pandas Series constructor rejects nullable/pyarrow-backed arrays produced via the fast-slow proxy.", diff --git a/ci/release/update-version.sh b/ci/release/update-version.sh index 239365ac3e86..7ad912e64c5c 100755 --- a/ci/release/update-version.sh +++ b/ci/release/update-version.sh @@ -140,7 +140,7 @@ fi # Centralized version file update echo "${NEXT_FULL_TAG}" > VERSION -# The cudf version file must be a copy, see https://github.com/rapidsai/cudf/pull/18198 +# The cudf version file must be a copy, see https://github.com/NVIDIA/cudf/pull/18198 echo "${NEXT_FULL_TAG}" > python/cudf/cudf/VERSION echo "${RAPIDS_BRANCH_NAME}" > RAPIDS_BRANCH @@ -237,7 +237,7 @@ sed_runner "s|/blob/\\bmain\\b/|/blob/${RAPIDS_BRANCH_NAME}/|g" python/custreamz # .devcontainer files find .devcontainer/ -type f -name devcontainer.json -print0 | while IFS= read -r -d '' filename; do sed_runner "s@rapidsai/devcontainers:[0-9.]*@rapidsai/devcontainers:${NEXT_SHORT_TAG}@g" "${filename}" - sed_runner "s@ghcr.io/rapidsai/cudf/devcontainer:[0-9.]*@ghcr.io/rapidsai/cudf/devcontainer:${NEXT_SHORT_TAG}@g" "${filename}" + sed_runner "s@ghcr.io/nvidia/cudf/devcontainer:[0-9.]*@ghcr.io/nvidia/cudf/devcontainer:${NEXT_SHORT_TAG}@g" "${filename}" sed_runner "s@rapidsai/devcontainers/features/cuda:[0-9.]*@rapidsai/devcontainers/features/cuda:${NEXT_SHORT_TAG_PEP440}@" "${filename}" sed_runner "s@rapidsai/devcontainers/features/rapids-build-utils:[0-9.]*@rapidsai/devcontainers/features/rapids-build-utils:${NEXT_SHORT_TAG_PEP440}@" "${filename}" sed_runner "s@rapids-\${localWorkspaceFolderBasename}-[0-9.]*@rapids-\${localWorkspaceFolderBasename}-${NEXT_SHORT_TAG}@g" "${filename}" diff --git a/ci/run_cudf_polars_polars_tests.sh b/ci/run_cudf_polars_polars_tests.sh index 740373ae3fae..63fb00844d25 100755 --- a/ci/run_cudf_polars_polars_tests.sh +++ b/ci/run_cudf_polars_polars_tests.sh @@ -33,7 +33,7 @@ DESELECTED_TESTS=( "tests/unit/io/test_write.py::test_write_async[read_parquet-]" # kvikio file creation error in CI "tests/unit/io/test_write.py::test_write_async[-0]" # kvikio file creation error in CI "tests/unit/io/test_write.py::test_write_async[-2]" # kvikio file creation error in CI - "tests/unit/operations/test_random.py::test_shuffle_group_by_reseed" # https://github.com/rapidsai/cudf/issues/22964 + "tests/unit/operations/test_random.py::test_shuffle_group_by_reseed" # https://github.com/NVIDIA/cudf/issues/22964 ) if [[ $(arch) == "aarch64" ]]; then @@ -80,7 +80,7 @@ python "${TIMEOUT_TOOL_PATH}" --enable-python 5400 \ py-polars/tests \ --inject-gpu-engine in-memory -# TODO(ResourceWarning): https://github.com/rapidsai/cudf/issues/22181 +# TODO(ResourceWarning): https://github.com/NVIDIA/cudf/issues/22181 echo "Run polars tests with injected SPMD GPU engine, small blocksize" CUDF_POLARS__EXECUTOR__TARGET_PARTITION_SIZE=805306368 \ CUDF_POLARS__EXECUTOR__FALLBACK_MODE=silent \ diff --git a/conda/recipes/cudf/recipe.yaml b/conda/recipes/cudf/recipe.yaml index 29d60f9eedde..db370ab24315 100644 --- a/conda/recipes/cudf/recipe.yaml +++ b/conda/recipes/cudf/recipe.yaml @@ -96,7 +96,7 @@ requirements: # lives in `dependencies.yaml::depends_on_numba_cuda_mlir`. - numba >=0.60.0,<0.65.0 - numpy >=2.0,<3.0 - - pyarrow>=19.0.0,<24 # https://github.com/rapidsai/cudf/issues/22229 + - pyarrow>=19.0.0,<24 # https://github.com/NVIDIA/cudf/issues/22229 - libcudf =${{ version }} - pylibcudf =${{ version }} - ${{ pin_compatible("rmm", upper_bound="x.x") }} diff --git a/conda/recipes/pylibcudf/recipe.yaml b/conda/recipes/pylibcudf/recipe.yaml index 8edcf4492de5..4040614f9f4d 100644 --- a/conda/recipes/pylibcudf/recipe.yaml +++ b/conda/recipes/pylibcudf/recipe.yaml @@ -95,7 +95,7 @@ requirements: - nvtx >=0.2.1 run_constraints: - numpy >=2.0,<3.0 - - pyarrow>=19.0.0,<24 # https://github.com/rapidsai/cudf/issues/22229 + - pyarrow>=19.0.0,<24 # https://github.com/NVIDIA/cudf/issues/22229 ignore_run_exports: from_package: - cuda-cudart-dev diff --git a/cpp/.clang-tidy b/cpp/.clang-tidy index 3e37a0d64ab9..b6044f0d6d67 100644 --- a/cpp/.clang-tidy +++ b/cpp/.clang-tidy @@ -9,7 +9,7 @@ # Purely stylistic, no benefit to rewriting everything # modernize-return-braced-init-list: # Stylistically we prefer to see the return type at the return site. -# See https://github.com/rapidsai/cudf/pull/16956#pullrequestreview-2341891672 +# See https://github.com/NVIDIA/cudf/pull/16956#pullrequestreview-2341891672 # for more information. # modernize-use-bool-literals: # Our tests use int flags for validity masks extensively and we prefer that diff --git a/cpp/cmake/thirdparty/get_cudf.cmake b/cpp/cmake/thirdparty/get_cudf.cmake index 036f26a7877d..6216ac3f9ecb 100644 --- a/cpp/cmake/thirdparty/get_cudf.cmake +++ b/cpp/cmake/thirdparty/get_cudf.cmake @@ -1,6 +1,6 @@ # ============================================================================= # cmake-format: off -# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 # cmake-format: on # ============================================================================= @@ -37,7 +37,7 @@ function(find_and_configure_cudf VERSION EXPORT_SET) BUILD_EXPORT_SET ${EXPORT_SET} INSTALL_EXPORT_SET ${EXPORT_SET} CPM_ARGS - GIT_REPOSITORY https://github.com/rapidsai/cudf.git + GIT_REPOSITORY https://github.com/NVIDIA/cudf.git GIT_TAG "${RAPIDS_BRANCH}" GIT_SHALLOW TRUE SOURCE_SUBDIR cpp OPTIONS "BUILD_TESTS OFF" "BUILD_BENCHMARKS OFF" diff --git a/cpp/cmake/thirdparty/patches/override.json b/cpp/cmake/thirdparty/patches/override.json index 0d2dede0aad3..60a5bf528536 100644 --- a/cpp/cmake/thirdparty/patches/override.json +++ b/cpp/cmake/thirdparty/patches/override.json @@ -12,7 +12,7 @@ }, { "file": "${current_json_dir}/arrow_thrift_0_23_0.diff", - "issue": "https://github.com/rapidsai/cudf/issues/23454" + "issue": "https://github.com/NVIDIA/cudf/issues/23454" } ], "source_subdir": "cpp" diff --git a/cpp/doxygen/developer_guide/TESTING.md b/cpp/doxygen/developer_guide/TESTING.md index 169eac5de095..87d4108ec6da 100644 --- a/cpp/doxygen/developer_guide/TESTING.md +++ b/cpp/doxygen/developer_guide/TESTING.md @@ -476,7 +476,7 @@ the null mask bits. libcudf employs a custom-built [preload library](https://man7.org/linux/man-pages/man8/ld.so.8.html) to validate its internal stream usage (the code may be found -[`here`](https://github.com/rapidsai/cudf/blob/main/cpp/tests/utilities/identify_stream_usage.cpp)). +[`here`](https://github.com/NVIDIA/cudf/blob/main/cpp/tests/utilities/identify_stream_usage.cpp)). This library wraps every asynchronous CUDA runtime API call that accepts a stream with a check to ensure that the passed CUDA stream is a valid one, immediately throwing an exception if an invalid stream is detected. Running tests with this library loaded immediately triggers errors if any test @@ -512,7 +512,7 @@ mode, however, can only be used to validate tests that are correctly passing `cudf::test::get_default_stream` to public libcudf APIs. In addition to the preload library, the test suite also implements a [custom memory -resource](https://github.com/rapidsai/cudf/blob/main/cpp/include/cudf_test/stream_checking_resource_adaptor.hpp) +resource](https://github.com/NVIDIA/cudf/blob/main/cpp/include/cudf_test/stream_checking_resource_adaptor.hpp) that performs analogous stream verification when its `do_allocate` method is called. During testing this rmm's default memory resource is set to use this adaptor for additional stream validation. diff --git a/cpp/examples/fetch_dependencies.cmake b/cpp/examples/fetch_dependencies.cmake index 23fee0138753..766e85996455 100644 --- a/cpp/examples/fetch_dependencies.cmake +++ b/cpp/examples/fetch_dependencies.cmake @@ -1,6 +1,6 @@ # ============================================================================= # cmake-format: off -# SPDX-FileCopyrightText: Copyright (c) 2023-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 # cmake-format: on # ============================================================================= @@ -19,7 +19,7 @@ include(${CMAKE_BINARY_DIR}/cmake/get_cpm.cmake) CPMFindPackage( NAME cudf FIND_PACKAGE_ARGUMENTS "PATHS ${cudf_ROOT} ${cudf_ROOT}/latest" GIT_REPOSITORY - https://github.com/rapidsai/cudf + https://github.com/NVIDIA/cudf GIT_TAG ${CUDF_TAG} GIT_SHALLOW TRUE diff --git a/cpp/include/cudf/detail/utilities/device_atomics.cuh b/cpp/include/cudf/detail/utilities/device_atomics.cuh index f84a18118da5..8429ced4d98c 100644 --- a/cpp/include/cudf/detail/utilities/device_atomics.cuh +++ b/cpp/include/cudf/detail/utilities/device_atomics.cuh @@ -367,7 +367,7 @@ __device__ __forceinline__ uint64_t calculate_carry_64(uint64_t old_val, __forceinline__ __device__ __int128_t atomic_add(__int128_t* address, __int128_t val) { // CUDA 13.0 miscompiles the native 128-bit CAS on Blackwell; the fix is confirmed in 13.3, so - // gate off all 13.x before 13.3. See https://github.com/rapidsai/cudf/issues/23150. + // gate off all 13.x before 13.3. See https://github.com/NVIDIA/cudf/issues/23150. #if __CUDA_ARCH__ >= 900 && \ !(__CUDA_ARCH__ >= 1000 && __CUDACC_VER_MAJOR__ == 13 && __CUDACC_VER_MINOR__ < 3) __int128_t expected, desired; diff --git a/cpp/include/cudf/detail/utilities/element_argminmax.cuh b/cpp/include/cudf/detail/utilities/element_argminmax.cuh index ba10ff0b1a11..a9bd876cf1e7 100644 --- a/cpp/include/cudf/detail/utilities/element_argminmax.cuh +++ b/cpp/include/cudf/detail/utilities/element_argminmax.cuh @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2022-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2022-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -29,7 +29,7 @@ struct element_argminmax_fn { __device__ inline bool out_of_bounds_or_null(size_type idx) const { - // The extra bounds checking is due to issue github.com/rapidsai/cudf/9156 and + // The extra bounds checking is due to issue github.com/NVIDIA/cudf/9156 and // github.com/NVIDIA/thrust/issues/1525 // where invalid random values may be passed here by thrust::reduce_by_key return idx < 0 || idx >= d_col.size() || (has_nulls && d_col.is_null_nocheck(idx)); diff --git a/cpp/include/cudf/io/datasource.hpp b/cpp/include/cudf/io/datasource.hpp index 67c44cee48aa..010a027bd25a 100644 --- a/cpp/include/cudf/io/datasource.hpp +++ b/cpp/include/cudf/io/datasource.hpp @@ -299,7 +299,7 @@ class datasource { * @param dst Address of the existing device memory * It must not be used asynchronously before the returned future is completed, * because the implementation is not guaranteed to follow stream-ordering. - * See https://github.com/rapidsai/cudf/pull/18279#issuecomment-2727726886 + * See https://github.com/NVIDIA/cudf/pull/18279#issuecomment-2727726886 * @param stream CUDA stream to use * * @return The number of bytes read as a future value (can be smaller than size) diff --git a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/bench_read.cpp b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/bench_read.cpp index 0e3a324805d9..6614fc0aaf5b 100644 --- a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/bench_read.cpp +++ b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/bench_read.cpp @@ -324,7 +324,7 @@ int main(int argc, char** argv) { rapidsmpf::ndsh::FinalizeMPI finalize{}; CUDF_CUDA_TRY(cudaFree(nullptr)); - // work around https://github.com/rapidsai/cudf/issues/20849 + // work around https://github.com/NVIDIA/cudf/issues/20849 cudf::initialize(); auto mr = rmm::mr::cuda_async_memory_resource{}; auto arguments = parse_arguments(argc, argv); diff --git a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q01.cpp b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q01.cpp index 226dd504ff55..9c259404fa20 100644 --- a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q01.cpp +++ b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q01.cpp @@ -264,7 +264,7 @@ int main(int argc, char** argv) { rapidsmpf::ndsh::FinalizeMPI finalize{}; CUDF_CUDA_TRY(cudaFree(nullptr)); - // work around https://github.com/rapidsai/cudf/issues/20849 + // work around https://github.com/NVIDIA/cudf/issues/20849 cudf::initialize(); auto mr = rmm::mr::cuda_async_memory_resource{}; auto arguments = rapidsmpf::ndsh::parse_arguments(argc, argv); diff --git a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q03.cpp b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q03.cpp index 14b2ecbb263c..783014e8ce0e 100644 --- a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q03.cpp +++ b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q03.cpp @@ -340,7 +340,7 @@ int main(int argc, char** argv) { rapidsmpf::ndsh::FinalizeMPI finalize{}; CUDF_CUDA_TRY(cudaFree(nullptr)); - // work around https://github.com/rapidsai/cudf/issues/20849 + // work around https://github.com/NVIDIA/cudf/issues/20849 cudf::initialize(); auto mr = rmm::mr::cuda_async_memory_resource{}; auto arguments = rapidsmpf::ndsh::parse_arguments(argc, argv); diff --git a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q04.cpp b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q04.cpp index 86bc8482507e..23d543291de9 100644 --- a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q04.cpp +++ b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q04.cpp @@ -238,7 +238,7 @@ int main(int argc, char** argv) { rapidsmpf::ndsh::FinalizeMPI finalize{}; CUDF_CUDA_TRY(cudaFree(nullptr)); - // work around https://github.com/rapidsai/cudf/issues/20849 + // work around https://github.com/NVIDIA/cudf/issues/20849 cudf::initialize(); auto mr = rmm::mr::cuda_async_memory_resource{}; auto arguments = rapidsmpf::ndsh::parse_arguments(argc, argv); diff --git a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q09.cpp b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q09.cpp index 7bb35a4115c2..630c1af23c07 100644 --- a/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q09.cpp +++ b/cpp/libcudf_streaming/benchmarks/streaming/ndsh/q09.cpp @@ -319,7 +319,7 @@ int main(int argc, char** argv) { rapidsmpf::ndsh::FinalizeMPI finalize{}; CUDF_CUDA_TRY(cudaFree(nullptr)); - // work around https://github.com/rapidsai/cudf/issues/20849 + // work around https://github.com/NVIDIA/cudf/issues/20849 cudf::initialize(); auto mr = rmm::mr::cuda_async_memory_resource{}; auto arguments = rapidsmpf::ndsh::parse_arguments(argc, argv); diff --git a/cpp/libcudf_streaming/src/parquet.cpp b/cpp/libcudf_streaming/src/parquet.cpp index 361e1739dab2..d4395d766e95 100644 --- a/cpp/libcudf_streaming/src/parquet.cpp +++ b/cpp/libcudf_streaming/src/parquet.cpp @@ -266,7 +266,7 @@ rapidsmpf::streaming::Actor produce_chunks( // Having acquire a ticket, let's move to a new thread. co_await ctx->executor()->schedule(); // TODO: This reads the metadata ntasks times. - // See https://github.com/rapidsai/cudf/issues/20311 + // See https://github.com/NVIDIA/cudf/issues/20311 auto [msg, exception] = [&]() -> std::pair { try { return {read_parquet_chunk(ctx, stream, chunk_options, chunk.sequence_number), nullptr}; diff --git a/cpp/libcudf_streaming/src/partition_utils.cpp b/cpp/libcudf_streaming/src/partition_utils.cpp index 871943190029..99458b77fee9 100644 --- a/cpp/libcudf_streaming/src/partition_utils.cpp +++ b/cpp/libcudf_streaming/src/partition_utils.cpp @@ -59,7 +59,7 @@ std::pair, std::unique_ptr> partition // Notice, the offset argument for split() and hash_partition() doesn't align. // hash_partition() returns the start offset of each partition thus we have to - // skip the first offset. See: . + // skip the first offset. See: . auto partition_offsets = cudf::host_span(offsets.data() + 1, offsets.size() - 2); diff --git a/cpp/src/interop/to_arrow_device.cu b/cpp/src/interop/to_arrow_device.cu index 91c172d0950a..2dc8bfe48c2c 100644 --- a/cpp/src/interop/to_arrow_device.cu +++ b/cpp/src/interop/to_arrow_device.cu @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -389,7 +389,7 @@ int dispatch_to_arrow_device_view::operator()(ArrowArray* out NANOARROW_RETURN_NOT_OK(initialize_array(tmp.get(), nanoarrow_type, column)); if (column.size() == 0) { - // https://github.com/rapidsai/cudf/pull/15047#discussion_r1546528552 + // https://github.com/NVIDIA/cudf/pull/15047#discussion_r1546528552 if (nanoarrow_type == NANOARROW_TYPE_LARGE_STRING) { auto zero = std::make_unique>(0, stream, mr); NANOARROW_RETURN_NOT_OK(set_buffer(std::move(zero), fixed_width_data_buffer_idx, tmp.get())); diff --git a/cpp/src/interop/to_arrow_host.cu b/cpp/src/interop/to_arrow_host.cu index 4b8cb2a6e42d..77685bb281f1 100644 --- a/cpp/src/interop/to_arrow_host.cu +++ b/cpp/src/interop/to_arrow_host.cu @@ -53,7 +53,7 @@ namespace { Enable Transparent Huge Pages (THP) for large (>4MB) allocations. `buf` is returned untouched. Enabling THP can improve performance of device-host memory transfers - significantly, see . + significantly, see . */ void enable_hugepage(ArrowBuffer* buffer) { diff --git a/cpp/src/io/avro/avro_gpu.cu b/cpp/src/io/avro/avro_gpu.cu index 4a0929c0484e..3d0ea0f868d0 100644 --- a/cpp/src/io/avro/avro_gpu.cu +++ b/cpp/src/io/avro/avro_gpu.cu @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ #include "avro_gpu.hpp" @@ -243,7 +243,7 @@ avro_decode_row(schemadesc_s const* schema, } break; // N.B. These aren't handled yet, see the discussion on - // https://github.com/rapidsai/cudf/pull/12788. The decoding logic + // https://github.com/NVIDIA/cudf/pull/12788. The decoding logic // is correct, though, so there's no harm in having them here. case type_timestamp_millis: [[fallthrough]]; case type_timestamp_micros: [[fallthrough]]; diff --git a/cpp/src/io/comp/common.hpp b/cpp/src/io/comp/common.hpp index 4c3212afda40..f67ae396580c 100644 --- a/cpp/src/io/comp/common.hpp +++ b/cpp/src/io/comp/common.hpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2024-2025, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -22,7 +22,7 @@ namespace cudf::io::detail { * * In the meantime, it is not entirely clear why such padding is needed. We need to further * investigate and implement a better fix rather than just padding the buffer. - * See https://github.com/rapidsai/cudf/issues/13605. + * See https://github.com/NVIDIA/cudf/issues/13605. */ constexpr std::size_t BUFFER_PADDING_MULTIPLE{8}; diff --git a/cpp/src/io/orc/stats_enc.cu b/cpp/src/io/orc/stats_enc.cu index 40871ed89571..ae89af34efb8 100644 --- a/cpp/src/io/orc/stats_enc.cu +++ b/cpp/src/io/orc/stats_enc.cu @@ -18,7 +18,7 @@ namespace cudf::io::orc::detail { using strings::detail::fixed_point_string_size; // Nanosecond statistics should not be enabled until the spec version is set correctly in the output -// files. See https://github.com/rapidsai/cudf/issues/14325 for more details +// files. See https://github.com/NVIDIA/cudf/issues/14325 for more details constexpr bool enable_nanosecond_statistics = true; constexpr unsigned int init_threads_per_group = 32; diff --git a/cpp/src/io/orc/writer_impl.cu b/cpp/src/io/orc/writer_impl.cu index 62ff8f165c3e..50b4acd4f804 100644 --- a/cpp/src/io/orc/writer_impl.cu +++ b/cpp/src/io/orc/writer_impl.cu @@ -934,7 +934,7 @@ encoded_data encode_columns(orc_table_view const& orc_table, [](auto valid_count) { return valid_count % 8; }), "There's currently a bug in encoding boolean columns. Suggested workaround is to convert " "to int8 type." - " Please see https://github.com/rapidsai/cudf/issues/6763 for more information."); + " Please see https://github.com/NVIDIA/cudf/issues/6763 for more information."); } hostdevice_2dvector chunk_streams( diff --git a/cpp/src/io/parquet/experimental/dictionary_page_filter.cu b/cpp/src/io/parquet/experimental/dictionary_page_filter.cu index 501d2cbab355..3df63facff25 100644 --- a/cpp/src/io/parquet/experimental/dictionary_page_filter.cu +++ b/cpp/src/io/parquet/experimental/dictionary_page_filter.cu @@ -198,7 +198,7 @@ __device__ __forceinline__ bool is_error_set(kernel_error::pointer error) __device__ __forceinline__ int32_t get_int32_type_len(LogicalType const& logical_type) { // Note: This function has been extracted from the snippet at: - // https://github.com/rapidsai/cudf/blob/c89c83c00c729a86c56570693b627f31408bc2c9/cpp/src/io/parquet/page_decode.cuh#L1278-L1287 + // https://github.com/NVIDIA/cudf/blob/c89c83c00c729a86c56570693b627f31408bc2c9/cpp/src/io/parquet/page_decode.cuh#L1278-L1287 // Check for smaller bitwidths if (logical_type.type == LogicalType::INTEGER) { @@ -223,7 +223,7 @@ __device__ __forceinline__ void decode_int96timestamp(uint8_t const* int96_ptr, int64_t* timestamp64) { // Note: This function has been modified from the original at - // https://github.com/rapidsai/cudf/blob/c89c83c00c729a86c56570693b627f31408bc2c9/cpp/src/io/parquet/page_data.cuh#L133-L198 + // https://github.com/NVIDIA/cudf/blob/c89c83c00c729a86c56570693b627f31408bc2c9/cpp/src/io/parquet/page_data.cuh#L133-L198 int64_t nanos = cudf::io::unaligned_load(int96_ptr); int64_t days = cudf::io::unaligned_load(int96_ptr + sizeof(int64_t)); @@ -447,7 +447,7 @@ __device__ T decode_fixed_width_value(PageInfo const& page, // Handle durations else if constexpr (cudf::is_duration()) { // Note: This function has been extracted from the snippet at: - // https://github.com/rapidsai/cudf/blob/594d26768ce86b9c2f389e851ae1afb77032c879/cpp/src/io/parquet/decode_fixed.cu#L159-L163 + // https://github.com/NVIDIA/cudf/blob/594d26768ce86b9c2f389e851ae1afb77032c879/cpp/src/io/parquet/decode_fixed.cu#L159-L163 // Reading INT32 TIME_MILLIS into 64-bit DURATION_MILLISECONDS // TIME_MILLIS is the only duration type stored as int32: diff --git a/cpp/src/io/parquet/experimental/hybrid_scan_impl.cpp b/cpp/src/io/parquet/experimental/hybrid_scan_impl.cpp index aab25c0e648b..c89cb798abc1 100644 --- a/cpp/src/io/parquet/experimental/hybrid_scan_impl.cpp +++ b/cpp/src/io/parquet/experimental/hybrid_scan_impl.cpp @@ -147,7 +147,7 @@ void hybrid_scan_reader_impl::select_columns(read_columns_mode read_columns_mode // Select only columns required by the options and filter. // Using as is from: - // https://github.com/rapidsai/cudf/blob/a8b25cd205dc5d04b9918dcb0b3abd6b8c4e4a74/cpp/src/io/parquet/reader_impl.cpp#L556-L569 + // https://github.com/NVIDIA/cudf/blob/a8b25cd205dc5d04b9918dcb0b3abd6b8c4e4a74/cpp/src/io/parquet/reader_impl.cpp#L556-L569 std::optional> filter_only_columns_names; if (options.get_filter().has_value() and select_column_names.has_value()) { filter_only_columns_names = parquet::detail::get_column_names_in_expression( diff --git a/cpp/src/io/parquet/reader_impl_helpers.hpp b/cpp/src/io/parquet/reader_impl_helpers.hpp index ed67ca05cfc3..37429364016d 100644 --- a/cpp/src/io/parquet/reader_impl_helpers.hpp +++ b/cpp/src/io/parquet/reader_impl_helpers.hpp @@ -357,7 +357,7 @@ class aggregate_reader_metadata { * Returns true iff at least one column chunk referenced by `filter_column_schemas` in the first * selected row group of any source carries any of `min` / `max` / `min_value` / `max_value` / * `null_count`. Inspecting one row group per source is sufficient; see - * https://github.com/rapidsai/cudf/pull/22664#issuecomment-4557500237. + * https://github.com/NVIDIA/cudf/pull/22664#issuecomment-4557500237. * * @param input_row_group_indices Selected row group indices, one vector per source * @param filter_column_schemas Zeroth-source schema indices of the columns referenced by the diff --git a/cpp/src/io/parquet/reader_impl_preprocess_utils.cu b/cpp/src/io/parquet/reader_impl_preprocess_utils.cu index 4113ed67f711..6d4b9f644c09 100644 --- a/cpp/src/io/parquet/reader_impl_preprocess_utils.cu +++ b/cpp/src/io/parquet/reader_impl_preprocess_utils.cu @@ -235,7 +235,7 @@ void generate_depth_remappings( // It's required to ignore unsupported encodings in this function // so that we can actually compile a list of all the unsupported encodings found // in the pages. That cannot be done here since we do not have the pages vector here. - // see https://github.com/rapidsai/cudf/pull/14453#pullrequestreview-1778346688 + // see https://github.com/NVIDIA/cudf/pull/14453#pullrequestreview-1778346688 if (auto const error = error_code.value_sync(stream); error != 0 and error != static_cast(decode_error::UNSUPPORTED_ENCODING)) { CUDF_FAIL("Parquet header parsing failed with code(s) while counting page headers " + diff --git a/cpp/src/io/parquet/writer_impl.cu b/cpp/src/io/parquet/writer_impl.cu index 30d8c2171e56..2ec556c2c81a 100644 --- a/cpp/src/io/parquet/writer_impl.cu +++ b/cpp/src/io/parquet/writer_impl.cu @@ -2738,7 +2738,7 @@ std::unique_ptr> writer::merge_row_group_metadata( // Remove any LogicalType::UNKNOWN annotations that were passed in as they can confuse // column type inferencing. - // See https://github.com/rapidsai/cudf/pull/14264#issuecomment-1778311615 + // See https://github.com/NVIDIA/cudf/pull/14264#issuecomment-1778311615 for (auto& se : md.schema) { if (se.logical_type.has_value() && se.logical_type.value().type == LogicalType::UNKNOWN) { se.logical_type = cuda::std::nullopt; diff --git a/cpp/src/io/statistics/column_statistics.cuh b/cpp/src/io/statistics/column_statistics.cuh index 27d886f832d4..638146c63018 100644 --- a/cpp/src/io/statistics/column_statistics.cuh +++ b/cpp/src/io/statistics/column_statistics.cuh @@ -88,7 +88,7 @@ struct calculate_group_statistics_functor { (IO != detail::io_file_format::PARQUET or !std::is_same_v)) { // Temporarily disable stats writing for int96 timestamps - // TODO: https://github.com/rapidsai/cudf/issues/10438 + // TODO: https://github.com/NVIDIA/cudf/issues/10438 if constexpr (cudf::is_timestamp() and IO == detail::io_file_format::PARQUET and INT96 == detail::is_int96_timestamp::YES) { return; @@ -302,7 +302,7 @@ CUDF_KERNEL void __launch_bounds__(block_size, 1) threadIdx.x); } // Temporarily disable stats writing for int96 timestamps - // TODO: https://github.com/rapidsai/cudf/issues/10438 + // TODO: https://github.com/NVIDIA/cudf/issues/10438 else { type_dispatcher( state.col.leaf_column->type(), diff --git a/cpp/src/io/utilities/config_utils.cpp b/cpp/src/io/utilities/config_utils.cpp index c07ad634279a..a7ba392cf724 100644 --- a/cpp/src/io/utilities/config_utils.cpp +++ b/cpp/src/io/utilities/config_utils.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -20,7 +20,7 @@ void set_up_kvikio() { static std::once_flag flag{}; std::call_once(flag, [] { - // Workaround for https://github.com/rapidsai/cudf/issues/14140, where cuFileDriverOpen errors + // Workaround for https://github.com/NVIDIA/cudf/issues/14140, where cuFileDriverOpen errors // out if no CUDA calls have been made before it. This is a no-op if the CUDA context is already // initialized. CUDF_CUDA_TRY(cudaFree(nullptr)); diff --git a/cpp/src/io/utilities/data_sink.cpp b/cpp/src/io/utilities/data_sink.cpp index d2f89c454c79..1876fecf0773 100644 --- a/cpp/src/io/utilities/data_sink.cpp +++ b/cpp/src/io/utilities/data_sink.cpp @@ -113,7 +113,7 @@ class host_buffer_sink : public data_sink { { auto const current_size = buffer_->size(); buffer_->resize(current_size + size); - // TODO: https://github.com/rapidsai/cudf/issues/21680 + // TODO: https://github.com/NVIDIA/cudf/issues/21680 // We need to replace this with memcpy_batch_async after fixing stream // ordering. The issue is that buffer_->resize() can reallocate, // invalidating pointers from previous async copies that are still diff --git a/cpp/src/jit/row_ir.cpp b/cpp/src/jit/row_ir.cpp index 74d4f581a243..73a4f065fee2 100644 --- a/cpp/src/jit/row_ir.cpp +++ b/cpp/src/jit/row_ir.cpp @@ -42,7 +42,7 @@ enum class [[nodiscard]] null_output : uint8_t { { // this implementation must be kept in sync with the implementation of // `cudf::ast::detail::operator_functor::fixed_point_result_scale` - // (https://github.com/rapidsai/cudf/blob/a5dccda20a74fe61e3c4491b0e74bdc0321d60d5/cpp/include/cudf/ast/detail/operator_functor.cuh#L161) + // (https://github.com/NVIDIA/cudf/blob/a5dccda20a74fe61e3c4491b0e74bdc0321d60d5/cpp/include/cudf/ast/detail/operator_functor.cuh#L161) switch (op) { // pseudo-opcode with no argument case opcode::GET_INPUT: return 0; diff --git a/cpp/src/partitioning/round_robin.cu b/cpp/src/partitioning/round_robin.cu index 5a6d3080fa0e..d314723b3cb7 100644 --- a/cpp/src/partitioning/round_robin.cu +++ b/cpp/src/partitioning/round_robin.cu @@ -178,7 +178,7 @@ std::pair, std::vector> round_robin_part auto np_max_size = nrows % num_partitions; // # partitions of max size // handle case when nr `mod` np == 0; - // fix for bug: https://github.com/rapidsai/cudf/issues/4043 + // fix for bug: https://github.com/NVIDIA/cudf/issues/4043 auto num_partitions_max_size = (np_max_size > 0 ? np_max_size : num_partitions); cudf::size_type max_partition_size = std::ceil( diff --git a/cpp/src/quantiles/quantiles_util.hpp b/cpp/src/quantiles/quantiles_util.hpp index 991e5766e959..d4362adda711 100644 --- a/cpp/src/quantiles/quantiles_util.hpp +++ b/cpp/src/quantiles/quantiles_util.hpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2025, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -27,7 +27,7 @@ CUDF_HOST_DEVICE inline Result linear(T lhs, T rhs, double frac) // Since the fraction part of double is 52 bits, // double cannot fully represent int64. // Underflow may occur when converting int64 to double - // detail: https://github.com/rapidsai/cudf/issues/1417 + // detail: https://github.com/NVIDIA/cudf/issues/1417 auto dlhs = convert_to_floating(lhs); auto drhs = convert_to_floating(rhs); diff --git a/cpp/src/reductions/nested_types_extrema_utils.cuh b/cpp/src/reductions/nested_types_extrema_utils.cuh index 78b40fe7624c..0e628800478a 100644 --- a/cpp/src/reductions/nested_types_extrema_utils.cuh +++ b/cpp/src/reductions/nested_types_extrema_utils.cuh @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2022-2025, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2022-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -37,7 +37,7 @@ struct row_arg_minmax_fn { // used with `thrust::reduce_by_key` or `thrust::scan_by_key`. __attribute__((noinline)) __device__ auto operator()(size_type lhs_idx, size_type rhs_idx) const { - // The extra bounds checking is due to issue github.com/rapidsai/cudf/issues/9156 and + // The extra bounds checking is due to issue github.com/NVIDIA/cudf/issues/9156 and // github.com/NVIDIA/thrust/issues/1525 // where invalid random values may be passed here by thrust::reduce_by_key if (lhs_idx < 0 || lhs_idx >= num_rows) { return rhs_idx; } diff --git a/cpp/src/search/contains_table_impl.cuh b/cpp/src/search/contains_table_impl.cuh index 792b9aa000f1..4eb7392d4f8d 100644 --- a/cpp/src/search/contains_table_impl.cuh +++ b/cpp/src/search/contains_table_impl.cuh @@ -149,8 +149,8 @@ void perform_contains(table_view const& haystack, // If the haystack table has nulls but they are compared unequal, don't insert them. // Otherwise, it was known to cause performance issue: - // - https://github.com/rapidsai/cudf/pull/6943 - // - https://github.com/rapidsai/cudf/pull/8277 + // - https://github.com/NVIDIA/cudf/pull/6943 + // - https://github.com/NVIDIA/cudf/pull/8277 set.insert_if_async(haystack_iter, haystack_iter + haystack.num_rows(), cuda::counting_iterator{0}, // stencil diff --git a/cpp/src/strings/copying/concatenate.cu b/cpp/src/strings/copying/concatenate.cu index bfde1f742698..e6fc99748cd8 100644 --- a/cpp/src/strings/copying/concatenate.cu +++ b/cpp/src/strings/copying/concatenate.cu @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -33,7 +33,7 @@ namespace cudf { namespace strings { namespace detail { -// Benchmark data, shared at https://github.com/rapidsai/cudf/pull/4703, shows +// Benchmark data, shared at https://github.com/NVIDIA/cudf/pull/4703, shows // that the single kernel optimization generally performs better, but when the // number of chars/col is beyond a certain threshold memcpy performs better. // This heuristic estimates which strategy will give better performance by diff --git a/cpp/src/strings/regex/regcomp.cpp b/cpp/src/strings/regex/regcomp.cpp index ecc7635ac3c5..7a33bb3ff411 100644 --- a/cpp/src/strings/regex/regcomp.cpp +++ b/cpp/src/strings/regex/regcomp.cpp @@ -1162,7 +1162,7 @@ void reprog::build_start_ids() * @brief Check a specific instruction for errors. * * Currently this is checking for an infinite-loop condition as documented in this issue: - * https://github.com/rapidsai/cudf/issues/10006 + * https://github.com/NVIDIA/cudf/issues/10006 * * Example instructions list created from pattern `(A?)+` * ``` diff --git a/cpp/tests/ast/transform_tests.cpp b/cpp/tests/ast/transform_tests.cpp index fe4960b9dfec..437788870761 100644 --- a/cpp/tests/ast/transform_tests.cpp +++ b/cpp/tests/ast/transform_tests.cpp @@ -1221,7 +1221,7 @@ using DecimalArithmeticParams = cudf::test::CrossProduct>; TYPED_TEST_SUITE(DecimalTests, DecimalArithmeticParams); -// Regression test for https://github.com/rapidsai/cudf/issues/21980 +// Regression test for https://github.com/NVIDIA/cudf/issues/21980 // Nested decimal expressions lose scale in intermediate return types, // causing "non-matching operand types" at parse time. TYPED_TEST(DecimalTests, NestedDecimalArithmetic) diff --git a/cpp/tests/column/column_test.cpp b/cpp/tests/column/column_test.cpp index 2b3da507b858..75d075b80be8 100644 --- a/cpp/tests/column/column_test.cpp +++ b/cpp/tests/column/column_test.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -631,7 +631,7 @@ TYPED_TEST(ListsColumnTest, ListsSlicedColumnViewConstructorWithNulls) // TODO: null mask equality is being checked separately because // expect_columns_equal doesn't do the check for lists columns. - // This is fixed in https://github.com/rapidsai/cudf/pull/5904, + // This is fixed in https://github.com/NVIDIA/cudf/pull/5904, // so we should remove this check after that's merged: CUDF_TEST_EXPECT_COLUMNS_EQUAL( cudf::mask_to_bools(result->view().null_mask(), 0, 4)->view(), diff --git a/cpp/tests/copying/split_tests.cpp b/cpp/tests/copying/split_tests.cpp index 72faa23df0c2..45ceeb728322 100644 --- a/cpp/tests/copying/split_tests.cpp +++ b/cpp/tests/copying/split_tests.cpp @@ -1754,7 +1754,7 @@ TEST_F(ContiguousSplitUntypedTest, DISABLED_VeryLargeColumnTestChunked) } // Disabled as this test requires about 7GB of device memory. -// See https://github.com/rapidsai/cudf/issues/20876 for more info. +// See https://github.com/NVIDIA/cudf/issues/20876 for more info. TEST_F(ContiguousSplitUntypedTest, DISABLED_ChunkedPackNextReturnValueOver2GB) { auto const mr = cudf::get_current_device_resource_ref(); diff --git a/cpp/tests/groupby/max_tests.cpp b/cpp/tests/groupby/max_tests.cpp index 3cb6b308648b..538a11ba2039 100644 --- a/cpp/tests/groupby/max_tests.cpp +++ b/cpp/tests/groupby/max_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -516,7 +516,7 @@ TYPED_TEST(groupby_max_floating_point_test, values_with_infinity) auto const expected_keys = int32s_col{1, 2}; auto const expected_vals = floats_col{inf, static_cast(2)}; - // Related issue: https://github.com/rapidsai/cudf/issues/11352 + // Related issue: https://github.com/NVIDIA/cudf/issues/11352 // The issue only occurs in sort-based cudf::aggregation. auto agg = cudf::make_max_aggregation(); test_single_agg( diff --git a/cpp/tests/groupby/min_tests.cpp b/cpp/tests/groupby/min_tests.cpp index 429de1c38672..8f59f8b59cf2 100644 --- a/cpp/tests/groupby/min_tests.cpp +++ b/cpp/tests/groupby/min_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -508,7 +508,7 @@ TYPED_TEST(groupby_min_floating_point_test, values_with_infinity) auto const expected_keys = int32s_col{1, 2}; auto const expected_vals = floats_col{-inf, static_cast(1)}; - // Related issue: https://github.com/rapidsai/cudf/issues/11352 + // Related issue: https://github.com/NVIDIA/cudf/issues/11352 // The issue only occurs in sort-based aggregation. auto agg = cudf::make_min_aggregation(); test_single_agg( diff --git a/cpp/tests/groupby/rank_scan_tests.cpp b/cpp/tests/groupby/rank_scan_tests.cpp index a123093fd528..b0d3a8443a38 100644 --- a/cpp/tests/groupby/rank_scan_tests.cpp +++ b/cpp/tests/groupby/rank_scan_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2021-2024, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -395,7 +395,7 @@ TYPED_TEST(typed_groupby_rank_scan_test, structsWithNullPushdown) CUDF_TEST_EXPECT_COLUMNS_EQUIVALENT(*agg_results[1].results[2], expected_percent_for_null); } -/* List support dependent on https://github.com/rapidsai/cudf/issues/8683 +/* List support dependent on https://github.com/NVIDIA/cudf/issues/8683 template struct list_groupby_rank_scan_test : public cudf::test::BaseFixture { }; diff --git a/cpp/tests/groupby/sum_tests.cpp b/cpp/tests/groupby/sum_tests.cpp index ef72f3f40742..8205a7ddcd1b 100644 --- a/cpp/tests/groupby/sum_tests.cpp +++ b/cpp/tests/groupby/sum_tests.cpp @@ -284,7 +284,7 @@ TEST_F(GroupByDecimal128ShmemAlignmentTest, Decimal128SumAfterInt32Sum) CUDF_TEST_EXPECT_COLUMNS_EQUIVALENT(expected, sorted_result->get_column(0)); } -// Regression test for https://github.com/rapidsai/cudf/issues/23150. +// Regression test for https://github.com/NVIDIA/cudf/issues/23150. // Blackwell returned incorrect sums when aggregating three DECIMAL128 columns. TEST_F(GroupByDecimal128ShmemAlignmentTest, MultiColumnDecimal128Sum) { diff --git a/cpp/tests/io/json/json_test.cpp b/cpp/tests/io/json/json_test.cpp index 161dca4f1669..545163a94a55 100644 --- a/cpp/tests/io/json/json_test.cpp +++ b/cpp/tests/io/json/json_test.cpp @@ -3681,7 +3681,7 @@ TEST_F(JsonReaderTest, DeviceWriteAsyncThrows) // schema. cuDF itself does NOT change the column contents on mismatch (existing // CHILD_NULL_ONLY behavior); the diagnostic is reported on a separate side channel so the // public `column_name_info`/`table_metadata` ABI is unaffected. -// See https://github.com/rapidsai/cudf/issues/22423. +// See https://github.com/NVIDIA/cudf/issues/22423. // Build schema for: struct>>. // Pin column order so cuDF emits columns in (c1, c2) regardless of JSON-encounter order. diff --git a/cpp/tests/io/orc_test.cpp b/cpp/tests/io/orc_test.cpp index 07b610acc291..a5b4476c5fc3 100644 --- a/cpp/tests/io/orc_test.cpp +++ b/cpp/tests/io/orc_test.cpp @@ -629,7 +629,7 @@ TEST_F(OrcWriterTest, negTimestampsNano) // This is a separate test because ORC format has a bug where writing a timestamp between -1 and 0 // seconds from UNIX epoch is read as that timestamp + 1 second. We mimic that behavior and so // this test has to hardcode test values which are < -1 second. - // Details: https://github.com/rapidsai/cudf/pull/5529#issuecomment-648768925 + // Details: https://github.com/NVIDIA/cudf/pull/5529#issuecomment-648768925 auto timestamps_ns = cudf::test::fixed_width_column_wrapper{ -131968727238000000, @@ -1195,7 +1195,7 @@ TEST_F(OrcWriterTest, SlicedValidMask) TEST_F(OrcReaderTest, ZeroColumnsPreservesRowCount) { GTEST_SKIP() << "Zero-column / N-row ORC reads are not yet supported. See " - "https://github.com/rapidsai/cudf/issues/22935)."; + "https://github.com/NVIDIA/cudf/issues/22935)."; constexpr cudf::size_type num_rows = 8; cudf::test::fixed_width_column_wrapper col{0, 1, 2, 3, 4, 5, 6, 7}; diff --git a/cpp/tests/io/parquet_reader_test.cpp b/cpp/tests/io/parquet_reader_test.cpp index 08e48dd213eb..fb0fea53a73a 100644 --- a/cpp/tests/io/parquet_reader_test.cpp +++ b/cpp/tests/io/parquet_reader_test.cpp @@ -131,7 +131,7 @@ TEST_F(ParquetReaderTest, UserBounds) TEST_F(ParquetReaderTest, ZeroColumnsPreservesRowCount) { GTEST_SKIP() << "Zero-column / N-row parquet reads are not yet supported. See " - "https://github.com/rapidsai/cudf/issues/22935)."; + "https://github.com/NVIDIA/cudf/issues/22935)."; srand(31337); auto const num_rows = 8; @@ -2841,7 +2841,7 @@ TEST_F(ParquetReaderTest, RepeatedNoAnnotations) CUDF_TEST_EXPECT_TABLES_EQUAL(result.tbl->view(), expected); } -// Regression test for https://github.com/rapidsai/cudf/issues/22541. +// Regression test for https://github.com/NVIDIA/cudf/issues/22541. // Schema (single-field inner repeated group must decode as list>): // required group root { // optional int32 primitive; @@ -4070,7 +4070,7 @@ void filter_unary_operation_typed_test() auto const expected = cudf::apply_boolean_mask(written_table, *predicate); // JIT does not support nullness-dependent operators such as IS_NULL - // Ref: https://github.com/rapidsai/cudf/issues/20177 + // Ref: https://github.com/NVIDIA/cudf/issues/20177 auto constexpr use_jit = false; // Reading with Predicate Pushdown @@ -4391,8 +4391,8 @@ TYPED_TEST(ParquetPredicatePushdownTestJIT, FilterTyped) filter_typed_test(); // JIT does not support decimals and nullness-dependent operators (IS_NULL) so we can't test: // `filter_unary_operation_typed_test()` and `decimal_stats_filter_test()`. - // Refs: https://github.com/rapidsai/cudf/issues/20177 and - // https://github.com/rapidsai/cudf/issues/21584 + // Refs: https://github.com/NVIDIA/cudf/issues/20177 and + // https://github.com/NVIDIA/cudf/issues/21584 } TEST_P(ParquetDecompressionTest, RoundTripBasic) diff --git a/cpp/tests/io/parquet_writer_test.cpp b/cpp/tests/io/parquet_writer_test.cpp index ec0267c3bfea..fe5745e254d1 100644 --- a/cpp/tests/io/parquet_writer_test.cpp +++ b/cpp/tests/io/parquet_writer_test.cpp @@ -934,7 +934,7 @@ TEST_F(ParquetWriterTest, FloatingPointWithNaNStatsOmittedAcrossFragments) TEST_F(ParquetWriterTest, FloatingPointWithNaNStatsOmittedNested) { - // NaN detection must reach a float leaf nested in a LIST column (rapidsai/cudf#22817). + // NaN detection must reach a float leaf nested in a LIST column (NVIDIA/cudf#22817). auto constexpr nanf = std::numeric_limits::quiet_NaN(); cudf::test::lists_column_wrapper list_col{{1.0f, nanf, 3.0f}, {4.0f, 5.0f}}; auto const expected = table_view{{list_col}}; diff --git a/cpp/tests/lists/combine/concatenate_list_elements_tests.cpp b/cpp/tests/lists/combine/concatenate_list_elements_tests.cpp index 849cb79f4932..6e758a782d65 100644 --- a/cpp/tests/lists/combine/concatenate_list_elements_tests.cpp +++ b/cpp/tests/lists/combine/concatenate_list_elements_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -797,7 +797,7 @@ TEST_F(ConcatenateListElementsTest, ListsOfListsOfStructsHavingListsWithNulls) } } -// Tests for the child.size()==0 early-exit guard added to fix rapidsai/cudf#22146. +// Tests for the child.size()==0 early-exit guard added to fix NVIDIA/cudf#22146. // When the inner list column has 0 rows, the offsets buffer may be unallocated; // the guard must build the result without dereferencing that buffer. @@ -861,7 +861,7 @@ TEST_F(ConcatenateListElementsTest, EmptyInnerListColumnWithNullRow) // Subcase D: child.size()==0 with a struct grandchild — verifies the guard preserves // child column structure (field columns) when the grandchild is a struct type. -// Addresses review feedback on rapidsai/cudf#22147 (pmattione-nvidia). +// Addresses review feedback on NVIDIA/cudf#22147 (pmattione-nvidia). TEST_F(ConcatenateListElementsTest, EmptyInnerListColumnChildSizeZeroWithStructGrandchild) { // list>> with 1 outer row that is an empty list-of-lists. diff --git a/cpp/tests/sort/rank_test.cpp b/cpp/tests/sort/rank_test.cpp index 6b0e3f3d5b52..27a24033e77c 100644 --- a/cpp/tests/sort/rank_test.cpp +++ b/cpp/tests/sort/rank_test.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -430,7 +430,7 @@ struct RankLarge : public cudf::test::BaseFixture {}; TEST_F(RankLarge, average_large) { - // testcase of https://github.com/rapidsai/cudf/issues/9703 + // testcase of https://github.com/NVIDIA/cudf/issues/9703 auto iter = cuda::counting_iterator{0}; cudf::test::fixed_width_column_wrapper col1(iter, iter + 10558); auto result = cudf::rank(col1, diff --git a/cpp/tests/sort/segmented_sort_tests.cpp b/cpp/tests/sort/segmented_sort_tests.cpp index 66763e27a77c..d778a4ba589e 100644 --- a/cpp/tests/sort/segmented_sort_tests.cpp +++ b/cpp/tests/sort/segmented_sort_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2021-2025, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -310,7 +310,7 @@ TEST_F(SegmentedSortInt, ErrorsMismatchArgSizes) CUDF_EXPECT_NO_THROW(cudf::stable_segmented_sort_by_key(input1, input1, segments)); } -// Test specifically verifies the patch added in https://github.com/rapidsai/cudf/pull/12234 +// Test specifically verifies the patch added in https://github.com/NVIDIA/cudf/pull/12234 // This test will fail if the CUB bug fix is not available or the patch has not been applied. TEST_F(SegmentedSortInt, Bool) { @@ -332,7 +332,7 @@ TEST_F(SegmentedSortInt, Bool) CUDF_TEST_EXPECT_COLUMNS_EQUIVALENT(result->view(), expected); } -// Specific test for fix in https://github.com/rapidsai/cudf/pull/16463 +// Specific test for fix in https://github.com/NVIDIA/cudf/pull/16463 TEST_F(SegmentedSortInt, UnbalancedOffsets) { auto h_input = std::vector(3535); diff --git a/cpp/tests/stream_compaction/drop_nulls_tests.cpp b/cpp/tests/stream_compaction/drop_nulls_tests.cpp index 88f023f5cdba..55267ed4bd09 100644 --- a/cpp/tests/stream_compaction/drop_nulls_tests.cpp +++ b/cpp/tests/stream_compaction/drop_nulls_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2024, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -83,7 +83,7 @@ TEST_F(DropNullsTest, MixedSetOfRows) TEST_F(DropNullsTest, LargeColumn) { // This test is a C++ repro of the failing Python in this issue: - // https://github.com/rapidsai/cudf/issues/5456 + // https://github.com/NVIDIA/cudf/issues/5456 // Specifically, there are two large columns, one nullable, one non-nullable using T = int32_t; using index_T = int64_t; diff --git a/cpp/tests/strings/replace_regex_tests.cpp b/cpp/tests/strings/replace_regex_tests.cpp index 8422264d4c36..dcdbd8ccb076 100644 --- a/cpp/tests/strings/replace_regex_tests.cpp +++ b/cpp/tests/strings/replace_regex_tests.cpp @@ -1,5 +1,5 @@ /* - * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 */ @@ -379,7 +379,7 @@ TEST_F(StringsReplaceRegexTest, ReplaceBackrefsWithEmptyCapture) cudf::test::strings_column_wrapper input({"one\ntwo", "three\n\n", "four\r\n"}); auto sv = cudf::strings_column_view(input); - // https://github.com/rapidsai/cudf/issues/13404 + // https://github.com/NVIDIA/cudf/issues/13404 auto pattern = std::string("(\r\n|\r)?$"); auto repl_template = std::string("[\\1]"); auto expected = @@ -388,7 +388,7 @@ TEST_F(StringsReplaceRegexTest, ReplaceBackrefsWithEmptyCapture) auto results = cudf::strings::replace_with_backrefs(sv, *prog, repl_template); CUDF_TEST_EXPECT_COLUMNS_EQUIVALENT(*results, expected); - // https://github.com/rapidsai/cudf/issues/22707 + // https://github.com/NVIDIA/cudf/issues/22707 pattern = std::string("^(a?)"); expected = cudf::test::strings_column_wrapper({"[]one\ntwo", "[]three\n\n", "[]four\r\n"}); prog = cudf::strings::regex_program::create(pattern); diff --git a/dependencies.yaml b/dependencies.yaml index 35044134c583..662b508b0535 100644 --- a/dependencies.yaml +++ b/dependencies.yaml @@ -832,10 +832,10 @@ dependencies: common: - output_types: [conda] packages: - - pyarrow>=19.0.0,<24 # https://github.com/rapidsai/cudf/issues/22229 + - pyarrow>=19.0.0,<24 # https://github.com/NVIDIA/cudf/issues/22229 - output_types: [requirements, pyproject] packages: - - pyarrow>=19.0.0,<24 # https://github.com/rapidsai/cudf/issues/22229 + - pyarrow>=19.0.0,<24 # https://github.com/NVIDIA/cudf/issues/22229 # 'cuda_version' intentionally does not contain fallback entries... we want # a loud error if an unsupported 'cuda' value is passed cuda_version: @@ -959,7 +959,7 @@ dependencies: # We always want nvJitLink >= whatever minor version was available in the build environment # nvJitLink tends to follow the CTK's minor versions, but check at https://docs.nvidia.com/cuda/cuda-toolkit-release-notes/index.html # - # ref: https://github.com/rapidsai/cudf/pull/20873 + # ref: https://github.com/NVIDIA/cudf/pull/20873 - matrix: cuda: "12.9" use_cuda_wheels: "true" @@ -1272,7 +1272,7 @@ dependencies: - output_types: [conda, requirements, pyproject] packages: # Upper bound on aiobotocore to avoid build issues in CI - # https://github.com/rapidsai/cudf/issues/20843 + # https://github.com/NVIDIA/cudf/issues/20843 - aiobotocore>=2.2.0,<3 - boto3>=1.21.21 - botocore>=1.24.21 diff --git a/docs/cudf/README.md b/docs/cudf/README.md index 7ca75680a008..960bf4c5657e 100644 --- a/docs/cudf/README.md +++ b/docs/cudf/README.md @@ -3,4 +3,4 @@ This directory contains the documentation of cuDF Python. For more information on how to write, build, and read the documentation, see -[the developer documentation](https://github.com/rapidsai/cudf/blob/HEAD/docs/cudf/source/cudf/developer_guide/documentation.md). +[the developer documentation](https://github.com/NVIDIA/cudf/blob/HEAD/docs/cudf/source/cudf/developer_guide/documentation.md). diff --git a/docs/cudf/source/conf.py b/docs/cudf/source/conf.py index 4b6cc014a71e..23990f461111 100644 --- a/docs/cudf/source/conf.py +++ b/docs/cudf/source/conf.py @@ -227,7 +227,7 @@ def clean_all_xml_files(path): html_theme_options = { "external_links": [], "icon_links": [], - "github_url": "https://github.com/rapidsai/cudf", + "github_url": "https://github.com/NVIDIA/cudf", "show_toc_level": 1, "navbar_align": "content", "navbar_center": "navbar-nav, version-switcher, navbar-external-links", @@ -739,7 +739,7 @@ def linkcode_resolve(domain, info) -> str | None: fn = os.path.relpath(fn, start=os.path.dirname(pkg_file)) return ( - f"https://github.com/rapidsai/cudf/blob/" + f"https://github.com/NVIDIA/cudf/blob/" f"{RAPIDS_BRANCH}/{source_path}/{fn}{linespec}" ) diff --git a/docs/cudf/source/cudf/10min.ipynb b/docs/cudf/source/cudf/10min.ipynb index 4692f07ef976..f43b657869b8 100644 --- a/docs/cudf/source/cudf/10min.ipynb +++ b/docs/cudf/source/cudf/10min.ipynb @@ -11,15 +11,15 @@ "\n", "## What are these Libraries?\n", "\n", - "[cuDF](https://github.com/rapidsai/cudf) is a Python GPU DataFrame library (built on the Apache Arrow columnar memory format) for loading, joining, aggregating, filtering, and otherwise manipulating tabular data using a DataFrame style API in the style of [pandas](https://pandas.pydata.org).\n", + "[cuDF](https://github.com/NVIDIA/cudf) is a Python GPU DataFrame library (built on the Apache Arrow columnar memory format) for loading, joining, aggregating, filtering, and otherwise manipulating tabular data using a DataFrame style API in the style of [pandas](https://pandas.pydata.org).\n", "\n", "[Dask](https://www.dask.org/) is a flexible library for parallel computing in Python that makes scaling out your workflow smooth and simple. On the CPU, Dask uses Pandas to execute operations in parallel on DataFrame partitions.\n", "\n", - "[Dask cuDF](https://github.com/rapidsai/cudf/tree/main/python/dask_cudf) extends Dask where necessary to allow its DataFrame partitions to be processed using cuDF GPU DataFrames instead of Pandas DataFrames. For instance, when you call `dask_cudf.read_csv(...)`, your cluster's GPUs do the work of parsing the CSV file(s) by calling [`cudf.read_csv()`](https://docs.rapids.ai/api/cudf/stable/cudf/api_docs/api/cudf.read_csv/).\n", + "[Dask cuDF](https://github.com/NVIDIA/cudf/tree/main/python/dask_cudf) extends Dask where necessary to allow its DataFrame partitions to be processed using cuDF GPU DataFrames instead of Pandas DataFrames. For instance, when you call `dask_cudf.read_csv(...)`, your cluster's GPUs do the work of parsing the CSV file(s) by calling [`cudf.read_csv()`](https://docs.rapids.ai/api/cudf/stable/cudf/api_docs/api/cudf.read_csv/).\n", "\n", "\n", "
\n", - "Note: This notebook uses the explicit Dask cuDF API (dask_cudf) for clarity. However, we strongly recommend that you use Dask's configuration infrastructure to set the \"dataframe.backend\" option to \"cudf\", and work with the Dask DataFrame API directly. Please see the Dask cuDF documentation for more information.\n", + "Note: This notebook uses the explicit Dask cuDF API (dask_cudf) for clarity. However, we strongly recommend that you use Dask's configuration infrastructure to set the \"dataframe.backend\" option to \"cudf\", and work with the Dask DataFrame API directly. Please see the Dask cuDF documentation for more information.\n", "
\n", "\n", "\n", diff --git a/docs/cudf/source/cudf/developer_guide/contributing_guide.md b/docs/cudf/source/cudf/developer_guide/contributing_guide.md index 2be09707b982..3f904e180d09 100644 --- a/docs/cudf/source/cudf/developer_guide/contributing_guide.md +++ b/docs/cudf/source/cudf/developer_guide/contributing_guide.md @@ -16,7 +16,7 @@ These linting checks must all pass when submitting a pull request, and the Linter configurations are primarily stored in `pyproject.toml`, shared among other Python projects, and extended with cudf specific configurations in `python/cudf/pyproject.toml` For more information on how to use pre-commit hooks, see the code formatting section of the -[overall contributing guide](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md#using-pre-commit-hooks). +[overall contributing guide](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md#using-pre-commit-hooks). ## Deprecating and removing code diff --git a/docs/cudf/source/cudf/developer_guide/documentation.md b/docs/cudf/source/cudf/developer_guide/documentation.md index eb6266fd9e51..6d31dcab5cfd 100644 --- a/docs/cudf/source/cudf/developer_guide/documentation.md +++ b/docs/cudf/source/cudf/developer_guide/documentation.md @@ -130,7 +130,7 @@ While the documentation pages for the major user-facing classes like `DataFrame` For example, {py:class}`cudf.CategoricalIndex` only includes `codes` and `categories` on its page, not the entire set of `Index` functionality. To accommodate these requirements, we take the following approach: -1. The default `autosummary` template for classes is overridden with a [simpler template that does not generate method or attribute documentation](https://github.com/rapidsai/cudf/blob/main/docs/cudf/source/_templates/autosummary/class.rst). In other words, we disable `autosummary`'s generation of Methods and Attributes lists. +1. The default `autosummary` template for classes is overridden with a [simpler template that does not generate method or attribute documentation](https://github.com/NVIDIA/cudf/blob/main/docs/cudf/source/_templates/autosummary/class.rst). In other words, we disable `autosummary`'s generation of Methods and Attributes lists. 2. We rely on `numpydoc` entirely for the classes that need their entire APIs listed (`DataFrame`/`Series`/etc). `numpydoc` will automatically populate Methods and Attributes section if (and only if) they are not already defined in the class's docstring. 3. For classes that should only include a subset of APIs, we include those explicitly in the class's documentation. When those lists exist, `numpydoc` will not override them. If either the Methods or Attributes section should be empty, that section must still be included but should simply contain "None". For example, the class documentation for `CategoricalIndex` could include something like the following: @@ -187,8 +187,8 @@ The following are required to build the documentation: - A RAPIDS-compatible GPU. This is necessary because the documentation execute code. - A working copy of cudf in the same build environment. If you are only making changes to documentation we recommend following the - [Documentation contributions guide](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md#documentation-contributions) otherwise follow the - [build instructions](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md#setting-up-your-build-environment). + [Documentation contributions guide](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md#documentation-contributions) otherwise follow the + [build instructions](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md#setting-up-your-build-environment). - Sphinx, numpydoc, and MyST-NB. Assuming you follow the instructions in the previous step, these should automatically be installed into your environment. diff --git a/docs/cudf/source/cudf/developer_guide/index.md b/docs/cudf/source/cudf/developer_guide/index.md index 72ede1447c51..6499b773d5f3 100644 --- a/docs/cudf/source/cudf/developer_guide/index.md +++ b/docs/cudf/source/cudf/developer_guide/index.md @@ -2,7 +2,7 @@ ```{note} This document assumes familiarity with the -[overall contributing guide](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md). +[overall contributing guide](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md). ``` cuDF is a GPU-accelerated, [pandas-like](https://pandas.pydata.org/) DataFrame library. diff --git a/docs/cudf/source/cudf/developer_guide/testing.md b/docs/cudf/source/cudf/developer_guide/testing.md index 0cf675f7321b..5fdaf42e205a 100644 --- a/docs/cudf/source/cudf/developer_guide/testing.md +++ b/docs/cudf/source/cudf/developer_guide/testing.md @@ -4,7 +4,7 @@ Tests in cuDF are written using [`pytest`](https://docs.pytest.org/en/latest/). Test coverage is measured using [`coverage.py`](https://coverage.readthedocs.io/en/latest/), with the [`pytest-cov`](https://github.com/pytest-dev/pytest-cov) plugin. -Code coverage reports are uploaded to [Codecov](https://app.codecov.io/gh/rapidsai/cudf). +Code coverage reports are uploaded to [Codecov](https://app.codecov.io/gh/NVIDIA/cudf). Each PR also indicates whether it increases or decreases test coverage. ### Configuring pytest diff --git a/docs/cudf/source/cudf_pandas/benchmarks.md b/docs/cudf/source/cudf_pandas/benchmarks.md index e5e17ed5cbc9..ee8d503ed057 100644 --- a/docs/cudf/source/cudf_pandas/benchmarks.md +++ b/docs/cudf/source/cudf_pandas/benchmarks.md @@ -146,7 +146,7 @@ tpchgen-cli parquet -o "${DATA_PATH}" -s ${SCALE_FACTOR} `tpchgen-cli` generates Decimal and `datetime.date` columns. pandas cannot use these types in arithmetic, so convert them to float64 and timestamp before running the benchmark. This -conversion step may not be needed in the future (see [#21204](https://github.com/rapidsai/cudf/issues/21204)). +conversion step may not be needed in the future (see [#21204](https://github.com/NVIDIA/cudf/issues/21204)). ```python from pathlib import Path diff --git a/docs/cudf/source/cudf_pandas/developer_guide.md b/docs/cudf/source/cudf_pandas/developer_guide.md index 20ad8bab7218..a1a0b6bc1ba4 100644 --- a/docs/cudf/source/cudf_pandas/developer_guide.md +++ b/docs/cudf/source/cudf_pandas/developer_guide.md @@ -3,7 +3,7 @@ The use of the cuDF pandas accelerator mode (`cudf.pandas`) is explained [in the The purpose of this document is to explain how the fast-slow proxy mechanism works and document internal environment variables that can be used to debug `cudf.pandas` itself. ## fast-slow proxy mechanism -The core of `cudf.pandas` is implemented through proxy types defined in [`fast_slow_proxy.py`](https://github.com/rapidsai/cudf/blob/5f45803b2a68b49d330d94e2f701791a7590612a/python/cudf/cudf/pandas/fast_slow_proxy.py), which link a pair of "fast" and "slow" libraries. +The core of `cudf.pandas` is implemented through proxy types defined in [`fast_slow_proxy.py`](https://github.com/NVIDIA/cudf/blob/5f45803b2a68b49d330d94e2f701791a7590612a/python/cudf/cudf/pandas/fast_slow_proxy.py), which link a pair of "fast" and "slow" libraries. `cudf.pandas` works by wrapping each "slow" type and its corresponding "fast" type in a new proxy type, also known as a fast-slow proxy type. The purpose of these proxy types is so we can first attempt computations on the fast object, and then fall back to the slow object if the fast version fails. While the core wrapping functionality is generic, the current usage mainly involves providing a proxy pair using cuDF and Pandas. @@ -77,7 +77,7 @@ DataFrame = make_final_proxy_type( ``` ### The Fallback Mechanism -Proxied calls are implemented with fallback via [`_fast_slow_function_call`](https://github.com/rapidsai/cudf/blob/57aeeb78d85e169ac18b82f51d2b1cbd01b0608d/python/cudf/cudf/pandas/fast_slow_proxy.py#L869). This implements the mechanism by which we attempt operations the fast way (using cuDF) and then fall back to the slow way (using Pandas) on failure. +Proxied calls are implemented with fallback via [`_fast_slow_function_call`](https://github.com/NVIDIA/cudf/blob/57aeeb78d85e169ac18b82f51d2b1cbd01b0608d/python/cudf/cudf/pandas/fast_slow_proxy.py#L869). This implements the mechanism by which we attempt operations the fast way (using cuDF) and then fall back to the slow way (using Pandas) on failure. The function looks like this: ```python def _fast_slow_function_call(func: Callable, *args, **kwargs): diff --git a/docs/cudf/source/cudf_polars/developer_docs.md b/docs/cudf/source/cudf_polars/developer_docs.md index 9251a6bbb373..069ee07ce00b 100644 --- a/docs/cudf/source/cudf_polars/developer_docs.md +++ b/docs/cudf/source/cudf_polars/developer_docs.md @@ -8,7 +8,7 @@ You will need: preferred configuration. Or else, use [rustup](https://rust-lang.org/tools/install/) 2. A [cudf development - environment](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md#setting-up-your-build-environment). + environment](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md#setting-up-your-build-environment). The combined devcontainer works, or whatever your favourite approach is. ```{note} diff --git a/docs/cudf/source/cudf_polars/index.md b/docs/cudf/source/cudf_polars/index.md index 7ca59279bd39..9f641f33a024 100644 --- a/docs/cudf/source/cudf_polars/index.md +++ b/docs/cudf/source/cudf_polars/index.md @@ -94,7 +94,7 @@ On a single GPU, you can run TB-scale workloads with significant speedups compar ``` -For more information on the benchmarks being run, see the PDS queries in the [cuDF GitHub repository](https://github.com/rapidsai/cudf/tree/main/python/cudf_polars/cudf_polars/streaming/benchmarks). +For more information on the benchmarks being run, see the PDS queries in the [cuDF GitHub repository](https://github.com/NVIDIA/cudf/tree/main/python/cudf_polars/cudf_polars/streaming/benchmarks). ## Learn More diff --git a/docs/cudf/source/cudf_polars/io_plugins.md b/docs/cudf/source/cudf_polars/io_plugins.md index bfc0447f5e2d..54151f8847ac 100644 --- a/docs/cudf/source/cudf_polars/io_plugins.md +++ b/docs/cudf/source/cudf_polars/io_plugins.md @@ -60,7 +60,7 @@ is used, but that flag is not carried into the GPU plan, so cudf-polars validate unconditionally. A source that deliberately yields a dtype different from its declared schema (only valid with `validate_schema=False`) therefore cannot run on the GPU and must be collected with the default Polars CPU engine. This is tracked -in [cudf#22917](https://github.com/rapidsai/cudf/issues/22917). +in [cudf#22917](https://github.com/NVIDIA/cudf/issues/22917). ## Rank-Aware Sources @@ -120,7 +120,7 @@ unwrapped instance is recognized; wrapping it in anything else (for example, a `functools.partial`, closure, lambda, or decorator) hides the instance, in which case the source is treated as rank-unaware and executes on rank 0 only. -This limitation is tracked in [cudf#22917](https://github.com/rapidsai/cudf/issues/22917). +This limitation is tracked in [cudf#22917](https://github.com/NVIDIA/cudf/issues/22917). (io-plugins-sized-chunks)= @@ -166,7 +166,7 @@ cudf-polars raises `NotImplementedError`. As a result, `n_rows` is always `None` on GPU engines. However, IO sources intended to also work with the default Polars CPU engine must still handle `n_rows` correctly. -Distributed support for a global row limit is tracked in [cudf#22918](https://github.com/rapidsai/cudf/issues/22918). +Distributed support for a global row limit is tracked in [cudf#22918](https://github.com/NVIDIA/cudf/issues/22918). ## Threading diff --git a/docs/cudf/source/developer_guide/index.md b/docs/cudf/source/developer_guide/index.md index 775087ae35da..2760aae9fafc 100644 --- a/docs/cudf/source/developer_guide/index.md +++ b/docs/cudf/source/developer_guide/index.md @@ -2,7 +2,7 @@ This section contains developer documentation for the cuDF project. For general contribution guidelines, see the -[contributing guide](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md). +[contributing guide](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md). Each sub-project maintains its own developer documentation: diff --git a/docs/cudf/source/pylibcudf/developer_docs.md b/docs/cudf/source/pylibcudf/developer_docs.md index 544128e678a4..f7798a4513cf 100644 --- a/docs/cudf/source/pylibcudf/developer_docs.md +++ b/docs/cudf/source/pylibcudf/developer_docs.md @@ -153,7 +153,7 @@ Some guidelines on what should be tested: - Exception: In special cases where constructing suitable large tests is difficult in C++ (such as creating suitable input data for I/O testing), tests may be added to pylibcudf instead. - Nullable data should always be tested. - Expected exceptions should be tested. Tests should be written from the user's perspective in mind, and if the API is not currently throwing the appropriate exception it should be updated. - - Important note: If the exception should be produced by libcudf, the underlying libcudf API should be updated to throw the desired exception in C++. Such changes may require consultation with libcudf devs in nontrivial cases. [This issue](https://github.com/rapidsai/cudf/issues/12885) provides an overview and an indication of acceptable exception types that should cover most use cases. In rare cases a new C++ exception may need to be introduced in [`error.hpp`](https://github.com/rapidsai/cudf/blob/main/cpp/include/cudf/utilities/error.hpp). If so, this exception will also need to be mapped to a suitable Python exception in `exception_handler.pxd`. + - Important note: If the exception should be produced by libcudf, the underlying libcudf API should be updated to throw the desired exception in C++. Such changes may require consultation with libcudf devs in nontrivial cases. [This issue](https://github.com/NVIDIA/cudf/issues/12885) provides an overview and an indication of acceptable exception types that should cover most use cases. In rare cases a new C++ exception may need to be introduced in [`error.hpp`](https://github.com/NVIDIA/cudf/blob/main/cpp/include/cudf/utilities/error.hpp). If so, this exception will also need to be mapped to a suitable Python exception in `exception_handler.pxd`. Some guidelines on how best to use pytests. - By default, fixtures producing device data containers should be of module scope and treated as immutable by tests. Allocating data on the GPU is expensive and slows tests. Almost all pylibcudf operations are out of place operations, so module-scoped fixtures should not typically be problematic to work with. Session-scoped fixtures would also work, but they are harder to reason about since they live in a different module, and if they need to change for any reason they could affect an arbitrarily large number of tests. Module scope is a good balance. diff --git a/docs/dask_cudf/source/conf.py b/docs/dask_cudf/source/conf.py index 9920f687caec..58d2bfd7395c 100644 --- a/docs/dask_cudf/source/conf.py +++ b/docs/dask_cudf/source/conf.py @@ -60,7 +60,7 @@ html_theme_options = { "external_links": [], - "github_url": "https://github.com/rapidsai/cudf", + "github_url": "https://github.com/NVIDIA/cudf", "show_toc_level": 1, "navbar_align": "right", "navbar_center": "navbar-nav, version-switcher, navbar-external-links", diff --git a/java/README.md b/java/README.md index 7315bbc987c0..686aece2a505 100644 --- a/java/README.md +++ b/java/README.md @@ -56,7 +56,7 @@ the following cmake options to the libcudf build: -DCUDF_LARGE_STRINGS_DISABLED=ON -DCUDF_USE_ARROW_STATIC=ON -DCUDF_ENABLE_ARROW_S3=OFF ``` These options: -- Disable large string support, see https://github.com/rapidsai/cudf/issues/16215 +- Disable large string support, see https://github.com/NVIDIA/cudf/issues/16215 - Statically link Arrow to libcudf to remove Arrow as a runtime dependency. After building libcudf, the Java bindings can be built via Maven, e.g.: diff --git a/java/ci/README.md b/java/ci/README.md index 49ab3b6074ec..7df4f399ca59 100644 --- a/java/ci/README.md +++ b/java/ci/README.md @@ -119,7 +119,7 @@ nvidia-docker run -it cudf-build:12.9.1-devel-rocky8 bash You can download the cuDF repo in the docker container or you can mount it into the container. Here I choose to download again in the container. ```bash -git clone --recursive https://github.com/rapidsai/cudf.git -b main +git clone --recursive https://github.com/NVIDIA/cudf.git -b main ``` ```bash diff --git a/java/pom.xml b/java/pom.xml index 8521de5567e4..3c5e946a317b 100644 --- a/java/pom.xml +++ b/java/pom.xml @@ -28,10 +28,10 @@ - scm:git:https://github.com/rapidsai/cudf.git - scm:git:git@github.com:rapidsai/cudf.git + scm:git:https://github.com/NVIDIA/cudf.git + scm:git:git@github.com:NVIDIA/cudf.git HEAD - https://github.com/rapidsai/cudf + https://github.com/NVIDIA/cudf diff --git a/java/src/main/java/ai/rapids/cudf/ColumnWriterOptions.java b/java/src/main/java/ai/rapids/cudf/ColumnWriterOptions.java index b9387b6b0d3f..17e57ae538de 100644 --- a/java/src/main/java/ai/rapids/cudf/ColumnWriterOptions.java +++ b/java/src/main/java/ai/rapids/cudf/ColumnWriterOptions.java @@ -1,6 +1,6 @@ /* * - * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. + * SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. * SPDX-License-Identifier: Apache-2.0 * */ @@ -77,9 +77,9 @@ protected AbstractStructBuilder() { // This child is needed as the first child of a List column meta due to how cudf has been // implemented. Cudf drops the first child from the meta if a column is a LIST. This is done // this way due to some complications in the parquet reader. There was change to fix this here: - // https://github.com/rapidsai/cudf/pull/7461/commits/5ce33b40abb87cc7b76b5efeb0a3a0215f9ef6fb + // https://github.com/NVIDIA/cudf/pull/7461/commits/5ce33b40abb87cc7b76b5efeb0a3a0215f9ef6fb // but it was reverted later on here: - // https://github.com/rapidsai/cudf/pull/7461/commits/f248eb7265de995a95f998d46d897fb0ae47f53e + // https://github.com/NVIDIA/cudf/pull/7461/commits/f248eb7265de995a95f998d46d897fb0ae47f53e static ColumnWriterOptions DUMMY_CHILD = new ColumnWriterOptions("DUMMY"); public static abstract class NestedBuilder { diff --git a/java/src/main/native/src/ColumnViewJni.cpp b/java/src/main/native/src/ColumnViewJni.cpp index b0d81a4a5ebc..0706fa6bc246 100644 --- a/java/src/main/native/src/ColumnViewJni.cpp +++ b/java/src/main/native/src/ColumnViewJni.cpp @@ -1390,7 +1390,7 @@ Java_ai_rapids_cudf_ColumnView_castTo(JNIEnv* env, jclass, jlong handle, jint ty } else if (cudf::is_timestamp(n_data_type) && cudf::is_numeric(column->type())) { // This is a temporary workaround to allow Java to cast from integral types into a timestamp // without forcing an intermediate duration column to be manifested. Ultimately this style of - // "reinterpret" casting will be supported via https://github.com/rapidsai/cudf/pull/5358 + // "reinterpret" casting will be supported via https://github.com/NVIDIA/cudf/pull/5358 if (n_data_type.id() == cudf::type_id::TIMESTAMP_DAYS) { if (column->type().id() != cudf::type_id::INT32) { JNI_THROW_NEW(env, @@ -1413,7 +1413,7 @@ Java_ai_rapids_cudf_ColumnView_castTo(JNIEnv* env, jclass, jlong handle, jint ty } else if (cudf::is_timestamp(column->type()) && cudf::is_numeric(n_data_type)) { // This is a temporary workaround to allow Java to cast from timestamp types to integral types // without forcing an intermediate duration column to be manifested. Ultimately this style of - // "reinterpret" casting will be supported via https://github.com/rapidsai/cudf/pull/5358 + // "reinterpret" casting will be supported via https://github.com/NVIDIA/cudf/pull/5358 cudf::data_type duration_type = cudf::jni::timestamp_to_duration(column->type()); cudf::column_view duration_view = cudf::column_view( duration_type, column->size(), column->head(), column->null_mask(), column->null_count()); diff --git a/java/src/test/java/ai/rapids/cudf/ColumnVectorTest.java b/java/src/test/java/ai/rapids/cudf/ColumnVectorTest.java index 2606f24bcba9..d877c5c193e8 100644 --- a/java/src/test/java/ai/rapids/cudf/ColumnVectorTest.java +++ b/java/src/test/java/ai/rapids/cudf/ColumnVectorTest.java @@ -2888,7 +2888,7 @@ void testFlattenLists() { } } - // Edge cases: empty/null outer rows — inner column has 0 rows (rapidsai/cudf#22146). + // Edge cases: empty/null outer rows — inner column has 0 rows (NVIDIA/cudf#22146). // These exercise the child.size()==0 early-exit guard added to concatenate_list_elements. // flatten([]) — 1 row: valid but empty list-of-lists; child.size()==0 triggers guard. @@ -3580,7 +3580,7 @@ void testCastDecimal64ToString() { * We can not simply create decimal string via `String.valueOf`, because castStringToDecimal doesn't * support scientific notations so far. * - * issue for scientific notation: https://github.com/rapidsai/cudf/issues/7665 + * issue for scientific notation: https://github.com/NVIDIA/cudf/issues/7665 */ private static String dumpDecimal(Long unscaledValue, int scale) { if (unscaledValue == null) return null; diff --git a/python/cudf/cudf/core/column/column.py b/python/cudf/cudf/core/column/column.py index d6e57b96576c..643a94134497 100644 --- a/python/cudf/cudf/core/column/column.py +++ b/python/cudf/cudf/core/column/column.py @@ -1074,7 +1074,7 @@ def from_cuda_array_interface(cls, arbitrary: Any) -> ColumnBase: arbitrary = cp.ascontiguousarray(arbitrary) # TODO: Can remove once from_cuda_array_interface can handle masks - # https://github.com/rapidsai/cudf/issues/19122 + # https://github.com/NVIDIA/cudf/issues/19122 if (mask := cai.get("mask", None)) is not None: cai_copy = cai.copy() cai_copy.pop("mask") @@ -1173,7 +1173,7 @@ def to_pandas( pandas_array = pandas_nullable_dtype.__from_arrow__(pa_array) return pd.Index(pandas_array, copy=False) else: - # xref https://github.com/rapidsai/cudf/issues/21120 + # xref https://github.com/NVIDIA/cudf/issues/21120 # TODO: Revisit using pa_array.to_pandas() once pandas 3.0 is supported np_array = pa_array.to_numpy(zero_copy_only=False, writable=True) return pd.Index( @@ -2189,7 +2189,7 @@ def isin(self, values: Sequence | ColumnBase) -> ColumnBase: # nulls, these nulls should be replaced by whether or not the # haystack contains a null. # TODO: this is unnecessary if we resolve - # https://github.com/rapidsai/cudf/issues/14515 by + # https://github.com/NVIDIA/cudf/issues/14515 by # providing a mode in which cudf::contains does not mask # the result. result = result.fillna(rhs.null_count > 0) @@ -3925,7 +3925,7 @@ def as_column( ): # TODO: Need to re-visit this cast and fill_null # calls while addressing the following issue: - # https://github.com/rapidsai/cudf/issues/14149 + # https://github.com/NVIDIA/cudf/issues/14149 arbitrary = arbitrary.cast(pa.float64()) arbitrary = pc.fill_null(arbitrary, np.nan) if ( diff --git a/python/cudf/cudf/core/column/decimal.py b/python/cudf/cudf/core/column/decimal.py index 78578608c7a6..77ea32d0e186 100644 --- a/python/cudf/cudf/core/column/decimal.py +++ b/python/cudf/cudf/core/column/decimal.py @@ -162,7 +162,7 @@ def __pow__(self, other: ColumnBinaryOperand) -> ColumnBase: ) # Decimals in libcudf don't support truediv, see - # https://github.com/rapidsai/cudf/pull/7435 for explanation. + # https://github.com/NVIDIA/cudf/pull/7435 for explanation. def __truediv__(self, other: ColumnBinaryOperand) -> ColumnBase: return self._binaryop(other, "__div__") diff --git a/python/cudf/cudf/core/column/string.py b/python/cudf/cudf/core/column/string.py index eb5d34a53637..2ee9267284fb 100644 --- a/python/cudf/cudf/core/column/string.py +++ b/python/cudf/cudf/core/column/string.py @@ -353,7 +353,7 @@ def strptime( if target_unit != "s" and len(without_nat): # libcudf parses directly into int64 values of the # target unit and silently wraps on overflow - # (see https://github.com/rapidsai/cudf/issues/23247). + # (see https://github.com/NVIDIA/cudf/issues/23247). # Parse to seconds first (which cannot realistically # overflow) and reject values whose whole-second part # falls outside the target unit's range, like pandas diff --git a/python/cudf/cudf/core/column/timedelta.py b/python/cudf/cudf/core/column/timedelta.py index 03c97e117372..b6280ec2a82e 100644 --- a/python/cudf/cudf/core/column/timedelta.py +++ b/python/cudf/cudf/core/column/timedelta.py @@ -220,7 +220,7 @@ def _binaryop(self, other: ColumnBinaryOperand, op: str) -> ColumnBase: def total_seconds(self) -> ColumnBase: conversion = unit_to_nanoseconds_conversion[self.time_unit] / 1e9 # Typecast to decimal128 to avoid floating point precision issues - # https://github.com/rapidsai/cudf/issues/17664 + # https://github.com/NVIDIA/cudf/issues/17664 result = ( (self.astype(self._UNDERLYING_DTYPE) * conversion) .astype( diff --git a/python/cudf/cudf/core/column_accessor.py b/python/cudf/cudf/core/column_accessor.py index 9c480812c046..039831748a00 100644 --- a/python/cudf/cudf/core/column_accessor.py +++ b/python/cudf/cudf/core/column_accessor.py @@ -769,7 +769,7 @@ def _select_by_label_slice(self, key: slice) -> Self: start, stop = key.start, key.stop if len(self) == 0: - # https://github.com/rapidsai/cudf/issues/18376 + # https://github.com/NVIDIA/cudf/issues/18376 # Any slice is valid when we have no columns return self._from_columns_like_self([], verify=False) diff --git a/python/cudf/cudf/core/dataframe.py b/python/cudf/cudf/core/dataframe.py index 92a43a6adedc..a08ed067c5ee 100644 --- a/python/cudf/cudf/core/dataframe.py +++ b/python/cudf/cudf/core/dataframe.py @@ -330,7 +330,7 @@ def _getitem_tuple_arg(self, arg, per_level=False): else: raise RuntimeError( "Should have been handled by now. Please raise Github issue " - "at https://github.com/rapidsai/cudf/issues" + "at https://github.com/NVIDIA/cudf/issues" ) @_performance_tracking @@ -6903,7 +6903,7 @@ def isin(self, values): dog False False """ # TODO: propagate nulls through isin - # https://github.com/rapidsai/cudf/issues/7556 + # https://github.com/NVIDIA/cudf/issues/7556 def make_false_column_like_self(): return as_column(False, length=len(self), dtype=np.dtype("bool")) diff --git a/python/cudf/cudf/core/groupby/groupby.py b/python/cudf/cudf/core/groupby/groupby.py index 79aaa7a47dbd..6edb6efdd6d0 100644 --- a/python/cudf/cudf/core/groupby/groupby.py +++ b/python/cudf/cudf/core/groupby/groupby.py @@ -1329,7 +1329,7 @@ def agg(self, func=None, *args, engine=None, engine_kwargs=None, **kwargs): result_col = ColumnBase.create(plc_result, create_dtype) if agg == "cumcount": # pandas 0-indexes cumulative count, see - # https://github.com/rapidsai/cudf/issues/10237 + # https://github.com/NVIDIA/cudf/issues/10237 result_col = result_col - 1 if cast_dtype is not None: result_col = result_col.astype(cast_dtype) @@ -2303,7 +2303,7 @@ def _normalize_aggs( """ aggs_per_column: Iterable[AggType | Iterable[AggType]] # TODO: Remove isinstance condition when the legacy dask_cudf API is removed. - # See https://github.com/rapidsai/cudf/pull/16528#discussion_r1715482302 for information. + # See https://github.com/NVIDIA/cudf/pull/16528#discussion_r1715482302 for information. if aggs or isinstance(aggs, dict): if isinstance(aggs, dict): if any( diff --git a/python/cudf/cudf/core/reshape.py b/python/cudf/cudf/core/reshape.py index b521a34a0e87..7a958b69607a 100644 --- a/python/cudf/cudf/core/reshape.py +++ b/python/cudf/cudf/core/reshape.py @@ -461,7 +461,7 @@ def _concat_impl( empty_inner = join == "inner" and any(obj.empty for obj in objs) if len(objs) == 0: - # TODO: https://github.com/rapidsai/cudf/issues/16550 + # TODO: https://github.com/NVIDIA/cudf/issues/16550 return cudf.DataFrame() # Don't need to align indices of all `objs` since we @@ -560,7 +560,7 @@ def _concat_impl( if len(objs) == 0: # If objs is empty, that indicates all of # objs are empty dataframes. - # TODO: https://github.com/rapidsai/cudf/issues/16550 + # TODO: https://github.com/NVIDIA/cudf/issues/16550 return cudf.DataFrame() elif len(objs) == 1: obj = objs[0] @@ -1362,9 +1362,9 @@ def _unstack( and cudf.get_option("mode.pandas_compatible") ): # We currently produce columns in the wrong order vs pandas - # See https://github.com/rapidsai/cudf/issues/20446. + # See https://github.com/NVIDIA/cudf/issues/20446. # We should plan to remove once we rewrite pivot and unstack - # for better performance. See https://github.com/rapidsai/cudf/issues/20469 + # for better performance. See https://github.com/NVIDIA/cudf/issues/20469 raise NotImplementedError( "Unstacking multiple index levels is not yet pandas compatible" ) diff --git a/python/cudf/cudf/core/series.py b/python/cudf/cudf/core/series.py index 92db077ed10f..dcc6f0df164b 100644 --- a/python/cudf/cudf/core/series.py +++ b/python/cudf/cudf/core/series.py @@ -5158,7 +5158,7 @@ def strftime(self, date_format: str, *args, **kwargs) -> Series: ) # TODO: Remove following validations - # once https://github.com/rapidsai/cudf/issues/5991 + # once https://github.com/NVIDIA/cudf/issues/5991 # is implemented not_implemented_formats = { "%c", @@ -5170,7 +5170,7 @@ def strftime(self, date_format: str, *args, **kwargs) -> Series: raise NotImplementedError( f"{d_format} date-time format is not " f"supported yet, Please follow this issue " - f"https://github.com/rapidsai/cudf/issues/5991 " + f"https://github.com/NVIDIA/cudf/issues/5991 " f"for tracking purposes." ) if isinstance(self.series.dtype, pd.ArrowDtype): diff --git a/python/cudf/cudf/io/csv.py b/python/cudf/cudf/io/csv.py index 9ca4d17db1ab..44fc88c12217 100644 --- a/python/cudf/cudf/io/csv.py +++ b/python/cudf/cudf/io/csv.py @@ -422,7 +422,7 @@ def to_csv( # TODO: Need to typecast categorical columns to the underlying # categories dtype to write the actual data to csv. Remove this # workaround once following issue is fixed: - # https://github.com/rapidsai/cudf/issues/6661 + # https://github.com/NVIDIA/cudf/issues/6661 if any( isinstance(dtype, CategoricalDtype) for _, dtype in df._dtypes ) or isinstance(df.index, CategoricalIndex): @@ -570,7 +570,7 @@ def _validate_args( def _get_plc_data_type_from_dtype(dtype: DtypeObj) -> plc.DataType: # TODO: Remove this work-around Dictionary types # in libcudf are fully mapped to categorical columns: - # https://github.com/rapidsai/cudf/issues/3960 + # https://github.com/NVIDIA/cudf/issues/3960 if isinstance(dtype, CategoricalDtype): # TODO: should we do this generally in dtype_to_pylibcudf_type? dtype = dtype.categories.dtype diff --git a/python/cudf/cudf/pandas/_benchmarks/utils.py b/python/cudf/cudf/pandas/_benchmarks/utils.py index 0058e6d95aec..73348fa491f7 100644 --- a/python/cudf/cudf/pandas/_benchmarks/utils.py +++ b/python/cudf/cudf/pandas/_benchmarks/utils.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 """Utility functions/classes for running the PDS-H and PDS-DS benchmarks.""" @@ -207,7 +207,7 @@ def from_index(cls, index: int) -> GPUInfo: except pynvml.NVMLError_NotSupported: # Happens on systems without traditional GPU memory (e.g., Grace Hopper), # where nvmlDeviceGetMemoryInfo is not supported. - # See: https://github.com/rapidsai/cudf/issues/19427 + # See: https://github.com/NVIDIA/cudf/issues/19427 return cls( name=pynvml.nvmlDeviceGetName(handle), index=index, diff --git a/python/cudf/cudf/pandas/_wrappers/pandas.py b/python/cudf/cudf/pandas/_wrappers/pandas.py index be20ddf8ea1a..ba96216bfb4d 100644 --- a/python/cudf/cudf/pandas/_wrappers/pandas.py +++ b/python/cudf/cudf/pandas/_wrappers/pandas.py @@ -14,7 +14,7 @@ # cuGraph third party integration test, test_cugraph_from_pandas_adjacency, # fails without this pyarrow.dataset import # I suspect it relates to pyarrow's pandas-shim that gets imported -# with this module https://github.com/rapidsai/cudf/issues/14521#issue-2015198786 +# with this module https://github.com/NVIDIA/cudf/issues/14521#issue-2015198786 import pyarrow.dataset as ds # noqa: F401 from pandas._libs.tslibs import offsets as liboffsets from pandas._testing import at, getitem, iat, iloc, loc, setitem @@ -2642,7 +2642,7 @@ def wrapped_init(self, data=None, *args, **kwargs): # proxy object of the same type. # It is a common case in `cuml` and `xgboost`. # For perf impact see: - # https://github.com/rapidsai/cudf/pull/17878/files#r1936469215 + # https://github.com/NVIDIA/cudf/pull/17878/files#r1936469215 self.__dict__.update(data.__dict__) return original_init(self, data, *args, **kwargs) diff --git a/python/cudf/cudf/pandas/fast_slow_proxy.py b/python/cudf/cudf/pandas/fast_slow_proxy.py index 60fa07dc37b4..a6fe972ec9cb 100644 --- a/python/cudf/cudf/pandas/fast_slow_proxy.py +++ b/python/cudf/cudf/pandas/fast_slow_proxy.py @@ -1956,7 +1956,7 @@ def _replace_closurevars( g_closure = tuple(types.CellType(val) for val in g_nonlocals.values()) - # https://github.com/rapidsai/cudf/issues/15548 + # https://github.com/NVIDIA/cudf/issues/15548 new_g_globals = f.__globals__.copy() new_g_globals.update(g_globals) diff --git a/python/cudf/cudf/pandas/profiler.py b/python/cudf/cudf/pandas/profiler.py index e9971c7cf7f7..e48e73f3b60a 100644 --- a/python/cudf/cudf/pandas/profiler.py +++ b/python/cudf/cudf/pandas/profiler.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2023-2025, NVIDIA CORPORATION & AFFILIATES. All rights reserved. +# SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations @@ -289,10 +289,10 @@ def print_per_function_stats(self): call_to_action = ( "To request GPU support for any of these functions, " "please file a Github issue here: " - "[link=https://github.com/rapidsai/cudf/issues/new?assignees" + "[link=https://github.com/NVIDIA/cudf/issues/new?assignees" "=&labels=%3F+-+Needs+Triage%2C+feature+request&projects=" "&template=pandas_function_request.md&title=%5BFEA%5D]" - "https://github.com/rapidsai/cudf/issues/new/choose" + "https://github.com/NVIDIA/cudf/issues/new/choose" "[/link]." ) console.print( diff --git a/python/cudf/cudf/pandas/scripts/pandas-testing-plugin.py b/python/cudf/cudf/pandas/scripts/pandas-testing-plugin.py index a903039d8c95..7b5461773c82 100644 --- a/python/cudf/cudf/pandas/scripts/pandas-testing-plugin.py +++ b/python/cudf/cudf/pandas/scripts/pandas-testing-plugin.py @@ -226,8 +226,8 @@ def pytest_unconfigure(config): "tests/arrays/sparse/test_array.py::TestSparseArrayAnalytics::test_ufunc_args": "TODO: Add a reason for failure", "tests/arrays/sparse/test_array.py::test_array_interface": "TODO: Add a reason for failure", "tests/arrays/sparse/test_constructors.py::TestConstructors::test_constructor_copy": "TODO: Add a reason for failure", - "tests/arrays/string_/test_string_arrow.py::test_pickle_roundtrip[na_value0]": "https://github.com/rapidsai/cudf/issues/18659#issuecomment-3710985854", - "tests/arrays/string_/test_string_arrow.py::test_pickle_roundtrip[nan]": "https://github.com/rapidsai/cudf/issues/18659#issuecomment-3710985854", + "tests/arrays/string_/test_string_arrow.py::test_pickle_roundtrip[na_value0]": "https://github.com/NVIDIA/cudf/issues/18659#issuecomment-3710985854", + "tests/arrays/string_/test_string_arrow.py::test_pickle_roundtrip[nan]": "https://github.com/NVIDIA/cudf/issues/18659#issuecomment-3710985854", "tests/arrays/test_datetimelike.py::TestDatetimeArray::test_array_interface[B]": "TODO: Add a reason for failure", "tests/arrays/test_datetimelike.py::TestDatetimeArray::test_array_interface[D]": "TODO: Add a reason for failure", "tests/arrays/test_datetimelike.py::TestDatetimeArray::test_array_interface[ME]": "TODO: Add a reason for failure", @@ -797,7 +797,7 @@ def pytest_unconfigure(config): "tests/extension/test_arrow.py::test_interpolate_not_numeric[bool]": "TODO: Add a reason for failure", "tests/extension/test_arrow.py::test_mul_numpy_nullable_with_pyarrow_float": "AssertionError: Attributes of Series are different", "tests/extension/test_arrow.py::test_ops_with_nan_is_na[False]": "assert not np.True_", - "tests/extension/test_arrow.py::test_pickle_roundtrip[string]": "https://github.com/rapidsai/cudf/issues/18659#issuecomment-3710985854", + "tests/extension/test_arrow.py::test_pickle_roundtrip[string]": "https://github.com/NVIDIA/cudf/issues/18659#issuecomment-3710985854", "tests/extension/test_arrow.py::test_pow_missing_operand": "TODO: Add a reason for failure", "tests/extension/test_arrow.py::test_round": "TODO: Add a reason for failure", "tests/extension/test_arrow.py::test_setitem_float_nan_is_na[False]": "Failed: DID NOT RAISE ", @@ -4176,7 +4176,7 @@ def pytest_unconfigure(config): "tests/generic/test_finalize.py::test_categorical_accessor[method4]": "pandas xfails (reason: not implemented), but xpasses with cudf.pandas", "tests/generic/test_finalize.py::test_categorical_accessor[method7]": "pandas xfails (reason: not implemented), but xpasses with cudf.pandas", "tests/generic/test_finalize.py::test_categorical_accessor[method8]": "pandas xfails (reason: not implemented), but xpasses with cudf.pandas", - "tests/groupby/test_groupby.py::test_groupby_all_nan_groups_drop": "Flaky: intermittent AssertionError (https://github.com/rapidsai/cudf/issues/22681)", + "tests/groupby/test_groupby.py::test_groupby_all_nan_groups_drop": "Flaky: intermittent AssertionError (https://github.com/NVIDIA/cudf/issues/22681)", "tests/groupby/test_groupby.py::test_ops_not_as_index[idxmin]": "pandas xfails, but xpasses with cudf.pandas", "tests/groupby/test_groupby.py::test_single_element_list_grouping[selection2]": "Flaky xfails (TODO: Validate with pandas 3)", "tests/groupby/test_numba.py::TestEngine::test_as_index_false_unsupported[max-min_count]": "cuDF computes reductions on GPU without numba; pandas-specific limitation does not apply", diff --git a/python/cudf/cudf/testing/dataset_generator.py b/python/cudf/cudf/testing/dataset_generator.py index f1ba9eb73450..887c8444e2f9 100644 --- a/python/cudf/cudf/testing/dataset_generator.py +++ b/python/cudf/cudf/testing/dataset_generator.py @@ -573,7 +573,7 @@ def rand_dataframe(dtypes_meta, rows, seed=0, use_threads=True): else: raise TypeError(f"Unsupported dtype: {dtype}") # TODO: Add List column support once - # https://github.com/rapidsai/cudf/pull/6075 + # https://github.com/NVIDIA/cudf/pull/6075 # is merged. df = get_dataframe( diff --git a/python/cudf/cudf/tests/dask/test_serialize.py b/python/cudf/cudf/tests/dask/test_serialize.py index cba1efe2e7ca..8c84b758111c 100644 --- a/python/cudf/cudf/tests/dask/test_serialize.py +++ b/python/cudf/cudf/tests/dask/test_serialize.py @@ -384,7 +384,7 @@ def test_serialize_string_check_buffer_sizes(): def test_serialize_sliced_string(): - # https://github.com/rapidsai/cudf/issues/7735 + # https://github.com/NVIDIA/cudf/issues/7735 data = ["hi", "hello", None] pd_series = pd.Series(data, dtype=pd.StringDtype()) gd_series = cudf.Series(data, dtype=pd_series.dtype) diff --git a/python/cudf/cudf/tests/dataframe/indexing/test_iloc.py b/python/cudf/cudf/tests/dataframe/indexing/test_iloc.py index 57c24180a160..b96a06e3cc37 100644 --- a/python/cudf/cudf/tests/dataframe/indexing/test_iloc.py +++ b/python/cudf/cudf/tests/dataframe/indexing/test_iloc.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 @@ -368,7 +368,7 @@ def test_dataframe_iloc_inplace_update_shape_mismatch_RHS_df(): def test_iloc_single_row_with_nullable_column(): - # see https://github.com/rapidsai/cudf/issues/11349 + # see https://github.com/NVIDIA/cudf/issues/11349 pdf = pd.DataFrame({"a": [0, 1, 2, 3], "b": [0.1, 0.2, None, 0.4]}) df = cudf.from_pandas(pdf) @@ -392,7 +392,7 @@ def test_boolean_mask_columns_iloc_series(): def test_iloc_column_boolean_mask_issue_13265(): - # https://github.com/rapidsai/cudf/issues/13265 + # https://github.com/NVIDIA/cudf/issues/13265 df = pd.DataFrame(np.arange(4).reshape(2, 2)) cdf = cudf.from_pandas(df) expect = df.iloc[:, [True, True]] @@ -401,8 +401,8 @@ def test_iloc_column_boolean_mask_issue_13265(): def test_iloc_repeated_column_label_issue_13266(): - # https://github.com/rapidsai/cudf/issues/13266 - # https://github.com/rapidsai/cudf/issues/13273 + # https://github.com/NVIDIA/cudf/issues/13266 + # https://github.com/NVIDIA/cudf/issues/13273 df = pd.DataFrame(np.arange(4).reshape(2, 2)) cdf = cudf.from_pandas(df) @@ -419,7 +419,7 @@ def test_iloc_repeated_column_label_issue_13266(): ids=["row_ellipsis", "column_ellipsis"], ) def test_iloc_ellipsis_as_slice_issue_13267(indexer): - # https://github.com/rapidsai/cudf/issues/13267 + # https://github.com/NVIDIA/cudf/issues/13267 df = pd.DataFrame(np.arange(4).reshape(2, 2)) cdf = cudf.from_pandas(df) @@ -443,7 +443,7 @@ def test_iloc_ellipsis_as_slice_issue_13267(indexer): ], ) def test_iloc_multiindex_lookup_as_label_issue_13515(indexer): - # https://github.com/rapidsai/cudf/issues/13515 + # https://github.com/NVIDIA/cudf/issues/13515 df = pd.DataFrame( {"a": [1, 1, 3], "b": [2, 3, 4], "c": [1, 6, 7], "d": [1, 8, 9]} ).set_index(["a", "b"]) diff --git a/python/cudf/cudf/tests/dataframe/indexing/test_loc.py b/python/cudf/cudf/tests/dataframe/indexing/test_loc.py index 41e9661af1c8..8cc339fe1657 100644 --- a/python/cudf/cudf/tests/dataframe/indexing/test_loc.py +++ b/python/cudf/cudf/tests/dataframe/indexing/test_loc.py @@ -98,7 +98,7 @@ def test_multiindex_wildcard_selection_partial(): pytest.param( ["7", "8"], marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/11298" + reason="https://github.com/NVIDIA/cudf/issues/11298" ), ), ], @@ -114,7 +114,7 @@ def test_loc_setitem_string_11298(value): assert_eq(df, cdf) -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/11944") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/11944") def test_loc_setitem_list_11944(): df = pd.DataFrame( data={"a": ["yes", "no"], "b": [["l1", "l2"], ["c", "d"]]} @@ -125,7 +125,7 @@ def test_loc_setitem_list_11944(): assert_eq(df, cdf) -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/12504") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/12504") def test_loc_setitem_extend_empty_12504(): df = pd.DataFrame(columns=["a"]) cdf = cudf.from_pandas(df) @@ -167,7 +167,7 @@ def test_loc_setitem_list_arg_missing_raises(): pdf.loc[[1], "a"] = 1 -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/12801") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/12801") def test_loc_setitem_add_column_partial_12801(): df = pd.DataFrame({"a": [0, 1, 2]}) cdf = cudf.from_pandas(df) @@ -969,7 +969,7 @@ def test_dataframe_indexing_setitem_np_cp_array( class TestLocIndexWithOrder: - # https://github.com/rapidsai/cudf/issues/12833 + # https://github.com/NVIDIA/cudf/issues/12833 @pytest.fixture(params=["increasing", "decreasing", "neither"]) def order(self, request): return request.param @@ -1049,7 +1049,7 @@ def test_loc_index_notinindex_slice(self, df, order, dtype, take_order): def test_loc_single_row_from_slice(): - # see https://github.com/rapidsai/cudf/issues/11930 + # see https://github.com/NVIDIA/cudf/issues/11930 pdf = pd.DataFrame({"a": [10, 20, 30], "b": [1, 2, 3]}).set_index("a") df = cudf.from_pandas(pdf) assert_eq(pdf.loc[5:10], df.loc[5:10]) @@ -1085,7 +1085,7 @@ def test_boolean_mask_columns_wrong_length(indexer, mask): @pytest.mark.parametrize("index_type", ["single", "slice"]) def test_loc_timestamp_issue_8585(index_type): rng = np.random.default_rng(seed=0) - # https://github.com/rapidsai/cudf/issues/8585 + # https://github.com/NVIDIA/cudf/issues/8585 start = pd.Timestamp("2021-03-12 00:00") end = pd.Timestamp("2021-03-12 11:00") timestamps = pd.date_range(start, end, periods=12) @@ -1110,20 +1110,20 @@ def test_loc_timestamp_issue_8585(index_type): pytest.param( "slice", marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/8585" + reason="https://github.com/NVIDIA/cudf/issues/8585" ), ), pytest.param( "date_range", marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/8585" + reason="https://github.com/NVIDIA/cudf/issues/8585" ), ), ], ) def test_loc_multiindex_timestamp_issue_8585(index_type): rng = np.random.default_rng(seed=0) - # https://github.com/rapidsai/cudf/issues/8585 + # https://github.com/NVIDIA/cudf/issues/8585 start = pd.Timestamp("2021-03-12 00:00") end = pd.Timestamp("2021-03-12 03:00") timestamps = pd.date_range(start, end, periods=4) @@ -1153,7 +1153,7 @@ def test_loc_multiindex_timestamp_issue_8585(index_type): "indexer", [(..., 0), (0, ...)], ids=["row_ellipsis", "column_ellipsis"] ) def test_loc_ellipsis_as_slice_issue_13268(indexer): - # https://github.com/rapidsai/cudf/issues/13268 + # https://github.com/NVIDIA/cudf/issues/13268 df = pd.DataFrame(np.arange(4).reshape(2, 2)) cdf = cudf.from_pandas(df) @@ -1163,12 +1163,12 @@ def test_loc_ellipsis_as_slice_issue_13268(indexer): @pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/13269 " - "and https://github.com/rapidsai/cudf/issues/13273" + reason="https://github.com/NVIDIA/cudf/issues/13269 " + "and https://github.com/NVIDIA/cudf/issues/13273" ) def test_loc_repeated_column_label_issue_13269(): - # https://github.com/rapidsai/cudf/issues/13269 - # https://github.com/rapidsai/cudf/issues/13273 + # https://github.com/NVIDIA/cudf/issues/13269 + # https://github.com/NVIDIA/cudf/issues/13273 df = pd.DataFrame(np.arange(4).reshape(2, 2)) cdf = cudf.from_pandas(df) @@ -1178,7 +1178,7 @@ def test_loc_repeated_column_label_issue_13269(): def test_loc_column_boolean_mask_issue_13270(): - # https://github.com/rapidsai/cudf/issues/13270 + # https://github.com/NVIDIA/cudf/issues/13270 df = pd.DataFrame(np.arange(4).reshape(2, 2)) cdf = cudf.from_pandas(df) expect = df.loc[:, [True, True]] @@ -1187,7 +1187,7 @@ def test_loc_column_boolean_mask_issue_13270(): def test_loc_unsorted_index_slice_lookup_keyerror_issue_12833(): - # https://github.com/rapidsai/cudf/issues/12833 + # https://github.com/NVIDIA/cudf/issues/12833 df = pd.DataFrame({"a": [1, 2, 3]}, index=[7, 0, 4]) cdf = cudf.from_pandas(df) @@ -1201,7 +1201,7 @@ def test_loc_unsorted_index_slice_lookup_keyerror_issue_12833(): @pytest.mark.parametrize("index", [range(5), list(range(5))]) def test_loc_missing_label_keyerror_issue_13379(index): - # https://github.com/rapidsai/cudf/issues/13379 + # https://github.com/NVIDIA/cudf/issues/13379 df = pd.DataFrame({"a": index}, index=index) cdf = cudf.from_pandas(df) # Check that pandas don't change their mind @@ -1214,7 +1214,7 @@ def test_loc_missing_label_keyerror_issue_13379(index): @pytest.mark.parametrize("series", [True, False], ids=["Series", "DataFrame"]) def test_loc_repeated_label_ordering_issue_13658(series): - # https://github.com/rapidsai/cudf/issues/13658 + # https://github.com/NVIDIA/cudf/issues/13658 values = range(2048) index = [1 for _ in values] if series: diff --git a/python/cudf/cudf/tests/dataframe/indexing/test_setitem.py b/python/cudf/cudf/tests/dataframe/indexing/test_setitem.py index f7884cb6d710..71999ccb25fb 100644 --- a/python/cudf/cudf/tests/dataframe/indexing/test_setitem.py +++ b/python/cudf/cudf/tests/dataframe/indexing/test_setitem.py @@ -119,7 +119,7 @@ def test_listcol_setitem_retain_dtype(): df1["b"] = df1["c"] # Performing a copy to trigger a copy dtype which is obtained by accessing # `ListColumn.children` that would have been corrupted in previous call - # prior to this fix: https://github.com/rapidsai/cudf/pull/10151/ + # prior to this fix: https://github.com/NVIDIA/cudf/pull/10151/ df2 = df1.copy() assert df2["a"].dtype == df["a"].dtype diff --git a/python/cudf/cudf/tests/dataframe/methods/test_apply.py b/python/cudf/cudf/tests/dataframe/methods/test_apply.py index b6ae38580f16..447fec51123d 100644 --- a/python/cudf/cudf/tests/dataframe/methods/test_apply.py +++ b/python/cudf/cudf/tests/dataframe/methods/test_apply.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import decimal import math @@ -423,7 +423,7 @@ def func(row): constant == 1 and arithmetic_op in {operator.pow, operator.ipow} ), - reason="https://github.com/rapidsai/cudf/issues/7478", + reason="https://github.com/NVIDIA/cudf/issues/7478", ) ) run_masked_udf_test(func, gdf, check_dtype=False) @@ -444,7 +444,7 @@ def func(row): (gdf["data"] == 1).any() and arithmetic_op in {operator.pow, operator.ipow} ), - reason="https://github.com/rapidsai/cudf/issues/7478", + reason="https://github.com/NVIDIA/cudf/issues/7478", ) ) run_masked_udf_test(func, gdf, check_dtype=False) diff --git a/python/cudf/cudf/tests/dataframe/methods/test_assign.py b/python/cudf/cudf/tests/dataframe/methods/test_assign.py index dc5174790752..ed3b06dba0ba 100644 --- a/python/cudf/cudf/tests/dataframe/methods/test_assign.py +++ b/python/cudf/cudf/tests/dataframe/methods/test_assign.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import numpy as np @@ -67,7 +67,7 @@ def test_categorical_assignment(codes, categories): # assign categorical array # needed for dask_cudf support for including file name # as a categorical column - # see issue: https://github.com/rapidsai/cudf/issues/2269 + # see issue: https://github.com/NVIDIA/cudf/issues/2269 pd_df = pd.DataFrame({"a": np.ones(len(data))}) cd_df = cudf.from_pandas(pd_df) diff --git a/python/cudf/cudf/tests/dataframe/methods/test_astype.py b/python/cudf/cudf/tests/dataframe/methods/test_astype.py index f0f4c737780a..64759849745c 100644 --- a/python/cudf/cudf/tests/dataframe/methods/test_astype.py +++ b/python/cudf/cudf/tests/dataframe/methods/test_astype.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import numpy as np @@ -151,7 +151,7 @@ def test_df_astype_numeric_to_all( def test_df_astype_string_to_other(as_dtype): if "datetime64" in as_dtype: # change None to "NaT" after this issue is fixed: - # https://github.com/rapidsai/cudf/issues/5117 + # https://github.com/NVIDIA/cudf/issues/5117 data = ["2001-01-01", "2002-02-02", "2000-01-05", None] elif as_dtype == "int32": data = [1, 2, 3] diff --git a/python/cudf/cudf/tests/dataframe/methods/test_reductions.py b/python/cudf/cudf/tests/dataframe/methods/test_reductions.py index 92e769dfd362..4a64088479b2 100644 --- a/python/cudf/cudf/tests/dataframe/methods/test_reductions.py +++ b/python/cudf/cudf/tests/dataframe/methods/test_reductions.py @@ -339,7 +339,7 @@ def test_dataframe_reductions(request, data, axis, func, skipna): }: request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/20628" + reason="https://github.com/NVIDIA/cudf/issues/20628" ) ) diff --git a/python/cudf/cudf/tests/dataframe/methods/test_replace.py b/python/cudf/cudf/tests/dataframe/methods/test_replace.py index 17bcc50a4809..b92ddd4ce8b3 100644 --- a/python/cudf/cudf/tests/dataframe/methods/test_replace.py +++ b/python/cudf/cudf/tests/dataframe/methods/test_replace.py @@ -175,7 +175,7 @@ def test_replace_multiple_rows(datadir): # pandas 3 reads nullable parquet columns as Float64; cast to float64 to # match cudf's representation before comparing. - # TODO: Remove this cast after https://github.com/rapidsai/cudf/issues/22018 is resolved. + # TODO: Remove this cast after https://github.com/NVIDIA/cudf/issues/22018 is resolved. pdf = pdf.astype( {c: "float64" for c, d in pdf.dtypes.items() if d == pd.Float64Dtype()} ) diff --git a/python/cudf/cudf/tests/dataframe/test_binops.py b/python/cudf/cudf/tests/dataframe/test_binops.py index e8a3898c0e94..01f9acb314af 100644 --- a/python/cudf/cudf/tests/dataframe/test_binops.py +++ b/python/cudf/cudf/tests/dataframe/test_binops.py @@ -420,7 +420,7 @@ def test_df_sr_binop(psr, colnames, binary_op): operator.mod, operator.pow, # comparison ops will temporarily XFAIL - # see PR https://github.com/rapidsai/cudf/pull/7491 + # see PR https://github.com/NVIDIA/cudf/pull/7491 pytest.param(operator.eq, marks=pytest.mark.xfail), pytest.param(operator.lt, marks=pytest.mark.xfail), pytest.param(operator.le, marks=pytest.mark.xfail), diff --git a/python/cudf/cudf/tests/general_functions/test_to_datetime.py b/python/cudf/cudf/tests/general_functions/test_to_datetime.py index 8a63323c186b..88ec2d204418 100644 --- a/python/cudf/cudf/tests/general_functions/test_to_datetime.py +++ b/python/cudf/cudf/tests/general_functions/test_to_datetime.py @@ -219,7 +219,7 @@ def test_to_datetime_units(data, unit): pd.Series(["1", "2", "1"]), "%m", marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/6109" + reason="https://github.com/NVIDIA/cudf/issues/6109" "https://github.com/pandas-dev/pandas/issues/35934" ), ), @@ -227,7 +227,7 @@ def test_to_datetime_units(data, unit): pd.Series(["14", "20", "10"]), "%d", marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/6109" + reason="https://github.com/NVIDIA/cudf/issues/6109" "https://github.com/pandas-dev/pandas/issues/35934" ), ), diff --git a/python/cudf/cudf/tests/groupby/test_agg.py b/python/cudf/cudf/tests/groupby/test_agg.py index 8e2a5ccfaf7f..79215962c1da 100644 --- a/python/cudf/cudf/tests/groupby/test_agg.py +++ b/python/cudf/cudf/tests/groupby/test_agg.py @@ -222,7 +222,7 @@ def test_groupby_agg_decimal(groupby_reduction_methods, request): # The unique is necessary because otherwise if there are duplicates idxmin # and idxmax may return different results than pandas (see - # https://github.com/rapidsai/cudf/issues/7756). This is not relevant to + # https://github.com/NVIDIA/cudf/issues/7756). This is not relevant to # the current version of the test, because idxmin and idxmax simply don't # work with pandas Series composed of Decimal objects (see # https://github.com/pandas-dev/pandas/issues/40685). However, if that is diff --git a/python/cudf/cudf/tests/groupby/test_apply.py b/python/cudf/cudf/tests/groupby/test_apply.py index 62dc09e3c2d0..80a1dc7690f5 100644 --- a/python/cudf/cudf/tests/groupby/test_apply.py +++ b/python/cudf/cudf/tests/groupby/test_apply.py @@ -229,7 +229,7 @@ def test_groupby_apply_jit_unary_reductions( and dtype.kind == "f" ) ), - reason=("https://github.com/rapidsai/cudf/issues/14860"), + reason=("https://github.com/NVIDIA/cudf/issues/14860"), ) ) dataset = groupby_jit_datasets[dataset].copy(deep=True) @@ -303,7 +303,7 @@ def test_groupby_apply_jit_reductions_special_vals( pytest.param( np.nan, marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/13832" + reason="https://github.com/NVIDIA/cudf/issues/13832" ), ), np.inf, @@ -367,7 +367,7 @@ def func(group): if np.dtype(dtype).kind == "f": # Correlation of floating types is not yet supported: - # https://github.com/rapidsai/cudf/issues/13839 + # https://github.com/NVIDIA/cudf/issues/13839 m = ( f"Series.corr\\(Series\\) is not " f"supported for \\({dtype}, {dtype}\\)" @@ -492,7 +492,7 @@ def test_groupby_apply_jit_args(func, args, groupby_jit_data_small): def test_groupby_apply_jit_block_divergence(): - # https://github.com/rapidsai/cudf/issues/12686 + # https://github.com/NVIDIA/cudf/issues/12686 df = cudf.DataFrame( { "a": [0, 0, 0, 1, 1, 1], diff --git a/python/cudf/cudf/tests/groupby/test_attributes.py b/python/cudf/cudf/tests/groupby/test_attributes.py index 773d26c7d900..e1953a652205 100644 --- a/python/cudf/cudf/tests/groupby/test_attributes.py +++ b/python/cudf/cudf/tests/groupby/test_attributes.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import numpy as np import pandas as pd @@ -10,7 +10,7 @@ def test_groups(): - # https://github.com/rapidsai/cudf/issues/14955 + # https://github.com/NVIDIA/cudf/issues/14955 df = cudf.DataFrame({"a": [1, 2] * 2}, index=[0] * 4) agg = df.groupby("a") pagg = df.to_pandas().groupby("a") diff --git a/python/cudf/cudf/tests/groupby/test_cov_corr.py b/python/cudf/cudf/tests/groupby/test_cov_corr.py index 1af91f81b57a..4d8afef82743 100644 --- a/python/cudf/cudf/tests/groupby/test_cov_corr.py +++ b/python/cudf/cudf/tests/groupby/test_cov_corr.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import numpy as np @@ -140,7 +140,7 @@ def test_groupby_cov_positive_semidefinite_matrix(): # rows containing missing values. # Note: cuDF currently matches pandas behavior in that the covariance # matrices are not guaranteed PSD (positive semi definite). - # https://github.com/rapidsai/cudf/pull/9889#discussion_r794158358 + # https://github.com/NVIDIA/cudf/pull/9889#discussion_r794158358 gdf = cudf.DataFrame( [[1, 2], [None, 4], [5, None], [7, 8]], columns=["v0", "v1"] ) diff --git a/python/cudf/cudf/tests/groupby/test_shift.py b/python/cudf/cudf/tests/groupby/test_shift.py index 1dd5108ea685..50101b21e024 100644 --- a/python/cudf/cudf/tests/groupby/test_shift.py +++ b/python/cudf/cudf/tests/groupby/test_shift.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 @@ -37,13 +37,13 @@ def test_groupby_shift_row(shift_perc, direction, fill_value): pytest.param( 0, marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/10608" + reason="https://github.com/NVIDIA/cudf/issues/10608" ), ), pytest.param( 42, marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/10608" + reason="https://github.com/NVIDIA/cudf/issues/10608" ), ), ], diff --git a/python/cudf/cudf/tests/indexes/timedeltaindex/test_binops.py b/python/cudf/cudf/tests/indexes/timedeltaindex/test_binops.py index fc25581c8fe7..640d25560ed5 100644 --- a/python/cudf/cudf/tests/indexes/timedeltaindex/test_binops.py +++ b/python/cudf/cudf/tests/indexes/timedeltaindex/test_binops.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2020-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import datetime @@ -190,7 +190,7 @@ def test_timedelta_index_ops_with_scalars( and 0 in ptdi.astype("int") and np.timedelta64(other_scalars).item() is not None ), - reason="Related to https://github.com/rapidsai/cudf/issues/5938", + reason="Related to https://github.com/NVIDIA/cudf/issues/5938", ) ) assert_eq(expected, actual) diff --git a/python/cudf/cudf/tests/input_output/test_avro.py b/python/cudf/cudf/tests/input_output/test_avro.py index 9583f56bc47d..05766346df00 100644 --- a/python/cudf/cudf/tests/input_output/test_avro.py +++ b/python/cudf/cudf/tests/input_output/test_avro.py @@ -228,7 +228,7 @@ def test_avro_decompression(set_decomp_env_vars, rows, codec): # N.B. rand_dataframe() is brutally slow for some reason. Switching to # np.random() speeds things up by a factor of 10. - # See also: https://github.com/rapidsai/cudf/issues/13128 + # See also: https://github.com/NVIDIA/cudf/issues/13128 df = rand_dataframe( [ {"dtype": "int32", "null_frequency": 0, "cardinality": 1000}, @@ -361,7 +361,7 @@ def test_alltypes_plain_avro(): # added simply to verify we fixed the problematic code path that was # causing CUDA kernel crashes. # - # See https://github.com/rapidsai/cudf/pull/12788#issuecomment-1468822875 + # See https://github.com/NVIDIA/cudf/pull/12788#issuecomment-1468822875 # for more information. relpath = "../../../../java/src/test/resources/alltypes_plain.avro" path = pathlib.Path(__file__).parent.parent.joinpath(relpath).resolve() @@ -594,7 +594,7 @@ def test_avro_reader_multiblock( # We don't use rand_dataframe() here, because it increases the # execution time of each test by a factor of 10 or more (it appears # to use a very costly approach to generating random data). - # See also: https://github.com/rapidsai/cudf/issues/13128 + # See also: https://github.com/NVIDIA/cudf/issues/13128 values = rng.random(total_rows).astype(dtype) bytes_per_row = values.dtype.itemsize diff --git a/python/cudf/cudf/tests/input_output/test_csv.py b/python/cudf/cudf/tests/input_output/test_csv.py index 2bf096c3d9c0..d8a2de57ee13 100644 --- a/python/cudf/cudf/tests/input_output/test_csv.py +++ b/python/cudf/cudf/tests/input_output/test_csv.py @@ -128,7 +128,7 @@ def test_csv_reader_numeric_data(numeric_types_as_str, tmp_path): @pytest.mark.skip( - reason="Disabled until https://github.com/rapidsai/cudf/pull/22094 is fixed" + reason="Disabled until https://github.com/NVIDIA/cudf/pull/22094 is fixed" ) @pytest.mark.parametrize("parse_dates", [["date2"], [0], ["date1", 1, "bad"]]) def test_csv_reader_datetime(parse_dates): @@ -1862,7 +1862,7 @@ def test_csv_write_empty_dataframe(idx, index): } ), marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/16533, np.nan/None coerced to NA since pandas 3" + reason="https://github.com/NVIDIA/cudf/issues/16533, np.nan/None coerced to NA since pandas 3" ), ), pd.DataFrame({"": [1, None, 3, 4]}), @@ -2042,7 +2042,7 @@ def test_csv_sep_error(): def test_to_csv_encoding_error(): # TODO: Remove this test once following - # issue is fixed: https://github.com/rapidsai/cudf/issues/2957 + # issue is fixed: https://github.com/NVIDIA/cudf/issues/2957 df = cudf.DataFrame({"a": ["你好", "test"]}) encoding = "utf-8-sig" error_message = ( diff --git a/python/cudf/cudf/tests/input_output/test_hdfs.py b/python/cudf/cudf/tests/input_output/test_hdfs.py index 2630d8164ebf..e08ed48b7c2f 100644 --- a/python/cudf/cudf/tests/input_output/test_hdfs.py +++ b/python/cudf/cudf/tests/input_output/test_hdfs.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2020-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from io import BytesIO @@ -13,7 +13,7 @@ from cudf.testing import assert_eq pytest.skip( - reason="https://github.com/rapidsai/cudf/issues/19633", + reason="https://github.com/NVIDIA/cudf/issues/19633", allow_module_level=True, ) diff --git a/python/cudf/cudf/tests/input_output/test_orc.py b/python/cudf/cudf/tests/input_output/test_orc.py index 1622e2dcb33f..a29d088fe96d 100644 --- a/python/cudf/cudf/tests/input_output/test_orc.py +++ b/python/cudf/cudf/tests/input_output/test_orc.py @@ -31,7 +31,7 @@ # Removal of these deprecated features is no longer imminent. They will not be # removed until a suitable alternative has been implemented. As a result, we # also do not want to stop testing them yet. -# https://github.com/rapidsai/cudf/issues/11519 +# https://github.com/NVIDIA/cudf/issues/11519 pytestmark = pytest.mark.filterwarnings( "ignore:(num_rows|skiprows) is deprecated and will be removed." ) @@ -1726,7 +1726,7 @@ def test_orc_columns_and_index_param(index_obj, index, columns): ), ], ) -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/12026") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/12026") def test_orc_columns_and_index_param_read_index(index_obj, index, columns): run_orc_columns_and_index_param(index_obj, index, columns) @@ -1846,7 +1846,7 @@ def test_orc_writer_negative_timestamp(): @pytest.mark.skip( - reason="Bug specific to rockylinux8: https://github.com/rapidsai/cudf/issues/15802", + reason="Bug specific to rockylinux8: https://github.com/NVIDIA/cudf/issues/15802", ) def test_orc_reader_apache_negative_timestamp(datadir): path = datadir / "TestOrcFile.apache_timestamp.orc" @@ -1918,7 +1918,7 @@ def test_reader_row_index_order(data): # Test the corner case where empty blocks are compressed # Decompressed data size is zero, even though compressed data size is non-zero -# For more information see https://github.com/rapidsai/cudf/issues/13608 +# For more information see https://github.com/NVIDIA/cudf/issues/13608 def test_orc_reader_empty_decomp_data(datadir): path = datadir / "TestOrcFile.Spark.EmptyDecompData.orc" @@ -2017,7 +2017,7 @@ def test_orc_reader_desynced_timestamp(datadir, inputfile): # is progressed faster than the SECONDARY stream (nanosecond) at the start of a row # group. In this case, the "run cache manager" in the decoder kernel is used to # orchestrate the dual-stream processing. - # For more information, see https://github.com/rapidsai/cudf/issues/17155. + # For more information, see https://github.com/NVIDIA/cudf/issues/17155. path = datadir / inputfile diff --git a/python/cudf/cudf/tests/input_output/test_parquet.py b/python/cudf/cudf/tests/input_output/test_parquet.py index bc7af5360010..bcee9a1023e4 100644 --- a/python/cudf/cudf/tests/input_output/test_parquet.py +++ b/python/cudf/cudf/tests/input_output/test_parquet.py @@ -1890,7 +1890,7 @@ def test_parquet_writer_bytes_io(simple_gdf, store_schema): ) def test_parquet_writer_row_group_size(tmp_path, row_group_size_kwargs): # Check that row_group_size options are exposed in Python - # See https://github.com/rapidsai/cudf/issues/10978 + # See https://github.com/NVIDIA/cudf/issues/10978 size = 20000 gdf = cudf.DataFrame({"a": range(size), "b": [1] * size}) @@ -3607,7 +3607,7 @@ def test_parquet_write_lz4(): def test_parquet_reader_zstd_huff_tables(datadir): # Ensure that this zstd-compressed file does not overrun buffers. The # problem was fixed in nvcomp 3.0.6. - # See https://github.com/rapidsai/cudf/issues/15096 + # See https://github.com/NVIDIA/cudf/issues/15096 fname = datadir / "zstd_huff_tables_bug.parquet" expected = pa.parquet.read_table(fname).to_pandas() @@ -4687,7 +4687,7 @@ def test_parquet_reader_mismatched_nullability_structs(tmp_path): @pytest.mark.skipif( pa.__version__ == "19.0.0", - reason="https://github.com/apache/arrow/issues/45283, https://github.com/rapidsai/cudf/issues/17806", + reason="https://github.com/apache/arrow/issues/45283, https://github.com/NVIDIA/cudf/issues/17806", ) @pytest.mark.parametrize( "stats_fname,bloom_filter_fname", diff --git a/python/cudf/cudf/tests/private_objects/test_column.py b/python/cudf/cudf/tests/private_objects/test_column.py index 387aa45d0dfb..0d88348880db 100644 --- a/python/cudf/cudf/tests/private_objects/test_column.py +++ b/python/cudf/cudf/tests/private_objects/test_column.py @@ -229,7 +229,7 @@ def test_column_series_cuda_array_dtype(data, float_types_as_str): def test_column_zero_length_slice(): - # see https://github.com/rapidsai/cudf/pull/4777 + # see https://github.com/NVIDIA/cudf/pull/4777 x = cudf.DataFrame({"a": [1]}) the_column = x[1:]["a"]._column diff --git a/python/cudf/cudf/tests/private_objects/test_nrt_stats.py b/python/cudf/cudf/tests/private_objects/test_nrt_stats.py index 68903dd0930d..c3671bb9f04c 100644 --- a/python/cudf/cudf/tests/private_objects/test_nrt_stats.py +++ b/python/cudf/cudf/tests/private_objects/test_nrt_stats.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import numpy as np import pytest @@ -17,7 +17,7 @@ from cudf.utils._numba import _CUDFNumbaConfig -@pytest.mark.skip(reason="https://github.com/rapidsai/cudf/issues/19880") +@pytest.mark.skip(reason="https://github.com/NVIDIA/cudf/issues/19880") def test_string_udf_basic(monkeypatch): monkeypatch.setattr(config, "CUDA_NRT_STATS", True) @@ -38,7 +38,7 @@ def double(st): assert stats.alloc - stats.free == 0 -@pytest.mark.skip(reason="https://github.com/rapidsai/cudf/issues/19880") +@pytest.mark.skip(reason="https://github.com/NVIDIA/cudf/issues/19880") def test_string_udf_conditional_allocations(monkeypatch): monkeypatch.setattr(config, "CUDA_NRT_STATS", True) @@ -59,7 +59,7 @@ def double(st): assert after_stats.alloc - before_stats.free == 1 -@pytest.mark.skip(reason="https://github.com/rapidsai/cudf/issues/19880") +@pytest.mark.skip(reason="https://github.com/NVIDIA/cudf/issues/19880") def test_string_udf_free_kernel(monkeypatch): monkeypatch.setattr(config, "CUDA_NRT_STATS", True) diff --git a/python/cudf/cudf/tests/reshape/test_concat.py b/python/cudf/cudf/tests/reshape/test_concat.py index 606f23a5bba8..039fe8e87b4c 100644 --- a/python/cudf/cudf/tests/reshape/test_concat.py +++ b/python/cudf/cudf/tests/reshape/test_concat.py @@ -1780,7 +1780,7 @@ def test_concat_list_column( def test_concat_categorical_ordering(): - # https://github.com/rapidsai/cudf/issues/11486 + # https://github.com/NVIDIA/cudf/issues/11486 sr = pd.Series( ["a", "b", "c", "d", "e", "a", "b", "c", "d", "e"], dtype="category" ) diff --git a/python/cudf/cudf/tests/series/accessors/test_str.py b/python/cudf/cudf/tests/series/accessors/test_str.py index 1281eea64967..1d3d1daada1f 100644 --- a/python/cudf/cudf/tests/series/accessors/test_str.py +++ b/python/cudf/cudf/tests/series/accessors/test_str.py @@ -2619,7 +2619,7 @@ def _cat_convert_seq_to_cudf(others): np.array(["f", "g", "h", "i", "j"]), ], marks=pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/5862" + reason="https://github.com/NVIDIA/cudf/issues/5862" ), ), pytest.param( @@ -2683,7 +2683,7 @@ def test_string_index_duplicate_str_cat(data, others, sep, na_rep, name): # TODO: Remove got.sort_values call once we have `join` param support # in `.str.cat` - # https://github.com/rapidsai/cudf/issues/5862 + # https://github.com/NVIDIA/cudf/issues/5862 assert_eq( expect.sort_values() if not isinstance(expect, str) else expect, @@ -2810,7 +2810,7 @@ def test_string_cat(ps_gs, others, sep, na_rep, index, request): request.applymarker( pytest.mark.xfail( is_any_others_series_with_string_index, - reason="https://github.com/rapidsai/cudf/issues/21123", + reason="https://github.com/NVIDIA/cudf/issues/21123", ) ) pd_others = others diff --git a/python/cudf/cudf/tests/series/indexing/test_iloc.py b/python/cudf/cudf/tests/series/indexing/test_iloc.py index 5d41ea31d47d..be904b14cc04 100644 --- a/python/cudf/cudf/tests/series/indexing/test_iloc.py +++ b/python/cudf/cudf/tests/series/indexing/test_iloc.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import cupy as cp @@ -290,7 +290,7 @@ def test_iloc_decimal(): @pytest.mark.parametrize("indexer", [[1], [0, 2]]) def test_iloc_integer_categorical_issue_13013(indexer): - # https://github.com/rapidsai/cudf/issues/13013 + # https://github.com/NVIDIA/cudf/issues/13013 s = pd.Series([0, 1, 2]) index = pd.Categorical(indexer) expect = s.iloc[index] @@ -300,7 +300,7 @@ def test_iloc_integer_categorical_issue_13013(indexer): def test_iloc_incorrect_boolean_mask_length_issue_13015(): - # https://github.com/rapidsai/cudf/issues/13015 + # https://github.com/NVIDIA/cudf/issues/13015 s = pd.Series([0, 1, 2]) with pytest.raises(IndexError): s.iloc[[True, False]] diff --git a/python/cudf/cudf/tests/series/indexing/test_loc.py b/python/cudf/cudf/tests/series/indexing/test_loc.py index 04f6f4cedc4f..d934118741e1 100644 --- a/python/cudf/cudf/tests/series/indexing/test_loc.py +++ b/python/cudf/cudf/tests/series/indexing/test_loc.py @@ -10,7 +10,7 @@ from cudf.testing import assert_eq -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/13031") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/13031") @pytest.mark.parametrize("other_index", [["1", "3", "2"], [1, 2, 3]]) def test_loc_setitem_series_index_alignment_13031(other_index): s = pd.Series([1, 2, 3], index=["1", "2", "3"]) @@ -352,7 +352,7 @@ def test_boolean_mask_wrong_length(indexer, mask): def test_loc_repeated_index_label_issue_8693(): - # https://github.com/rapidsai/cudf/issues/8693 + # https://github.com/NVIDIA/cudf/issues/8693 s = pd.Series([1, 2, 3, 4], index=[0, 1, 1, 2]) cs = cudf.from_pandas(s) expect = s.loc[1] @@ -382,7 +382,7 @@ def test_series_iloc_float_int(arg): @pytest.mark.parametrize("indexer", [[1], [0, 2]]) def test_loc_integer_categorical_issue_13014(indexer): - # https://github.com/rapidsai/cudf/issues/13014 + # https://github.com/NVIDIA/cudf/issues/13014 s = pd.Series([0, 1, 2]) index = pd.Categorical(indexer) expect = s.loc[index] @@ -396,7 +396,7 @@ def test_loc_integer_categorical_issue_13014(indexer): def test_loc_categorical_ordering_mismatch_issue_13652( index_is_ordered, label_is_ordered ): - # https://github.com/rapidsai/cudf/issues/13652 + # https://github.com/NVIDIA/cudf/issues/13652 s = cudf.Series( [0, 2, 8, 4, 2], index=cudf.CategoricalIndex( @@ -414,7 +414,7 @@ def test_loc_categorical_ordering_mismatch_issue_13652( def test_loc_categorical_no_integer_fallback_issue_13653(): - # https://github.com/rapidsai/cudf/issues/13653 + # https://github.com/NVIDIA/cudf/issues/13653 s = cudf.Series( [1, 2], index=cudf.CategoricalIndex([3, 4], categories=[3, 4]) ) diff --git a/python/cudf/cudf/tests/series/indexing/test_setitem.py b/python/cudf/cudf/tests/series/indexing/test_setitem.py index a2278082860f..9e2215cf54d4 100644 --- a/python/cudf/cudf/tests/series/indexing/test_setitem.py +++ b/python/cudf/cudf/tests/series/indexing/test_setitem.py @@ -223,7 +223,7 @@ def test_series_set_item_range_index(ps, arg): assert_eq(psr, gsr, check_index_type=True) -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/7448") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/7448") def test_iloc_setitem_7448(): index = pd.MultiIndex.from_product([(1, 2), (3, 4)]) expect = cudf.Series([1, 2, 3, 4], index=index) diff --git a/python/cudf/cudf/tests/series/methods/test_apply.py b/python/cudf/cudf/tests/series/methods/test_apply.py index 47960032df6f..bb9076e531aa 100644 --- a/python/cudf/cudf/tests/series/methods/test_apply.py +++ b/python/cudf/cudf/tests/series/methods/test_apply.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import operator @@ -234,7 +234,7 @@ def func(x): constant is cudf.NA and arithmetic_op in {operator.pow, operator.ipow} ), - reason="https://github.com/rapidsai/cudf/issues/7478", + reason="https://github.com/NVIDIA/cudf/issues/7478", ) ) run_masked_udf_series(func, data, check_dtype=False) @@ -258,7 +258,7 @@ def func(x): constant in {1} and arithmetic_op in {operator.pow, operator.ipow} ), - reason="https://github.com/rapidsai/cudf/issues/7478", + reason="https://github.com/NVIDIA/cudf/issues/7478", ) ) run_masked_udf_series(func, data, check_dtype=False) diff --git a/python/cudf/cudf/tests/series/methods/test_astype.py b/python/cudf/cudf/tests/series/methods/test_astype.py index ef37cf00923b..19c8ee2da6db 100644 --- a/python/cudf/cudf/tests/series/methods/test_astype.py +++ b/python/cudf/cudf/tests/series/methods/test_astype.py @@ -1393,7 +1393,7 @@ def test_typecast_from_float_to_decimal( request.applymarker( pytest.mark.xfail( float_types_as_str == "float32" and to_dtype.precision > 12, - reason="https://github.com/rapidsai/cudf/issues/14169", + reason="https://github.com/NVIDIA/cudf/issues/14169", ) ) got = data.astype(float_types_as_str) @@ -1639,7 +1639,7 @@ def test_astype_aware_to_naive_raises(): ) @pytest.mark.parametrize("dtype", ["int32", "int64", "uint64"]) def test_string_astype_int_pep515_underscores(data, dtype): - # https://github.com/rapidsai/cudf/issues/12047 + # https://github.com/NVIDIA/cudf/issues/12047 # Python (PEP 515) allows underscores between digits, so pandas # parses "123_1" as 1231; cudf should match. got = cudf.Series(data).astype(dtype) @@ -1658,7 +1658,7 @@ def test_string_astype_int_pep515_underscores(data, dtype): ) @pytest.mark.parametrize("dtype", ["int32", "int64", "uint64"]) def test_string_astype_int_invalid_underscores_raises(data, dtype): - # https://github.com/rapidsai/cudf/issues/12047 + # https://github.com/NVIDIA/cudf/issues/12047 # Underscores not surrounded by digits are invalid (PEP 515); # both pandas and cudf must reject them. assert_exceptions_equal( diff --git a/python/cudf/cudf/tests/series/methods/test_memory_usage.py b/python/cudf/cudf/tests/series/methods/test_memory_usage.py index 683e7ed61f23..64c0d0340d94 100644 --- a/python/cudf/cudf/tests/series/methods/test_memory_usage.py +++ b/python/cudf/cudf/tests/series/methods/test_memory_usage.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import cudf @@ -19,7 +19,7 @@ def test_empty_nested_list_uninitialized_offsets_memory_usage(): [[[1, 2], [3]], []], dtype=cudf.ListDtype(cudf.ListDtype("int64")) ) # Unlike in Arrow a 0-size list column produced by libcudf could have no offsets - # allocated and thus be truly empty (https://github.com/rapidsai/cudf/issues/16164) + # allocated and thus be truly empty (https://github.com/NVIDIA/cudf/issues/16164) assert ser.iloc[:0].memory_usage() == 0 diff --git a/python/cudf/cudf/tests/series/test_binops.py b/python/cudf/cudf/tests/series/test_binops.py index aaae6365cda2..2a40cc644610 100644 --- a/python/cudf/cudf/tests/series/test_binops.py +++ b/python/cudf/cudf/tests/series/test_binops.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import datetime import decimal @@ -298,7 +298,7 @@ def test_timedelta_series_ops_with_scalars( reason=( "timedelta modulo by zero is dubiously defined in " "both pandas and cuDF " - "(see https://github.com/rapidsai/cudf/issues/5938)" + "(see https://github.com/NVIDIA/cudf/issues/5938)" ), ), ), @@ -1525,7 +1525,7 @@ def test_binop_bool_uint(request, binary_op_method, rhs): if binary_op_method in {"rmod", "rfloordiv"}: request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/12162" + reason="https://github.com/NVIDIA/cudf/issues/12162" ), ) psr = pd.Series([True, False, False]) @@ -1554,7 +1554,7 @@ def test_floordiv_zero_float64( @pytest.mark.parametrize("scalar_divisor", [False, True]) -@pytest.mark.xfail(reason="https://github.com/rapidsai/cudf/issues/12162") +@pytest.mark.xfail(reason="https://github.com/NVIDIA/cudf/issues/12162") def test_floordiv_zero_bool(scalar_divisor): sr = pd.Series([True, True, False], dtype=np.bool_) cr = cudf.from_pandas(sr) diff --git a/python/cudf/cudf/tests/series/test_constructors.py b/python/cudf/cudf/tests/series/test_constructors.py index ad4457bee4f6..a47fa49eaf4d 100644 --- a/python/cudf/cudf/tests/series/test_constructors.py +++ b/python/cudf/cudf/tests/series/test_constructors.py @@ -450,7 +450,7 @@ def test_default_float_bitwidth_construction(default_float_bitwidth, data): def test_series_ordered_dedup(): - # part of https://github.com/rapidsai/cudf/issues/11486 + # part of https://github.com/NVIDIA/cudf/issues/11486 rng = np.random.default_rng(seed=0) sr = cudf.Series(rng.integers(0, 100, 1000)) # pandas unique() preserves order diff --git a/python/cudf/cudf/tests/test_no_cuinit.py b/python/cudf/cudf/tests/test_no_cuinit.py index b6dc12c84eaf..19f3acc392c0 100644 --- a/python/cudf/cudf/tests/test_no_cuinit.py +++ b/python/cudf/cudf/tests/test_no_cuinit.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2023-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2023-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import subprocess @@ -55,7 +55,7 @@ def test_rapids_no_initialize_cuinit( # Intercepting the call to cuInit programmatically is tricky since # the way it is resolved from dynamic libraries by # cuda-python/numba/cupy is multitudinous (see discussion at - # https://github.com/rapidsai/cudf/pull/12361 which does this, but + # https://github.com/NVIDIA/cudf/pull/12361 which does this, but # needs provide hooks that override dlsym, cuGetProcAddress, and # cuInit. # Instead, we just run under GDB and see if we hit a breakpoint diff --git a/python/cudf/cudf_pandas_tests/third_party_integration_tests/dependencies.yaml b/python/cudf/cudf_pandas_tests/third_party_integration_tests/dependencies.yaml index 6db9b550e7a9..c7d4aa83d05e 100644 --- a/python/cudf/cudf_pandas_tests/third_party_integration_tests/dependencies.yaml +++ b/python/cudf/cudf_pandas_tests/third_party_integration_tests/dependencies.yaml @@ -63,7 +63,7 @@ files: - py_version - test_base - test_stumpy - # TODO: Re-enable https://github.com/rapidsai/cudf/issues/21875 + # TODO: Re-enable https://github.com/NVIDIA/cudf/issues/21875 # test_tensorflow: # output: none # includes: diff --git a/python/cudf/pyproject.toml b/python/cudf/pyproject.toml index e9fef847e955..c9e275f545d6 100644 --- a/python/cudf/pyproject.toml +++ b/python/cudf/pyproject.toml @@ -84,7 +84,7 @@ cudf-pandas-tests = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" Documentation = "https://docs.rapids.ai/api/cudf/stable/" [tool.pydistcheck] diff --git a/python/cudf_kafka/pyproject.toml b/python/cudf_kafka/pyproject.toml index 125d35f77c12..3e5ff58eb98f 100644 --- a/python/cudf_kafka/pyproject.toml +++ b/python/cudf_kafka/pyproject.toml @@ -30,7 +30,7 @@ test = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" Documentation = "https://docs.rapids.ai/api/cudf/stable/" [tool.ruff] diff --git a/python/cudf_polars/cudf_polars/containers/column.py b/python/cudf_polars/cudf_polars/containers/column.py index bca694352f6c..3790057ac4eb 100644 --- a/python/cudf_polars/cudf_polars/containers/column.py +++ b/python/cudf_polars/cudf_polars/containers/column.py @@ -580,7 +580,7 @@ def nan_count(self, stream: Stream) -> int: """ result: int if self.size > 0 and plc.traits.is_floating_point(self.obj.type()): - # See https://github.com/rapidsai/cudf/issues/20202 for we type ignore + # See https://github.com/NVIDIA/cudf/issues/20202 for we type ignore result = plc.reduce.reduce( # type: ignore[assignment] plc.unary.is_nan(self.obj, stream=stream), plc.aggregation.sum(), diff --git a/python/cudf_polars/cudf_polars/dsl/expressions/datetime.py b/python/cudf_polars/cudf_polars/dsl/expressions/datetime.py index 3fd568d4b2d3..86cc4a76a9c4 100644 --- a/python/cudf_polars/cudf_polars/dsl/expressions/datetime.py +++ b/python/cudf_polars/cudf_polars/dsl/expressions/datetime.py @@ -184,7 +184,7 @@ def __init__( every = cast("Literal", self.children[1]).value match = re.fullmatch(r"(\d+)(ns|us|ms|s|m|h|d)", every) if match is None or int(match.group(1)) != 1: - # https://github.com/rapidsai/cudf/issues/18654 to support non-1 buckets + # https://github.com/NVIDIA/cudf/issues/18654 to support non-1 buckets raise NotImplementedError(f"Unsupported bucket: {every!r}") self.options = (self._TRUNCATE_FREQ_MAP[match.group(2)],) diff --git a/python/cudf_polars/cudf_polars/dsl/expressions/string.py b/python/cudf_polars/cudf_polars/dsl/expressions/string.py index be747c608837..765fc7da0038 100644 --- a/python/cudf_polars/cudf_polars/dsl/expressions/string.py +++ b/python/cudf_polars/cudf_polars/dsl/expressions/string.py @@ -954,7 +954,7 @@ def do_evaluate( plc_col, 0, stream=df.stream ).to_py(stream=df.stream) - # See https://github.com/rapidsai/cudf/issues/20202 for we type ignore + # See https://github.com/NVIDIA/cudf/issues/20202 for we type ignore format = _infer_datetime_format(first_valid_data) # type: ignore[arg-type] if not format: raise InvalidOperationError( @@ -1013,7 +1013,7 @@ def do_evaluate( (char,) = self.options # TODO: Maybe accept a string scalar in # cudf::strings::pad to avoid DtoH transfer - # See https://github.com/rapidsai/cudf/issues/20202 + # See https://github.com/NVIDIA/cudf/issues/20202 width_py = width_col.obj.to_scalar(stream=df.stream).to_py(stream=df.stream) assert width_py is not None pad_width = int(width_py) diff --git a/python/cudf_polars/cudf_polars/dsl/expressions/struct.py b/python/cudf_polars/cudf_polars/dsl/expressions/struct.py index 82f9b7ad1087..18885ec9724a 100644 --- a/python/cudf_polars/cudf_polars/dsl/expressions/struct.py +++ b/python/cudf_polars/cudf_polars/dsl/expressions/struct.py @@ -37,8 +37,8 @@ class Name(IntEnum): PrefixFields = auto() SuffixFields = auto() JsonEncode = auto() - WithFields = auto() # TODO: https://github.com/rapidsai/cudf/issues/19284 - MapFieldNames = auto() # TODO: https://github.com/rapidsai/cudf/issues/19285 + WithFields = auto() # TODO: https://github.com/NVIDIA/cudf/issues/19284 + MapFieldNames = auto() # TODO: https://github.com/NVIDIA/cudf/issues/19285 FieldByIndex = auto() MultipleFields = ( auto() @@ -108,7 +108,7 @@ def do_evaluate( dtype=self.dtype, ) elif self.name == StructFunction.Name.JsonEncode: - # Once https://github.com/rapidsai/cudf/issues/19338 is implemented, + # Once https://github.com/NVIDIA/cudf/issues/19338 is implemented, # we can use do this conversion on host. if column.size == 0: # write_json emits no lines for an empty input, which makes diff --git a/python/cudf_polars/cudf_polars/dsl/ir.py b/python/cudf_polars/cudf_polars/dsl/ir.py index 559777f43c7d..160dad312068 100644 --- a/python/cudf_polars/cudf_polars/dsl/ir.py +++ b/python/cudf_polars/cudf_polars/dsl/ir.py @@ -438,7 +438,7 @@ def run_source_function( # We pass predicate=None and apply any pushed predicate on the # GPU in process_chunk. # TODO: forward the pushed predicate to a RankAwareSource so a GPU-aware source - # can apply it at read time. See https://github.com/rapidsai/cudf/issues/22917. + # can apply it at read time. See https://github.com/NVIDIA/cudf/issues/22917. if rank_aware_source is not None: source_chunks = rank_aware_source( with_columns, None, None, None, rank=rank, nranks=nranks @@ -510,7 +510,7 @@ def process_chunk( # Validate against the declared (output) schema. Polars performs this # check for register_io_source(..., validate_schema=True), but the flag is # not exposed to the GPU plan, so we always validate. - # See https://github.com/rapidsai/cudf/issues/23043 + # See https://github.com/NVIDIA/cudf/issues/23043 declared = pl.Schema( {name: dtype.polars_type for name, dtype in schema.items()} ) @@ -573,7 +573,7 @@ def _parquet_physical_types( paths: list[str], columns: list[str] | None ) -> dict[str, plc.DataType]: # TODO: Use prefetched metadata - # https://github.com/rapidsai/cudf/issues/22940 + # https://github.com/NVIDIA/cudf/issues/22940 metadata = plc.io.parquet_metadata.read_parquet_metadata(plc.io.SourceInfo(paths)) column_types = metadata.schema().column_types() @@ -901,7 +901,7 @@ def _get_parquet_row_count_from_metadata( cached_parquet_info: list[CachedParquetInfo] | None, ) -> int: # Zero-width parquet files lose their row count when read through - # pylibcudf. See https://github.com/rapidsai/cudf/issues/21428 + # pylibcudf. See https://github.com/NVIDIA/cudf/issues/21428 if parquet_options.prefetch_file_metadata: if cached_parquet_info is None: raise AssertionError( @@ -1736,7 +1736,7 @@ def do_evaluate( df = df.select(projection) # Zero-width dataframes lose their row count when converted through - # pylibcudf. See https://github.com/rapidsai/cudf/issues/21428 + # pylibcudf. See https://github.com/NVIDIA/cudf/issues/21428 if len(schema) == 0: return DataFrame([], stream=context.get_cuda_stream(), num_rows=height) diff --git a/python/cudf_polars/cudf_polars/dsl/translate.py b/python/cudf_polars/cudf_polars/dsl/translate.py index 8316879b0c12..f814fdf32367 100644 --- a/python/cudf_polars/cudf_polars/dsl/translate.py +++ b/python/cudf_polars/cudf_polars/dsl/translate.py @@ -247,7 +247,7 @@ def unsupported_operations_error(self) -> NotImplementedError | None: return None unique_errors = sorted({str(e): e for e in self.errors}.values(), key=str) # TODO: Display these errors in user-friendly way, tracked in - # https://github.com/rapidsai/cudf/issues/17051 + # https://github.com/NVIDIA/cudf/issues/17051 formatted_errors = "\n".join( f"- {e.__class__.__name__}: {e}" for e in unique_errors ) @@ -426,7 +426,7 @@ def _(node: plrs._ir_nodes.PythonScan, translator: Translator, schema: Schema) - ) if nrows is not None: # A global row limit cannot be enforced independently per rank; tracked - # in https://github.com/rapidsai/cudf/issues/22918. + # in https://github.com/NVIDIA/cudf/issues/22918. raise NotImplementedError( "A row limit (head/limit) on a PythonScan source is not supported." ) @@ -1063,12 +1063,12 @@ def _( # cudf-polars has no concept of chunking, so we can just # drop it. # Note: This could be a plan hook for explicit repartition for streaming engines - # https://github.com/rapidsai/cudf/pull/23192#discussion_r3553113408 + # https://github.com/NVIDIA/cudf/pull/23192#discussion_r3553113408 (child,) = children return child if name == "fused": # TODO: fuse into a single kernel via JIT transform, see - # https://github.com/rapidsai/cudf/issues/21456. We don't use + # https://github.com/NVIDIA/cudf/issues/21456. We don't use # libcudf AST here because it widens the dtype for integer types # narrower than int32 (e.g. int8*int8 to int32), then fails # with a type mismatch when doing the add/sub with the third operand. diff --git a/python/cudf_polars/cudf_polars/dsl/utils/aggregations.py b/python/cudf_polars/cudf_polars/dsl/utils/aggregations.py index 4b66ebb1846a..aad2d1205cce 100644 --- a/python/cudf_polars/cudf_polars/dsl/utils/aggregations.py +++ b/python/cudf_polars/cudf_polars/dsl/utils/aggregations.py @@ -258,7 +258,7 @@ def decompose_single_agg( # first_non_null extracts the selected value, and separately # sum the predicate to validate item cardinality. # TODO: Use libcudf predicated aggregations when available: - # https://github.com/rapidsai/cudf/issues/22947 + # https://github.com/NVIDIA/cudf/issues/22947 aggs, _ = decompose_single_agg( expr.NamedExpr(next(name_generator), selected), name_generator, diff --git a/python/cudf_polars/cudf_polars/dsl/utils/io.py b/python/cudf_polars/cudf_polars/dsl/utils/io.py index 75373afeed96..091fa6bb403c 100644 --- a/python/cudf_polars/cudf_polars/dsl/utils/io.py +++ b/python/cudf_polars/cudf_polars/dsl/utils/io.py @@ -68,7 +68,7 @@ def _prefetch_parquet_footers_for_paths(paths: list[str]) -> list[CachedParquetI metadata The list of ``FileMetaData`` objects for the ``paths``. """ - # TODO: https://github.com/rapidsai/cudf/issues/22734, use object metadata from polars + # TODO: https://github.com/NVIDIA/cudf/issues/22734, use object metadata from polars # For now, we'll just use kvikio to explicitly get the size. sizes: list[int | None] = [] diff --git a/python/cudf_polars/cudf_polars/engine/rank_local_store.py b/python/cudf_polars/cudf_polars/engine/rank_local_store.py index 29b5cd89dcf6..2d334467f5a3 100644 --- a/python/cudf_polars/cudf_polars/engine/rank_local_store.py +++ b/python/cudf_polars/cudf_polars/engine/rank_local_store.py @@ -79,7 +79,7 @@ def pop(self, query_id: uuid.UUID, rank: int) -> DataFrame: "scanned more than once (for example a self-join, or " "collecting the same LazyFrame twice). Call engine.execute() " "again for a fresh result. Re-scan support is tracked as " - "future work (see https://github.com/rapidsai/cudf/issues/23115)." + "future work (see https://github.com/NVIDIA/cudf/issues/23115)." ) from None return df diff --git a/python/cudf_polars/cudf_polars/quent/_types.py b/python/cudf_polars/cudf_polars/quent/_types.py index 37299162a05b..e1894afc022d 100644 --- a/python/cudf_polars/cudf_polars/quent/_types.py +++ b/python/cudf_polars/cudf_polars/quent/_types.py @@ -3,7 +3,7 @@ """Quent telemetry tracing.""" -# TODO: https://github.com/rapidsai/cudf/issues/22949 use the bindings from quent +# TODO: https://github.com/NVIDIA/cudf/issues/22949 use the bindings from quent from __future__ import annotations diff --git a/python/cudf_polars/cudf_polars/streaming/actor_graph/collectives/sort.py b/python/cudf_polars/cudf_polars/streaming/actor_graph/collectives/sort.py index 8bfb1c8a86ab..e8de685d45cc 100644 --- a/python/cudf_polars/cudf_polars/streaming/actor_graph/collectives/sort.py +++ b/python/cudf_polars/cudf_polars/streaming/actor_graph/collectives/sort.py @@ -492,7 +492,7 @@ def _build_order_scheme( strict_boundaries = ( n_rows == 0 # TODO: Use unique_count_table - # Requires https://github.com/rapidsai/cudf/pull/22487 + # Requires https://github.com/NVIDIA/cudf/pull/22487 or plc.stream_compaction.unique( by_table, list(range(n_keys)), diff --git a/python/cudf_polars/cudf_polars/streaming/actor_graph/io.py b/python/cudf_polars/cudf_polars/streaming/actor_graph/io.py index e89035e12e2a..44957443e726 100644 --- a/python/cudf_polars/cudf_polars/streaming/actor_graph/io.py +++ b/python/cudf_polars/cudf_polars/streaming/actor_graph/io.py @@ -326,7 +326,7 @@ def _find_rank_aware_source(scan_fn: Callable[..., Any]) -> RankAwareSource | No This reaches into Polars' ``register_io_source`` closure layout (the captured source object). It is the only available hook today. When Polars exposes a supported way to thread state into a source this should move to it. See - https://github.com/rapidsai/cudf/issues/22917. + https://github.com/NVIDIA/cudf/issues/22917. """ for cell in getattr(scan_fn, "__closure__", ()): source = cell.cell_contents diff --git a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q1.py b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q1.py index a5aebfa87dda..5d5a71c9b361 100644 --- a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q1.py +++ b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q1.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 """Query 1.""" @@ -76,7 +76,7 @@ def polars_impl(run_config: RunConfig) -> QueryResult: .group_by(["sr_customer_sk", "sr_store_sk"]) .agg( # Polars sum() returns 0 for all-null groups; SQL returns NULL. - # See https://github.com/rapidsai/cudf/issues/19560. + # See https://github.com/NVIDIA/cudf/issues/19560. pl.when(pl.col("sr_return_amt").count() > 0) .then(pl.col("sr_return_amt").sum()) .otherwise(None) diff --git a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q24.py b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q24.py index fe506bf767cf..04590757c9fa 100644 --- a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q24.py +++ b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q24.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION. All rights reserved. +# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 @@ -139,7 +139,7 @@ def polars_impl(run_config: RunConfig) -> QueryResult: ) .agg( # Polars sum() returns 0 for all-null groups; SQL returns NULL. - # See https://github.com/rapidsai/cudf/issues/19560. + # See https://github.com/NVIDIA/cudf/issues/19560. pl.when(pl.col(amountone).count() > 0) .then(pl.col(amountone).sum()) .otherwise(None) @@ -159,7 +159,7 @@ def polars_impl(run_config: RunConfig) -> QueryResult: .group_by(["c_last_name", "c_first_name", "s_store_name"]) .agg( # Polars sum() returns 0 for all-null groups; SQL returns NULL. - # See https://github.com/rapidsai/cudf/issues/19560. + # See https://github.com/NVIDIA/cudf/issues/19560. pl.when(pl.col("netpaid").count() > 0) .then(pl.col("netpaid").sum()) .otherwise(None) diff --git a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q56.py b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q56.py index 646ad01c8f7a..be80304cd47b 100644 --- a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q56.py +++ b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q56.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 """Query 56.""" @@ -169,7 +169,7 @@ def polars_impl(run_config: RunConfig) -> QueryResult: .group_by("i_item_id") .agg( # Polars sum() returns 0 for all-null groups; SQL returns NULL. - # See https://github.com/rapidsai/cudf/issues/19560. + # See https://github.com/NVIDIA/cudf/issues/19560. pl.when(pl.col(str(ch["ext_col"])).count() > 0) .then(pl.col(str(ch["ext_col"])).sum()) .otherwise(None) @@ -188,7 +188,7 @@ def polars_impl(run_config: RunConfig) -> QueryResult: .group_by("i_item_id") .agg( # Polars sum() returns 0 for all-null groups; SQL returns NULL. - # See https://github.com/rapidsai/cudf/issues/19560. + # See https://github.com/NVIDIA/cudf/issues/19560. pl.when(pl.col("total_sales").count() > 0) .then(pl.col("total_sales").sum()) .otherwise(None) diff --git a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q64.py b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q64.py index b7046d2e65ac..9a60220e67bf 100644 --- a/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q64.py +++ b/python/cudf_polars/cudf_polars/streaming/benchmarks/pdsds_queries/q64.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 """Query 64.""" @@ -363,7 +363,7 @@ def build_cross_sales_for_year(target_year: int) -> pl.LazyFrame: [ pl.len().alias("cnt"), # Polars sum() returns 0 for all-null groups; SQL returns NULL. - # See https://github.com/rapidsai/cudf/issues/19560. + # See https://github.com/NVIDIA/cudf/issues/19560. pl.when(pl.col("ss_wholesale_cost").count() > 0) .then(pl.col("ss_wholesale_cost").sum()) .otherwise(None) diff --git a/python/cudf_polars/cudf_polars/streaming/benchmarks/utils.py b/python/cudf_polars/cudf_polars/streaming/benchmarks/utils.py index b07d400d35d0..6846c48ce9a3 100644 --- a/python/cudf_polars/cudf_polars/streaming/benchmarks/utils.py +++ b/python/cudf_polars/cudf_polars/streaming/benchmarks/utils.py @@ -355,7 +355,7 @@ def from_index(cls, index: int) -> GPUInfo: except pynvml.NVMLError_NotSupported: # Happens on systems without traditional GPU memory (e.g., Grace Hopper), # where nvmlDeviceGetMemoryInfo is not supported. - # See: https://github.com/rapidsai/cudf/issues/19427 + # See: https://github.com/NVIDIA/cudf/issues/19427 return cls( name=pynvml.nvmlDeviceGetName(handle), index=index, diff --git a/python/cudf_polars/cudf_polars/streaming/expressions.py b/python/cudf_polars/cudf_polars/streaming/expressions.py index 887e9e9fecce..8cb072cee6dc 100644 --- a/python/cudf_polars/cudf_polars/streaming/expressions.py +++ b/python/cudf_polars/cudf_polars/streaming/expressions.py @@ -524,7 +524,7 @@ def _decompose( # All child IRs were Empty. Use an Empty({}) with # count=1 to ensure that scalar expressions still # produce one output partition with a single row - # See: https://github.com/rapidsai/cudf/pull/20409 + # See: https://github.com/NVIDIA/cudf/pull/20409 input_ir = Empty({}) partition_info[input_ir] = PartitionInfo(count=1) diff --git a/python/cudf_polars/cudf_polars/streaming/parallel.py b/python/cudf_polars/cudf_polars/streaming/parallel.py index 2ac5c8c2eef7..302aa75fa1b2 100644 --- a/python/cudf_polars/cudf_polars/streaming/parallel.py +++ b/python/cudf_polars/cudf_polars/streaming/parallel.py @@ -366,7 +366,7 @@ def _( expr.is_pointwise for expr in traversal([ir.mask.value]) ): # TODO: Use expression decomposition to lower Filter - # See: https://github.com/rapidsai/cudf/issues/20076 + # See: https://github.com/NVIDIA/cudf/issues/20076 return _lower_ir_fallback( ir, rec, msg="This filter is not supported for multiple partitions." ) diff --git a/python/cudf_polars/cudf_polars/streaming/rank_aware_source.py b/python/cudf_polars/cudf_polars/streaming/rank_aware_source.py index c7fcae5ee7e8..135bb26c13f2 100644 --- a/python/cudf_polars/cudf_polars/streaming/rank_aware_source.py +++ b/python/cudf_polars/cudf_polars/streaming/rank_aware_source.py @@ -33,7 +33,7 @@ class SizedChunks: This is an interim cudf-polars API. It exists because Polars' IO-source contract returns a plain iterator with no length. It may be removed if Polars exposes a supported way to report the chunk count. Tracked in - https://github.com/rapidsai/cudf/issues/22917. + https://github.com/NVIDIA/cudf/issues/22917. Parameters ---------- @@ -91,7 +91,7 @@ class RankAwareSource(abc.ABC): This is an interim cudf-polars API. It exists because Polars' IO-source contract has no way to thread rank information into a source. It may be removed if Polars exposes a supported mechanism. Tracked in - https://github.com/rapidsai/cudf/issues/22917. + https://github.com/NVIDIA/cudf/issues/22917. The `rank` and `nranks` arguments default to `0` and `1` for the in-memory cudf-polars engine, the default Polars engine, and single-rank streaming. @@ -110,7 +110,7 @@ class RankAwareSource(abc.ABC): unwrapped instance is recognized; wrapping it in anything else (a `functools.partial`, closure, lambda, or decorator) hides it, in which case the source is treated as rank-unaware and runs on rank 0 only. This - limitation is tracked in https://github.com/rapidsai/cudf/issues/22917. + limitation is tracked in https://github.com/NVIDIA/cudf/issues/22917. Examples -------- @@ -171,7 +171,7 @@ def __call__( Maximum number of rows requested from the scan. cudf-polars does not support a pushed-down row limit and rejects it during translation, so this is always ``None`` on a GPU engine (see - https://github.com/rapidsai/cudf/issues/22918). It may be non-None + https://github.com/NVIDIA/cudf/issues/22918). It may be non-None on the default Polars CPU engine, where the source must honor it. batch_size Optional hint for the number of rows to yield per chunk. diff --git a/python/cudf_polars/cudf_polars/testing/inject_gpu_engine.py b/python/cudf_polars/cudf_polars/testing/inject_gpu_engine.py index b4527ecda137..95b1bb96737f 100644 --- a/python/cudf_polars/cudf_polars/testing/inject_gpu_engine.py +++ b/python/cudf_polars/cudf_polars/testing/inject_gpu_engine.py @@ -216,7 +216,7 @@ def pytest_report_header(config: pytest.Config) -> str: "tests/unit/io/test_parquet.py::test_binary_offset_roundtrip": "binary offset type unsupported", "tests/unit/lazyframe/test_engine_selection.py::test_engine_import_error_raises[gpu]": "Expect this to pass because cudf-polars is installed", "tests/unit/lazyframe/test_engine_selection.py::test_engine_import_error_raises[engine1]": "Expect this to pass because cudf-polars is installed", - "tests/unit/lazyframe/test_lazyframe.py::test_round[dtype2-123.55-1-123.6]": "libcudf HALF_EVEN rounding bug for Float64 with decimal_places > 0. See https://github.com/rapidsai/cudf/issues/21319", + "tests/unit/lazyframe/test_lazyframe.py::test_round[dtype2-123.55-1-123.6]": "libcudf HALF_EVEN rounding bug for Float64 with decimal_places > 0. See https://github.com/NVIDIA/cudf/issues/21319", "tests/unit/lazyframe/test_lazyframe.py::test_cast_frame": "Casting that raises not supported on GPU", "tests/unit/lazyframe/test_lazyframe.py::test_lazy_cache_hit": "Debug output on stderr doesn't match", "tests/unit/operations/aggregation/test_aggregations.py::test_binary_op_agg_context_no_simplify_expr_12423": "groupby-agg of just literals should not produce collect_list", @@ -256,8 +256,8 @@ def pytest_report_header(config: pytest.Config) -> str: "tests/unit/lazyframe/test_predicates.py::test_predicate_pushdown_split_pushable": "Casting that raises not supported on GPU", "tests/unit/lazyframe/test_predicates.py::test_filter_contradiction_fallible_error_handling": "Casting that raises not supported on GPU", "tests/unit/sql/test_miscellaneous.py::test_read_csv": "Incorrect handling of missing_is_null in read_csv", - "tests/unit/lazyframe/test_cse.py::test_cse_predicate_self_join[False]": "Debug output on stderr doesn't match, see https://github.com/rapidsai/cudf/issues/22967", - "tests/unit/lazyframe/test_cse.py::test_cse_predicate_self_join[True]": "Debug output on stderr doesn't match, see https://github.com/rapidsai/cudf/issues/22967", + "tests/unit/lazyframe/test_cse.py::test_cse_predicate_self_join[False]": "Debug output on stderr doesn't match, see https://github.com/NVIDIA/cudf/issues/22967", + "tests/unit/lazyframe/test_cse.py::test_cse_predicate_self_join[True]": "Debug output on stderr doesn't match, see https://github.com/NVIDIA/cudf/issues/22967", "tests/unit/io/test_scan_row_deletion.py::test_scan_row_deletion_skips_file_with_all_rows_deleted": "The test intentionally corrupts the parquet file, so we cannot read the row count from the header.", "tests/unit/io/test_multiscan.py::test_multiscan_row_index[scan_csv-write_csv-csv]": "Debug output on stderr doesn't match", "tests/unit/io/test_lazy_parquet.py::test_parquet_schema_arg[True-columns]": "allow_missing_columns argument in read_parquet not translated in IR", @@ -279,11 +279,11 @@ def pytest_report_header(config: pytest.Config) -> str: "tests/unit/operations/test_slice.py::test_schema_slice_on_literal_23999[lit1-0-len1-False]": "List literal loses nesting in slice: cudf#19610", "tests/unit/operations/test_slice.py::test_schema_slice_on_literal_23999[lit1-offset1-0-False]": "List literal loses nesting in slice: cudf#19610", "tests/unit/operations/test_slice.py::test_schema_slice_on_literal_23999[lit1-offset1-len1-False]": "List literal loses nesting in slice: cudf#19610", - "tests/unit/functions/test_concat.py::test_concat_with_empty_dataframes_strict_25725": "https://github.com/rapidsai/cudf/issues/21644", - "tests/unit/sql/test_window_functions.py::test_over_with_cumulative_window_funcs": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", - "tests/unit/sql/test_window_functions.py::test_window_cumulative_agg_with_nulls": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", - "tests/unit/sql/test_window_functions.py::test_window_named_window": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", - "tests/unit/operations/test_window.py::test_over_literal_cum_sum_26800": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", + "tests/unit/functions/test_concat.py::test_concat_with_empty_dataframes_strict_25725": "https://github.com/NVIDIA/cudf/issues/21644", + "tests/unit/sql/test_window_functions.py::test_over_with_cumulative_window_funcs": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", + "tests/unit/sql/test_window_functions.py::test_window_cumulative_agg_with_nulls": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", + "tests/unit/sql/test_window_functions.py::test_window_named_window": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", + "tests/unit/operations/test_window.py::test_over_literal_cum_sum_26800": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", "tests/unit/operations/namespaces/array/test_array.py::test_array_idx_size_limit_eval": "polars-internal IdxSize chunking debug assertion does not apply with the GPU engine", "tests/unit/operations/aggregation/test_aggregations.py::test_implode_and_agg": "implode + agg returns a mismatched dtype", "tests/unit/operations/aggregation/test_aggregations.py::test_duration_aggs": "Unsupported libcudf reduction operator for Duration dtype", @@ -323,14 +323,14 @@ def pytest_report_header(config: pytest.Config) -> str: "tests/unit/constructors/test_constructors.py::test_init_pydantic_2x": "pydantic deprecation warning", "tests/unit/constructors/test_constructors.py::test_init_structured_objects_nested[_TestFooPD-_TestBarPD-_TestBazPD]": "pydantic deprecation warning", "tests/unit/series/test_series.py::test_init_structured_objects": "pydantic deprecation warning", - "tests/unit/series/test_describe.py::test_series_describe_float": "https://github.com/rapidsai/cudf/issues/19324", - "tests/unit/series/test_describe.py::test_series_describe_int": "https://github.com/rapidsai/cudf/issues/19324", + "tests/unit/series/test_describe.py::test_series_describe_float": "https://github.com/NVIDIA/cudf/issues/19324", + "tests/unit/series/test_describe.py::test_series_describe_int": "https://github.com/NVIDIA/cudf/issues/19324", "tests/unit/streaming/test_streaming.py::test_streaming_apply": "https://github.com/pola-rs/polars/issues/22558", "tests/unit/operations/test_group_by_dynamic.py::test_group_by_dynamic_agg_bad_input_types[str]": "TODO: Need to investigate why this fails in CI but passes locally. We should fallback to CPU for group_by_dynamic", "tests/unit/expr/test_exprs.py::test_exp_log1p[Float16-Float16]": "Flaky test: Small floating-point precision differences in exp/log1p results", # TODO: Investigate why these tests fail in CI but pass locally. "tests/unit/io/test_delta.py::test_scan_delta_extract_table_statistics_df": "schemas mismatch: dtypes different", - "tests/unit/io/test_partition.py::test_sink_partitioned_no_columns_in_file_25535[scan_parquet-sink_parquet]": "Incorrect row count. Related to https://github.com/rapidsai/cudf/issues/21428", + "tests/unit/io/test_partition.py::test_sink_partitioned_no_columns_in_file_25535[scan_parquet-sink_parquet]": "Incorrect row count. Related to https://github.com/NVIDIA/cudf/issues/21428", "tests/unit/operations/test_group_by.py::test_unique_head_tail_26429[0]": "ZeroDivisionError: division by zero", # Flaky deadlock test, may occur on rtxpro6000 only "tests/unit/io/test_lazy_parquet.py::test_scan_parquet_in_mem_to_streaming_dispatch_deadlock_22641": "Flaky deadlock, may occur on rtxpro6000 only", @@ -381,7 +381,7 @@ def pytest_report_header(config: pytest.Config) -> str: # 1) Tests that are too slow with --inject-gpu-engine-blocksize=small due to many small partitions for large data STREAMING_ENGINE_TESTS_TO_SKIP: Mapping[str, str] = { "tests/unit/operations/aggregation/test_aggregations.py::test_boolean_aggs": "float difference in std/var in the unit of least precision", - # No deterministic key sort (https://github.com/rapidsai/cudf/issues/21641): + # No deterministic key sort (https://github.com/NVIDIA/cudf/issues/21641): # passes on some streaming runs and fails on others, so skip rather than # xfail to avoid a flaky XPASS/FAIL. "tests/unit/operations/test_group_by.py::test_group_by_unique_parametric[n_unique-True-True]": "non-deterministic key sort under the streaming engine", @@ -446,28 +446,28 @@ def pytest_report_header(config: pytest.Config) -> str: # xfail for tests that produce different results than CPU Polars STREAMING_ENGINE_EXPECTED_FAILURES: Mapping[str, str] = { - "tests/unit/functions/range/test_linear_space.py::test_linear_space_num_samples_expr": "https://github.com/rapidsai/cudf/issues/22072", - "tests/unit/functions/test_concat.py::test_concat_horizontal_zero_width_height_mismatch_26876": "https://github.com/rapidsai/cudf/issues/21644", + "tests/unit/functions/range/test_linear_space.py::test_linear_space_num_samples_expr": "https://github.com/NVIDIA/cudf/issues/22072", + "tests/unit/functions/test_concat.py::test_concat_horizontal_zero_width_height_mismatch_26876": "https://github.com/NVIDIA/cudf/issues/21644", "tests/unit/functions/test_concat.py::test_concat_horizontally_strict": "Correct polars.exceptions.ShapeError raised but it's in a ExceptionGroup", "tests/unit/functions/test_when_then.py::test_mismatched_height_should_raise[ternary_expr0-df0]": "Correct polars.exceptions.ShapeError raised but it's in a ExceptionGroup", "tests/unit/functions/test_when_then.py::test_mismatched_height_should_raise[ternary_expr0-df1]": "Correct polars.exceptions.ShapeError raised but it's in a ExceptionGroup", "tests/unit/functions/test_when_then.py::test_mismatched_height_should_raise[ternary_expr1-df0]": "Correct polars.exceptions.ShapeError raised but it's in a ExceptionGroup", "tests/unit/functions/test_when_then.py::test_mismatched_height_should_raise[ternary_expr1-df1]": "Correct polars.exceptions.ShapeError raised but it's in a ExceptionGroup", - "tests/unit/operations/test_slice.py::test_slice_pushdown_literal_projection_14349": "https://github.com/rapidsai/cudf/issues/22072", + "tests/unit/operations/test_slice.py::test_slice_pushdown_literal_projection_14349": "https://github.com/NVIDIA/cudf/issues/22072", "tests/unit/operations/test_group_by.py::test_group_by_lit_series": "Incorrect broadcasting of literals in groupby-agg", - "tests/unit/operations/test_group_by.py::test_group_by_series_partitioned": "https://github.com/rapidsai/cudf/issues/22072", - "tests/unit/operations/test_group_by.py::test_partitioned_group_by_chunked": "https://github.com/rapidsai/cudf/issues/22072", - "tests/unit/operations/test_group_by.py::test_unique_head_tail_26429[1]": "https://github.com/rapidsai/cudf/issues/22075", - "tests/unit/operations/test_group_by.py::test_unique_head_tail_26429[4]": "https://github.com/rapidsai/cudf/issues/22075", + "tests/unit/operations/test_group_by.py::test_group_by_series_partitioned": "https://github.com/NVIDIA/cudf/issues/22072", + "tests/unit/operations/test_group_by.py::test_partitioned_group_by_chunked": "https://github.com/NVIDIA/cudf/issues/22072", + "tests/unit/operations/test_group_by.py::test_unique_head_tail_26429[1]": "https://github.com/NVIDIA/cudf/issues/22075", + "tests/unit/operations/test_group_by.py::test_unique_head_tail_26429[4]": "https://github.com/NVIDIA/cudf/issues/22075", "tests/unit/operations/aggregation/test_aggregations.py::test_item_too_many": "Correct polars.exceptions.ComputeError raised but it's in an ExceptionGroup", "tests/unit/operations/aggregation/test_aggregations.py::test_single_empty": "Correct polars.exceptions.ComputeError raised but it's in an ExceptionGroup", - "tests/unit/operations/test_join.py::test_empty_outer_join_22206": "https://github.com/rapidsai/cudf/issues/22084", + "tests/unit/operations/test_join.py::test_empty_outer_join_22206": "https://github.com/NVIDIA/cudf/issues/22084", "tests/unit/operations/test_replace.py::test_replace_invalid_old_dtype": "Correct InvalidOperationError raised but it's in an ExceptionGroup", - "tests/unit/operations/test_window.py::test_over_literal_cum_sum_26800": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", - "tests/unit/sql/test_joins.py::test_cross_join_unnest_from_cte": "https://github.com/rapidsai/cudf/issues/22073", - "tests/unit/sql/test_window_functions.py::test_over_with_cumulative_window_funcs": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", - "tests/unit/sql/test_window_functions.py::test_window_cumulative_agg_with_nulls": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", - "tests/unit/sql/test_window_functions.py::test_window_multiple_named_window": "TODO: https://github.com/rapidsai/cudf/pull/22048#discussion_r3238041970", + "tests/unit/operations/test_window.py::test_over_literal_cum_sum_26800": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", + "tests/unit/sql/test_joins.py::test_cross_join_unnest_from_cte": "https://github.com/NVIDIA/cudf/issues/22073", + "tests/unit/sql/test_window_functions.py::test_over_with_cumulative_window_funcs": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", + "tests/unit/sql/test_window_functions.py::test_window_cumulative_agg_with_nulls": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", + "tests/unit/sql/test_window_functions.py::test_window_multiple_named_window": "TODO: https://github.com/NVIDIA/cudf/pull/22048#discussion_r3238041970", "tests/unit/functions/test_concat.py::test_concat_horizontal_lazy_strict_raises_shape_error_27415": "horizontal-concat strict height-mismatch raised inside an ExceptionGroup under the streaming engine", "tests/unit/io/test_io_plugin.py::test_defer_validate_true": "correct SchemaError raised but wrapped in an ExceptionGroup under the streaming engine", "tests/unit/operations/test_slice.py::test_hconcat_tail_unequal_heights_strict_raises_27552": "horizontal-concat strict height-mismatch raised inside an ExceptionGroup under the streaming engine", diff --git a/python/cudf_polars/docs/overview.md b/python/cudf_polars/docs/overview.md index 1a5a51d206f8..daf249902630 100644 --- a/python/cudf_polars/docs/overview.md +++ b/python/cudf_polars/docs/overview.md @@ -8,7 +8,7 @@ You will need: preferred configuration. Or else, use [rustup](https://rust-lang.org/tools/install/) 2. A [cudf development - environment](https://github.com/rapidsai/cudf/blob/main/CONTRIBUTING.md#setting-up-your-build-environment). + environment](https://github.com/NVIDIA/cudf/blob/main/CONTRIBUTING.md#setting-up-your-build-environment). The combined devcontainer works, or whatever your favourite approach is. :::{note} diff --git a/python/cudf_polars/pyproject.toml b/python/cudf_polars/pyproject.toml index 44825a801a42..aebde634ec00 100644 --- a/python/cudf_polars/pyproject.toml +++ b/python/cudf_polars/pyproject.toml @@ -68,7 +68,7 @@ dask = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" [tool.setuptools.dynamic] version = {file = "cudf_polars/VERSION"} @@ -100,7 +100,7 @@ exclude_also = [ "assert_never\\(" ] # The cudf_polars test suite are missing some coverage. -# See https://github.com/rapidsai/cudf/issues/22568. +# See https://github.com/NVIDIA/cudf/issues/22568. omit = [ "cudf_polars/testing/inject_gpu_engine.py", "cudf_polars/streaming/**", diff --git a/python/cudf_polars/tests/containers/test_column.py b/python/cudf_polars/tests/containers/test_column.py index 19b249e2687e..c2af58f5d955 100644 --- a/python/cudf_polars/tests/containers/test_column.py +++ b/python/cudf_polars/tests/containers/test_column.py @@ -228,7 +228,7 @@ def test_serialize_cache_miss(): assert frames[0].nbytes > 0 assert frames[1].nbytes > 0 - # https://github.com/rapidsai/cudf/pull/18953 + # https://github.com/NVIDIA/cudf/pull/18953 # In a multi-GPU setup, we might attempt to deserialize a column # whose type we haven't seen before. polars lets you use either the # class (`pl.Int8`) or an instance (`pl.Int8()`) in most places diff --git a/python/cudf_polars/tests/expressions/test_agg.py b/python/cudf_polars/tests/expressions/test_agg.py index b363619e192f..81b7614598a1 100644 --- a/python/cudf_polars/tests/expressions/test_agg.py +++ b/python/cudf_polars/tests/expressions/test_agg.py @@ -57,7 +57,7 @@ def is_sorted(request): @pytest.fixture def xfail_if_sorted(is_sorted, request): - # See https://github.com/rapidsai/cudf/pull/20791#issuecomment-3750528419 + # See https://github.com/NVIDIA/cudf/pull/20791#issuecomment-3750528419 if is_sorted: request.applymarker( pytest.mark.xfail(reason="See https://github.com/pola-rs/polars/pull/24981") diff --git a/python/cudf_polars/tests/expressions/test_casting.py b/python/cudf_polars/tests/expressions/test_casting.py index ac869bda9601..f3e828b429af 100644 --- a/python/cudf_polars/tests/expressions/test_casting.py +++ b/python/cudf_polars/tests/expressions/test_casting.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations @@ -94,7 +94,7 @@ def test_cast_to_string_unsupported(engine: pl.GPUEngine): def test_float_to_decimal_rounding(engine: pl.GPUEngine): - # See https://github.com/rapidsai/cudf/pull/21450 + # See https://github.com/NVIDIA/cudf/pull/21450 df = pl.LazyFrame( { "foo": [Decimal("16954168.35")], diff --git a/python/cudf_polars/tests/expressions/test_gather.py b/python/cudf_polars/tests/expressions/test_gather.py index 9b4d4186befb..cd1c319e256d 100644 --- a/python/cudf_polars/tests/expressions/test_gather.py +++ b/python/cudf_polars/tests/expressions/test_gather.py @@ -79,7 +79,7 @@ def test_gather_out_of_bounds(engine_raise_on_fail: pl.GPUEngine, negative): pytest.param( pl.lit([7]), marks=pytest.mark.xfail( - reason="List literal loses nesting in gather: https://github.com/rapidsai/cudf/issues/19610" + reason="List literal loses nesting in gather: https://github.com/NVIDIA/cudf/issues/19610" ), ), pl.lit([[7]]), diff --git a/python/cudf_polars/tests/streaming/test_dataframescan.py b/python/cudf_polars/tests/streaming/test_dataframescan.py index 0fd27f2b57e8..a6e86045f2b6 100644 --- a/python/cudf_polars/tests/streaming/test_dataframescan.py +++ b/python/cudf_polars/tests/streaming/test_dataframescan.py @@ -87,7 +87,7 @@ def test_dataframescan_concat(request, df, streaming_engine_factory): # polars-CPU [A, B]. request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/22376", + reason="https://github.com/NVIDIA/cudf/issues/22376", strict=False, ) ) diff --git a/python/cudf_polars/tests/streaming/test_groupby.py b/python/cudf_polars/tests/streaming/test_groupby.py index 9bfb363147ce..2127fe03753a 100644 --- a/python/cudf_polars/tests/streaming/test_groupby.py +++ b/python/cudf_polars/tests/streaming/test_groupby.py @@ -266,7 +266,7 @@ def test_groupby_then_slice(streaming_engine, zlice: tuple[int, int]) -> None: def test_groupby_on_equality(streaming_engine) -> None: - # See: https://github.com/rapidsai/cudf/issues/19152 + # See: https://github.com/NVIDIA/cudf/issues/19152 df = pl.LazyFrame( { "key1": [1, 1, 1, 2, 3, 1, 4, 6, 7], diff --git a/python/cudf_polars/tests/streaming/test_join.py b/python/cudf_polars/tests/streaming/test_join.py index 2733b051925a..a8311af71ed3 100644 --- a/python/cudf_polars/tests/streaming/test_join.py +++ b/python/cudf_polars/tests/streaming/test_join.py @@ -192,7 +192,7 @@ def test_join_and_slice(request, zlice, streaming_engine_factory): # than the CPU baseline. request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/22405", + reason="https://github.com/NVIDIA/cudf/issues/22405", strict=False, ) ) @@ -212,7 +212,7 @@ def test_join_and_slice(request, zlice, streaming_engine_factory): ) q = left.join(right, on="a", how="inner").slice(*zlice) # Check that we get the correct row count - # See: https://github.com/rapidsai/cudf/issues/19153 + # See: https://github.com/NVIDIA/cudf/issues/19153 with warns_on_spmd( streaming_engine, UserWarning, diff --git a/python/cudf_polars/tests/streaming/test_metadata.py b/python/cudf_polars/tests/streaming/test_metadata.py index 62226513001f..c9a17fce6ee8 100644 --- a/python/cudf_polars/tests/streaming/test_metadata.py +++ b/python/cudf_polars/tests/streaming/test_metadata.py @@ -83,7 +83,7 @@ def test_rapidsmpf_join_metadata( # ``self._handle`` across worker/actor processes, so the # ``metadata_collector`` round-trip fails on Dask and Ray. # - # When https://github.com/rapidsai/cudf/pull/22394 lands, dedup of + # When https://github.com/NVIDIA/cudf/pull/22394 lands, dedup of # replicated outputs moves to the Dask/Ray frontends and the # ``duplicated`` flag's semantics change to "every rank holds the # data". Revisit the ``len(metadata_collector) == 1`` and @@ -255,7 +255,7 @@ def test_get_partitioning_moduli(partitioning, key_indices, nranks, expected) -> 4, NormalizedPartitioning(None, None), ), - # Resolves https://github.com/rapidsai/cudf/issues/21742 + # Resolves https://github.com/NVIDIA/cudf/issues/21742 ( Partitioning(inter_rank=HashScheme((0,), 8), local="inherit"), (1,), diff --git a/python/cudf_polars/tests/streaming/test_parallel.py b/python/cudf_polars/tests/streaming/test_parallel.py index 7940bcda0a59..b5310d6d7f47 100644 --- a/python/cudf_polars/tests/streaming/test_parallel.py +++ b/python/cudf_polars/tests/streaming/test_parallel.py @@ -44,7 +44,7 @@ def test_rename_multi(mapping, streaming_engine): def test_rename_concat(streaming_engine) -> None: - # https://github.com/rapidsai/cudf/pull/19121#issuecomment-2959305678 + # https://github.com/NVIDIA/cudf/pull/19121#issuecomment-2959305678 q = pl.concat( [ pl.LazyFrame({"a": [1, 2, 3]}).rename({"a": "A"}), diff --git a/python/cudf_polars/tests/streaming/test_rolling.py b/python/cudf_polars/tests/streaming/test_rolling.py index de1682c173eb..4f6d7e4527ff 100644 --- a/python/cudf_polars/tests/streaming/test_rolling.py +++ b/python/cudf_polars/tests/streaming/test_rolling.py @@ -212,7 +212,7 @@ def test_over_noncol_key_fallback(request, streaming_engine_factory, expr) -> No # invisible to ``pytest.warns``. request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/22405", + reason="https://github.com/NVIDIA/cudf/issues/22405", strict=False, ) ) @@ -344,7 +344,7 @@ def test_over_in_filter_unsupported(request, streaming_engine_factory) -> None: # doesn't preserve row order. request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/22405", + reason="https://github.com/NVIDIA/cudf/issues/22405", strict=False, ) ) diff --git a/python/cudf_polars/tests/streaming/test_select.py b/python/cudf_polars/tests/streaming/test_select.py index 16719d50ed24..2b83911e7d5b 100644 --- a/python/cudf_polars/tests/streaming/test_select.py +++ b/python/cudf_polars/tests/streaming/test_select.py @@ -171,7 +171,7 @@ def test_select_parquet_fast_count(tmp_path, df, engine): def test_select_literal(engine): - # See: https://github.com/rapidsai/cudf/issues/19147 + # See: https://github.com/NVIDIA/cudf/issues/19147 ldf = pl.LazyFrame({"a": list(range(10))}) q = ldf.select(pl.lit(2).pow(pl.lit(-3, dtype=pl.Float32))) assert_gpu_result_equal(q, engine=engine) diff --git a/python/cudf_polars/tests/streaming/test_union.py b/python/cudf_polars/tests/streaming/test_union.py index ecb8f1ebfe19..7837481ace3a 100644 --- a/python/cudf_polars/tests/streaming/test_union.py +++ b/python/cudf_polars/tests/streaming/test_union.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations @@ -10,7 +10,7 @@ def test_union_shared_fanout_no_deadlock(streaming_engine): # union actor can deadlock when input branches share a fanout. - # See https://github.com/rapidsai/cudf/issues/21750 + # See https://github.com/NVIDIA/cudf/issues/21750 n = 100 df = pl.LazyFrame({"key": list(range(50)) * (n // 50), "val": list(range(n))}) gb = df.group_by("key").agg(pl.col("val").sum()) diff --git a/python/cudf_polars/tests/test_cache.py b/python/cudf_polars/tests/test_cache.py index 1bfb0a2b646c..f710b11017d8 100644 --- a/python/cudf_polars/tests/test_cache.py +++ b/python/cudf_polars/tests/test_cache.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations @@ -72,7 +72,7 @@ def test_union_cache_nodes(): # (repeated 2 times) # Check that the concatenated Cache nodes are the same object - # See: https://github.com/rapidsai/cudf/issues/19766 + # See: https://github.com/NVIDIA/cudf/issues/19766 assert isinstance(qir, ir.Union) assert isinstance(qir.children[0], ir.Cache) assert isinstance(qir.children[1], ir.Cache) diff --git a/python/cudf_polars/tests/test_engine_execute.py b/python/cudf_polars/tests/test_engine_execute.py index 7fd5a15ea93d..6e3d4bfcb6f0 100644 --- a/python/cudf_polars/tests/test_engine_execute.py +++ b/python/cudf_polars/tests/test_engine_execute.py @@ -224,7 +224,7 @@ def test_spmd_execute_self_join_raises(spmd_engine): reused = result.lazy() # Polars does not dedup a reused plugin source, so the self-join scans the # partition twice; the second read finds it already consumed. Re-scan support - # is tracked as future work (see https://github.com/rapidsai/cudf/issues/23115). + # is tracked as future work (see https://github.com/NVIDIA/cudf/issues/23115). with pytest.RaisesGroup( pytest.RaisesExc(RuntimeError, match="consumed on read"), allow_unwrapped=True, diff --git a/python/cudf_polars/tests/test_groupby.py b/python/cudf_polars/tests/test_groupby.py index 4030727e54e0..f1f501f56cc2 100644 --- a/python/cudf_polars/tests/test_groupby.py +++ b/python/cudf_polars/tests/test_groupby.py @@ -163,7 +163,7 @@ def test_groupby_sorted_keys( pytest.mark.xfail( is_streaming_engine(engine), strict=False, - reason="https://github.com/rapidsai/cudf/issues/21642 - no deterministic sort for keys", + reason="https://github.com/NVIDIA/cudf/issues/21642 - no deterministic sort for keys", ) ) sorted_keys = [ @@ -463,7 +463,7 @@ def test_groupby_nan_minmax_raises(engine: pl.GPUEngine, op): condition=POLARS_VERSION_LT_140, reason="polars < 1.40 emits the list literal at its inner dtype, " "nested per group, which we mishandle: " - "https://github.com/rapidsai/cudf/issues/19610", + "https://github.com/NVIDIA/cudf/issues/19610", ), ), pl.col("float") * (1 - pl.col("int")), @@ -712,7 +712,7 @@ def test_groupby_sum_decimal_null_group( @pytest.mark.xfail( raises=AssertionError, - reason="https://github.com/rapidsai/cudf/issues/19610", + reason="https://github.com/NVIDIA/cudf/issues/19610", ) def test_groupby_literal_agg(engine: pl.GPUEngine): df = pl.LazyFrame({"c0": [True, False]}) diff --git a/python/cudf_polars/tests/test_mapfunction.py b/python/cudf_polars/tests/test_mapfunction.py index 3e476afc9b0e..25842dde8843 100644 --- a/python/cudf_polars/tests/test_mapfunction.py +++ b/python/cudf_polars/tests/test_mapfunction.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations @@ -101,7 +101,7 @@ def test_with_row_index_defaults(engine: pl.GPUEngine): def test_unique_hash(): - # https://github.com/rapidsai/cudf/pull/19121#issuecomment-2959305678 + # https://github.com/NVIDIA/cudf/pull/19121#issuecomment-2959305678 a = pl.LazyFrame({"a": [1, 2, 3]}).rename({"a": "A"}) b = pl.LazyFrame({"a": [4, 5, 6]}).rename({"a": "A"}) ir_a = Translator(a._ldf.visit(), pl.GPUEngine()).translate_ir() diff --git a/python/cudf_polars/tests/test_tracing.py b/python/cudf_polars/tests/test_tracing.py index 5b0f6c772f27..c9abc02c6500 100644 --- a/python/cudf_polars/tests/test_tracing.py +++ b/python/cudf_polars/tests/test_tracing.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from __future__ import annotations @@ -68,7 +68,7 @@ def test_trace_basic( def test_import_without_structlog(timeout_seconds: int) -> None: # This test could avoid the subprocess by monkeypatching sys.modules, but - # that was flaky. https://github.com/rapidsai/cudf/pull/22012#issuecomment-4284536686 + # that was flaky. https://github.com/NVIDIA/cudf/pull/22012#issuecomment-4284536686 # has more details. code = textwrap.dedent("""\ import sys diff --git a/python/cudf_polars/tests/testing/test_asserts.py b/python/cudf_polars/tests/testing/test_asserts.py index 0a6cadcf3e60..7e4130002ee5 100644 --- a/python/cudf_polars/tests/testing/test_asserts.py +++ b/python/cudf_polars/tests/testing/test_asserts.py @@ -465,7 +465,7 @@ def test_assert_tpch_result_equal_sort_keys_raises_not_sorted() -> None: def test_assert_tpch_result_equal_grouped_float_sort( sort_by: list[tuple[str, bool]], drop_columns: list[str] ) -> None: - # https://github.com/rapidsai/cudf/issues/22129 + # https://github.com/NVIDIA/cudf/issues/22129 # Same non-float values with float values reordered inside each non-float group. left = pl.DataFrame({"a": [1, 1, 1], "b": [2, 2, 2], "c": [1.0, 2.0, 3.0]}) right = pl.DataFrame({"a": [1, 1, 1], "b": [2, 2, 2], "c": [1.0, 2.999, 2.0]}) diff --git a/python/cudf_streaming/pyproject.toml b/python/cudf_streaming/pyproject.toml index b71674ee9ec1..f7cf81f83999 100644 --- a/python/cudf_streaming/pyproject.toml +++ b/python/cudf_streaming/pyproject.toml @@ -35,7 +35,7 @@ test = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" Documentation = "https://docs.rapids.ai/api/cudf/stable/" [tool.ruff] diff --git a/python/custreamz/custreamz/tests/test_dataframes.py b/python/custreamz/custreamz/tests/test_dataframes.py index 2c343be7ec0c..60b1f6270a24 100644 --- a/python/custreamz/custreamz/tests/test_dataframes.py +++ b/python/custreamz/custreamz/tests/test_dataframes.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 """ @@ -29,7 +29,7 @@ def client(): client = Client(processes=False, asynchronous=False) # Fix flaky tests seen in workflows like - # https://github.com/rapidsai/cudf/actions/runs/15119048978/job/42498435703?pr=18870#step:9:1722 + # https://github.com/NVIDIA/cudf/actions/runs/15119048978/job/42498435703?pr=18870#step:9:1722 # These manifest as a RecursionError in https://github.com/dask/distributed/blob/a890b85c8f107f7c8664ef96270ef8c25a2b31e4/distributed/gc.py#L201 # There isn't a public API for whether it's enabled or disabled. We'll just # assume that it's enabled and disable it for the duration of the tests. diff --git a/python/custreamz/pyproject.toml b/python/custreamz/pyproject.toml index 7f9a0f4c5f94..05c009341f91 100644 --- a/python/custreamz/pyproject.toml +++ b/python/custreamz/pyproject.toml @@ -46,7 +46,7 @@ test = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" [tool.rapids-build-backend] build-backend = "setuptools.build_meta" diff --git a/python/dask_cudf/dask_cudf/_expr/collection.py b/python/dask_cudf/dask_cudf/_expr/collection.py index 5091e4d7da39..a354eb25cfba 100644 --- a/python/dask_cudf/dask_cudf/_expr/collection.py +++ b/python/dask_cudf/dask_cudf/_expr/collection.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2024-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import warnings @@ -32,8 +32,8 @@ def _prepare_cov_corr(self, min_periods, numeric_only): # to 2 by default (which is not supported by cudf) # TODO: Remove when cudf supports both min_periods # and numeric_only - # See: https://github.com/rapidsai/cudf/issues/12626 - # See: https://github.com/rapidsai/cudf/issues/9009 + # See: https://github.com/NVIDIA/cudf/issues/12626 + # See: https://github.com/NVIDIA/cudf/issues/9009 self._meta.cov(min_periods=min_periods) frame = self @@ -44,7 +44,7 @@ def _prepare_cov_corr(self, min_periods, numeric_only): return frame, min_periods # var can be removed if cudf#15179 is addressed. - # See: https://github.com/rapidsai/cudf/issues/14935 + # See: https://github.com/NVIDIA/cudf/issues/14935 def var( self, axis=0, diff --git a/python/dask_cudf/dask_cudf/_expr/expr.py b/python/dask_cudf/dask_cudf/_expr/expr.py index eae79ace678e..ff5b4f6daf85 100644 --- a/python/dask_cudf/dask_cudf/_expr/expr.py +++ b/python/dask_cudf/dask_cudf/_expr/expr.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2024-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2024-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import functools @@ -32,7 +32,7 @@ class RenameAxisCudf(RenameAxis): # TODO: Remove this after rename_axis is supported in cudf - # (See: https://github.com/rapidsai/cudf/issues/16895) + # (See: https://github.com/NVIDIA/cudf/issues/16895) @staticmethod def operation(df, index=no_default, **kwargs): if index != no_default: @@ -71,7 +71,7 @@ def _simplify_down(self): # This can be removed after cudf#15176 is addressed. -# See: https://github.com/rapidsai/cudf/issues/15176 +# See: https://github.com/NVIDIA/cudf/issues/15176 class PatchCumulativeBlockwise(CumulativeBlockwise): @property def _args(self) -> list: @@ -142,7 +142,7 @@ def reduction_chunk(cls, x, skipna=True, numeric_only=False): avg = x.sum(**kwargs) / n if numeric_only: # Workaround for cudf bug - # (see: https://github.com/rapidsai/cudf/issues/13731) + # (see: https://github.com/NVIDIA/cudf/issues/13731) x = x[n.index] m2 = ((x - avg) ** 2).sum(**kwargs) return n, avg, m2 @@ -186,7 +186,7 @@ def _patched_var( # Temporary work-around for missing cudf + categorical support -# See: https://github.com/rapidsai/cudf/issues/11795 +# See: https://github.com/NVIDIA/cudf/issues/11795 # TODO: Fix RepartitionQuantiles and remove this in cudf>24.06 _original_get_divisions = _shuffle_module._get_divisions diff --git a/python/dask_cudf/dask_cudf/_expr/groupby.py b/python/dask_cudf/dask_cudf/_expr/groupby.py index b596fe7779e4..aadc846bc44b 100644 --- a/python/dask_cudf/dask_cudf/_expr/groupby.py +++ b/python/dask_cudf/dask_cudf/_expr/groupby.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2025-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import functools @@ -519,7 +519,7 @@ def _translate_arg(arg): # separate tasks. # (2) We need to work around missing `observed=False` # support: -# https://github.com/rapidsai/cudf/issues/15173 +# https://github.com/NVIDIA/cudf/issues/15173 class GroupBy(DXGroupBy): diff --git a/python/dask_cudf/dask_cudf/_legacy/io/parquet.py b/python/dask_cudf/dask_cudf/_legacy/io/parquet.py index 6c72da8171c9..086356873d09 100644 --- a/python/dask_cudf/dask_cudf/_legacy/io/parquet.py +++ b/python/dask_cudf/dask_cudf/_legacy/io/parquet.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import contextlib import itertools @@ -106,7 +106,7 @@ def _read_paths( ) except RuntimeError as err: # TODO: Remove try/except after null-schema issue is resolved - # (See: https://github.com/rapidsai/cudf/issues/12702) + # (See: https://github.com/NVIDIA/cudf/issues/12702) if len(paths) > 1: df = cudf.concat( [ @@ -385,7 +385,7 @@ def write_metadata(parts, fmd, fs, path, append=False, **kwargs): metadata_path = fs.sep.join([path, "_metadata"]) _meta = [] if append and fmd is not None: - # Convert to bytes: + # Convert to bytes: if isinstance(fmd, pq.FileMetaData): with BytesIO() as myio: fmd.write_metadata_file(myio) diff --git a/python/dask_cudf/dask_cudf/io/tests/test_csv.py b/python/dask_cudf/dask_cudf/io/tests/test_csv.py index a8c9e7ef8060..8a20c21ab0f3 100644 --- a/python/dask_cudf/dask_cudf/io/tests/test_csv.py +++ b/python/dask_cudf/dask_cudf/io/tests/test_csv.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import gzip @@ -243,7 +243,7 @@ def test_read_csv_skipfooter(csv_end_bad_lines): def test_read_csv_skipfooter_arrow_string_fail(request, csv_end_bad_lines): request.applymarker( pytest.mark.xfail( - reason="https://github.com/rapidsai/cudf/issues/14915", + reason="https://github.com/NVIDIA/cudf/issues/14915", ) ) ddf_cpu = dd.read_csv( diff --git a/python/dask_cudf/dask_cudf/io/tests/test_orc.py b/python/dask_cudf/dask_cudf/io/tests/test_orc.py index 19f7032cad85..b2190caebbf3 100644 --- a/python/dask_cudf/dask_cudf/io/tests/test_orc.py +++ b/python/dask_cudf/dask_cudf/io/tests/test_orc.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2018-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2018-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import glob @@ -107,7 +107,7 @@ def test_read_orc_filtered(tmpdir, engine, predicate, expected_len): def test_read_orc_first_file_empty(tmpdir): # Write a 3-file dataset where the first file is empty - # See: https://github.com/rapidsai/cudf/issues/8011 + # See: https://github.com/NVIDIA/cudf/issues/8011 path = str(tmpdir) os.makedirs(path, exist_ok=True) df1 = cudf.DataFrame({"id": [1, 2], "float": [1.0, 2.0]}) diff --git a/python/dask_cudf/dask_cudf/io/tests/test_parquet.py b/python/dask_cudf/dask_cudf/io/tests/test_parquet.py index 65bb16419f4a..7863ef9e0d48 100644 --- a/python/dask_cudf/dask_cudf/io/tests/test_parquet.py +++ b/python/dask_cudf/dask_cudf/io/tests/test_parquet.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import glob @@ -598,7 +598,7 @@ def test_null_partition(tmpdir): def test_nullable_schema_mismatch(tmpdir): - # See: https://github.com/rapidsai/cudf/issues/12702 + # See: https://github.com/NVIDIA/cudf/issues/12702 path0 = str(tmpdir.join("test.0.parquet")) path1 = str(tmpdir.join("test.1.parquet")) cudf.DataFrame.from_dict({"a": [1, 2, 3]}).to_parquet(path0) diff --git a/python/dask_cudf/dask_cudf/tests/test_core.py b/python/dask_cudf/dask_cudf/tests/test_core.py index fd8c2e386298..55aedc065a07 100644 --- a/python/dask_cudf/dask_cudf/tests/test_core.py +++ b/python/dask_cudf/dask_cudf/tests/test_core.py @@ -476,7 +476,7 @@ def test_repartition_hash(by, npartitions, max_branch): def test_repartition_no_extra_row(): - # see https://github.com/rapidsai/cudf/issues/11930 + # see https://github.com/NVIDIA/cudf/issues/11930 gdf = cudf.DataFrame({"a": [10, 20, 30], "b": [1, 2, 3]}).set_index("a") ddf = dask_cudf.from_cudf(gdf, npartitions=1) ddf_new = ddf.repartition([0, 5, 10, 30], force=True) @@ -760,7 +760,7 @@ def test_large_numbers_var(): def test_index_map_partitions(): - # https://github.com/rapidsai/cudf/issues/6738 + # https://github.com/NVIDIA/cudf/issues/6738 ddf = dd.from_pandas(pd.DataFrame({"a": range(10)}), npartitions=2) mins_pd = ddf.index.map_partitions(M.min, meta=ddf.index).compute() @@ -804,7 +804,7 @@ def test_merging_categorical_columns(): def test_correct_meta(): # Need these local imports in this specific order. - # For context: https://github.com/rapidsai/cudf/issues/7946 + # For context: https://github.com/NVIDIA/cudf/issues/7946 import pandas as pd from dask import dataframe as dd @@ -829,7 +829,7 @@ def test_categorical_dtype_round_trip(): assert ds.dtype.ordered is False # Below validations are required, see: - # https://github.com/rapidsai/cudf/issues/11487#issuecomment-1208912383 + # https://github.com/NVIDIA/cudf/issues/11487#issuecomment-1208912383 actual = ds.compute() expected = pds.compute() assert actual.dtype.ordered == expected.dtype.ordered @@ -922,7 +922,7 @@ def test_cov_corr(op, numeric_only): ddf = dd.from_pandas(df, npartitions=2) res = getattr(ddf, op)(numeric_only=numeric_only) # Use to_pandas until cudf supports numeric_only - # (See: https://github.com/rapidsai/cudf/issues/12626) + # (See: https://github.com/NVIDIA/cudf/issues/12626) expect = getattr(df.to_pandas(), op)(numeric_only=numeric_only) dd.assert_eq(res, expect) diff --git a/python/dask_cudf/dask_cudf/tests/test_groupby.py b/python/dask_cudf/dask_cudf/tests/test_groupby.py index c317e0e3009a..373320b08b58 100644 --- a/python/dask_cudf/dask_cudf/tests/test_groupby.py +++ b/python/dask_cudf/dask_cudf/tests/test_groupby.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 @@ -516,7 +516,7 @@ def test_groupby_reset_index_string_name(): def test_groupby_categorical_key(): - # See https://github.com/rapidsai/cudf/issues/4608 + # See https://github.com/NVIDIA/cudf/issues/4608 df = dask.datasets.timeseries(seed=1) gddf = df.to_backend("cudf") gddf["name"] = gddf["name"].astype("category") diff --git a/python/dask_cudf/dask_cudf/tests/test_join.py b/python/dask_cudf/dask_cudf/tests/test_join.py index 49ee61718952..fed8063c4136 100644 --- a/python/dask_cudf/dask_cudf/tests/test_join.py +++ b/python/dask_cudf/dask_cudf/tests/test_join.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2019-2024, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 from functools import partial @@ -377,7 +377,7 @@ def test_single_partition(): def test_issue_12773(): - # https://github.com/rapidsai/cudf/issues/12773 + # https://github.com/NVIDIA/cudf/issues/12773 df1 = cudf.DataFrame({"a": ["a", "b"], "b": [1, 2]}) df2 = cudf.DataFrame({"a": ["a", "c"], "b": [2, 3]}) diff --git a/python/dask_cudf/dask_cudf/tests/test_onehot.py b/python/dask_cudf/dask_cudf/tests/test_onehot.py index 698e694c0674..70f63b6723c6 100644 --- a/python/dask_cudf/dask_cudf/tests/test_onehot.py +++ b/python/dask_cudf/dask_cudf/tests/test_onehot.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2019-2025, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import pandas as pd @@ -107,7 +107,7 @@ def test_get_dummies_large(): def test_get_dummies_categorical(): - # https://github.com/rapidsai/cudf/issues/7111 + # https://github.com/NVIDIA/cudf/issues/7111 gdf = cudf.DataFrame({"A": ["a", "b", "b"], "B": [1, 2, 3]}) pdf = gdf.to_pandas() diff --git a/python/dask_cudf/dask_cudf/tests/test_reductions.py b/python/dask_cudf/dask_cudf/tests/test_reductions.py index d97fefb7ea91..bad6426e61b1 100644 --- a/python/dask_cudf/dask_cudf/tests/test_reductions.py +++ b/python/dask_cudf/dask_cudf/tests/test_reductions.py @@ -1,4 +1,4 @@ -# SPDX-FileCopyrightText: Copyright (c) 2021-2024, NVIDIA CORPORATION. +# SPDX-FileCopyrightText: Copyright (c) 2021-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved. # SPDX-License-Identifier: Apache-2.0 import pytest @@ -75,7 +75,7 @@ def test_rowwise_reductions(data, op): @pytest.mark.parametrize("skipna", [True, False]) def test_var_nulls(skipna): # Copied from 10min example notebook - # See: https://github.com/rapidsai/cudf/pull/15347 + # See: https://github.com/NVIDIA/cudf/pull/15347 s = cudf.Series([1, 2, 3, None, 4]) ds = dask_cudf.from_cudf(s, npartitions=2) dd.assert_eq(s.var(skipna=skipna), ds.var(skipna=skipna)) diff --git a/python/dask_cudf/pyproject.toml b/python/dask_cudf/pyproject.toml index bc8bc7b6f59e..bd0789370185 100644 --- a/python/dask_cudf/pyproject.toml +++ b/python/dask_cudf/pyproject.toml @@ -59,7 +59,7 @@ test = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" [tool.rapids-build-backend] build-backend = "setuptools.build_meta" diff --git a/python/libcudf/pyproject.toml b/python/libcudf/pyproject.toml index 875f7a1516aa..947c9fe608a8 100644 --- a/python/libcudf/pyproject.toml +++ b/python/libcudf/pyproject.toml @@ -34,7 +34,7 @@ dependencies = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" [project.entry-points."cmake.prefix"] libcudf = "libcudf" diff --git a/python/libcudf_streaming/pyproject.toml b/python/libcudf_streaming/pyproject.toml index 7b09f36258f6..06492b4e8d69 100644 --- a/python/libcudf_streaming/pyproject.toml +++ b/python/libcudf_streaming/pyproject.toml @@ -32,7 +32,7 @@ dependencies = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" [project.entry-points."cmake.prefix"] libcudf_streaming = "libcudf_streaming" diff --git a/python/pylibcudf/pylibcudf/column.pyx b/python/pylibcudf/pylibcudf/column.pyx index f7f2a59925e0..0a156b0afb7f 100644 --- a/python/pylibcudf/pylibcudf/column.pyx +++ b/python/pylibcudf/pylibcudf/column.pyx @@ -1316,7 +1316,7 @@ cdef class Column: and reference_child.size() == child.size() and reference_child.null_count() == child.null_count() # We assume the null masks are equivalent but may be expensive to - # check: https://github.com/rapidsai/cudf/pull/19357#issuecomment-3071033448 + # check: https://github.com/NVIDIA/cudf/pull/19357#issuecomment-3071033448 for child in children ): raise ValueError( diff --git a/python/pylibcudf/pylibcudf/interop.pyx b/python/pylibcudf/pylibcudf/interop.pyx index b31792f13b2f..8614fa650aa8 100644 --- a/python/pylibcudf/pylibcudf/interop.pyx +++ b/python/pylibcudf/pylibcudf/interop.pyx @@ -72,8 +72,8 @@ cpdef Table from_dlpack( # Note: A copy is always performed when converting the dlpack # data to a libcudf table. We also delete the dlpack_tensor pointer # as the pointer is not deleted by libcudf's from_dlpack function. - # TODO: https://github.com/rapidsai/cudf/issues/10874 - # TODO: https://github.com/rapidsai/cudf/issues/10849 + # TODO: https://github.com/NVIDIA/cudf/issues/10874 + # TODO: https://github.com/NVIDIA/cudf/issues/10849 with nogil: c_result = cpp_from_dlpack(dlpack_tensor, _cs, mr.get_mr()) diff --git a/python/pylibcudf/pylibcudf/scalar.pyx b/python/pylibcudf/pylibcudf/scalar.pyx index 1b80a62bed96..88cfb1924f98 100644 --- a/python/pylibcudf/pylibcudf/scalar.pyx +++ b/python/pylibcudf/pylibcudf/scalar.pyx @@ -99,7 +99,7 @@ __all__ = ["Scalar"] # the best we can do is to grab the current memory resource at the time of # construction and keep it alive until the Scalar is destroyed (for potential # problems with this approach, see https://github.com/rapidsai/rmm/issues/1515; -# the solution will be to address https://github.com/rapidsai/cudf/issues/15170 +# the solution will be to address https://github.com/NVIDIA/cudf/issues/15170 # and also pass mrs all the way down to every rmm Python API to avoid its # default mrs). This is done in the `__cinit__` method below. # diff --git a/python/pylibcudf/pylibcudf/strings/capitalize.pyx b/python/pylibcudf/pylibcudf/strings/capitalize.pyx index 2c5683980191..b8190f79c1ca 100644 --- a/python/pylibcudf/pylibcudf/strings/capitalize.pyx +++ b/python/pylibcudf/pylibcudf/strings/capitalize.pyx @@ -28,7 +28,7 @@ cpdef Column capitalize( object stream=None, DeviceMemoryResource mr=None, # TODO: default scalar values - # https://github.com/rapidsai/cudf/issues/15505 + # https://github.com/NVIDIA/cudf/issues/15505 ): """Returns a column of capitalized strings. diff --git a/python/pylibcudf/pylibcudf/strings/replace.pyx b/python/pylibcudf/pylibcudf/strings/replace.pyx index 5603ac849f0c..25f8eb45b208 100644 --- a/python/pylibcudf/pylibcudf/strings/replace.pyx +++ b/python/pylibcudf/pylibcudf/strings/replace.pyx @@ -133,7 +133,7 @@ cpdef Column replace_multiple( cpdef Column replace_slice( Column input, # TODO: default scalar values - # https://github.com/rapidsai/cudf/issues/15505 + # https://github.com/NVIDIA/cudf/issues/15505 Scalar repl=None, size_type start=0, size_type stop=-1, diff --git a/python/pylibcudf/pyproject.toml b/python/pylibcudf/pyproject.toml index ca8f1ce9a04a..3a6f0dfe2745 100644 --- a/python/pylibcudf/pyproject.toml +++ b/python/pylibcudf/pyproject.toml @@ -58,7 +58,7 @@ numpy = [ ] # This list was generated by `rapids-dependency-file-generator`. To make changes, edit ../../dependencies.yaml and run `rapids-dependency-file-generator`. [project.urls] -Homepage = "https://github.com/rapidsai/cudf" +Homepage = "https://github.com/NVIDIA/cudf" Documentation = "https://docs.rapids.ai/api/cudf/stable/" [tool.ruff] diff --git a/python/pylibcudf/tests/io/test_parquet.py b/python/pylibcudf/tests/io/test_parquet.py index be56621c2bc9..c659d65b8050 100644 --- a/python/pylibcudf/tests/io/test_parquet.py +++ b/python/pylibcudf/tests/io/test_parquet.py @@ -617,7 +617,7 @@ def test_write_parquet( ): _, pa_table = table_data if len(pa_table) == 0 and partitions is not None: - pytest.skip("https://github.com/rapidsai/cudf/issues/17361") + pytest.skip("https://github.com/NVIDIA/cudf/issues/17361") plc_table = plc.Table.from_arrow(pa_table) table_meta = plc.io.types.TableInputMetadata(plc_table) sink = plc.io.SinkInfo([io.BytesIO()]) diff --git a/skills/accelerated-computing-cudf/SKILL.md b/skills/accelerated-computing-cudf/SKILL.md index 41fcff67caaf..26ddef528e09 100644 --- a/skills/accelerated-computing-cudf/SKILL.md +++ b/skills/accelerated-computing-cudf/SKILL.md @@ -199,5 +199,5 @@ Use WebFetch to retrieve detailed API signatures, parameter descriptions, and ex - **cuDF Documentation:** https://docs.rapids.ai/api/cudf/stable/ - **dask-cuDF API Reference:** https://docs.rapids.ai/api/dask-cudf/stable/api/ -- **GitHub:** https://github.com/rapidsai/cudf -- **CHANGELOG:** https://github.com/rapidsai/cudf/blob/main/CHANGELOG.md +- **GitHub:** https://github.com/NVIDIA/cudf +- **CHANGELOG:** https://github.com/NVIDIA/cudf/blob/main/CHANGELOG.md diff --git a/skills/accelerated-computing-cudf/evals/files/source-cudf-null-fillna-semantics/NOTICE.md b/skills/accelerated-computing-cudf/evals/files/source-cudf-null-fillna-semantics/NOTICE.md index bf77166c3d9b..791570030a53 100644 --- a/skills/accelerated-computing-cudf/evals/files/source-cudf-null-fillna-semantics/NOTICE.md +++ b/skills/accelerated-computing-cudf/evals/files/source-cudf-null-fillna-semantics/NOTICE.md @@ -2,7 +2,7 @@ This task is source-inspired by cuDF null-handling tests. -- Source: https://github.com/rapidsai/cudf/blob/235f69a6fcef/python/cudf/cudf/tests/dataframe/methods/test_fillna.py +- Source: https://github.com/NVIDIA/cudf/blob/235f69a6fcef/python/cudf/cudf/tests/dataframe/methods/test_fillna.py - Upstream project: RAPIDS cuDF - License: Apache-2.0 - Local changes: original pandas fixture written for benchmark scoring; no upstream code copied. diff --git a/skills/accelerated-computing-cudf/skill-card.md b/skills/accelerated-computing-cudf/skill-card.md index 7097caaa744a..0f5c72d23ad0 100644 --- a/skills/accelerated-computing-cudf/skill-card.md +++ b/skills/accelerated-computing-cudf/skill-card.md @@ -24,7 +24,7 @@ Mitigation: Review and scan skill before deployment.
- [dask-cuDF Patterns](references/dask-cudf-patterns.md)
- [cuDF Documentation](https://docs.rapids.ai/api/cudf/stable/)
- [dask-cuDF API Reference](https://docs.rapids.ai/api/dask-cudf/stable/api/)
-- [cuDF GitHub Repository](https://github.com/rapidsai/cudf)
+- [cuDF GitHub Repository](https://github.com/NVIDIA/cudf)
## Skill Output: