Repository navigation
DRAFT Bench 25767 - #26368
Draft
kosiew wants to merge 7 commits into
Draft
DRAFT Bench 25767#26368kosiew wants to merge 7 commits into
kosiew wants to merge 7 commits into
Conversation
…harness - Added benchmarks/sql_benchmarks/set_operations/ using shared Criterion harness
Signed-off-by: Stefan Wang <1fannnw@gmail.com>
Signed-off-by: Stefan Wang <1fannnw@gmail.com>
Both sides used the same name for their synthetic row number column, so the sides were always requalified as `left` and `right`, and the final projection rebuilt every output column as `left.<name>`. Queries that referred to the original qualifier after the set operation failed, and two columns with the same name from different relations on one side collapsed into duplicate `left.x` outputs. Give each side its own row number name so requalification only happens when the user's columns conflict, join on the columns the requalified schemas actually contain, and alias the output back to the left input's qualifiers and names. Signed-off-by: Stefan Wang <1fannnw@gmail.com>
Document that `LogicalPlanBuilder::intersect` and `except` with `is_all = true` do not preserve duplicate counts and point callers at `intersect_all` and `except_all`. Name the operation in the planning error raised when `row_number` is not registered, and add an upgrade guide entry for custom context providers and session states. The Substrait `intersect_rel` helper is now only used for distinct intersections, so drop its `is_all` parameter. Signed-off-by: Stefan Wang <1fannnw@gmail.com>
Add sqllogictest cases for min and max duplicate counts with NULLs, rows that are NULL in only some columns, INTERSECT ALL binding tighter than EXCEPT ALL, and the two qualifier regressions. Record the +0.0 / -0.0 case as a known issue: window partitioning puts the two values in separate partitions, so the counts are wrong until apache#25766 is fixed. Signed-off-by: Stefan Wang <1fannnw@gmail.com>
Count each distinct row on both sides, join the counts, and repeat each row min(m, n) or m - n times with unnest(range(..)). This replaces the per-row row_number window, which was slow when most rows are distinct. Grouping also treats 0.0 and -0.0 as equal, so those counts are now right. Signed-off-by: Stefan Wang <1fannnw@gmail.com>
github-actions
Bot
added
documentation
Contributor
Author
|
run benchmark sql env:
BENCH_NAME: set_operations
PARTITIONS: 4
baseline:
ref: 2cba9d87ee |
Contributor
Author
|
show benchmark queue |
|
Hi @kosiew, you asked to view the benchmark queue (#26368 (comment)).
File an issue against this benchmark runner |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing bench-25767 (5a8f209) to 2cba9d8 diff Run configurationrun benchmark sql
env:
BENCH_NAME: "set_operations"
PARTITIONS: "4"
baseline:
ref: "2cba9d87ee"Results will be posted here when complete File an issue against this benchmark runner |
Codecov Report❌ Patch coverage is Additional details and impacted files@@ Coverage Diff @@
## main #26368 +/- ##
========================================
Coverage 82.84% 82.85%
========================================
Files 1151 1151
Lines 455961 456143 +182
Branches 455961 456143 +182
========================================
+ Hits 377760 377929 +169
+ Misses 55245 55244 -1
- Partials 22956 22970 +14 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
|
🤖 Benchmark completed (GKE) | trigger Instance: Comparing bench-25767 (5a8f209) to 2cba9d8 diff Run configurationrun benchmark sql
env:
BENCH_NAME: "set_operations"
PARTITIONS: "4"
baseline:
ref: "2cba9d87ee"CPU Details (lscpu)Details
Resource Usagesql — branch
File an issue against this benchmark runner |
Contributor
Author
|
run benchmark sql env:
BENCH_NAME: set_operations
PARTITIONS: 4
baseline:
ref: fa66d51c1b |
|
🤖 Benchmark running (GKE) | trigger CPU Details (lscpu)Comparing bench-25767 (5a8f209) to fa66d51 diff Run configurationrun benchmark sql
env:
BENCH_NAME: "set_operations"
PARTITIONS: "4"
baseline:
ref: "fa66d51c1b"Results will be posted here when complete File an issue against this benchmark runner |
|
🤖 Benchmark completed (GKE) | trigger Instance: Comparing bench-25767 (5a8f209) to fa66d51 diff Run configurationrun benchmark sql
env:
BENCH_NAME: "set_operations"
PARTITIONS: "4"
baseline:
ref: "fa66d51c1b"CPU Details (lscpu)Details
Resource Usagesql — branch
File an issue against this benchmark runner |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
rebased
benchmarking 25767