From 14240b0d16ecc9062c45b76397f3f3519d634e75 Mon Sep 17 00:00:00 2001
From: Florian Pfaff <6773539+FlorianPfaff@users.noreply.github.com>
Date: Mon, 13 Jul 2026 11:46:54 +0200
Subject: [PATCH 1/2] Align heterogeneous evaluation result columns

---
 .../evaluation/group_results_by_filter.py     | 21 ++++++++++++-------
 1 file changed, 14 insertions(+), 7 deletions(-)

diff --git a/src/pyrecest/evaluation/group_results_by_filter.py b/src/pyrecest/evaluation/group_results_by_filter.py
index 2b5542c057..440a34e765 100644
--- a/src/pyrecest/evaluation/group_results_by_filter.py
+++ b/src/pyrecest/evaluation/group_results_by_filter.py
@@ -20,13 +20,20 @@ def group_results_by_filter(data):
         # Remove the 'name' key-value pair from the entry
         entry_values = {k: v for k, v in entry.items() if k != "name"}
 
-        # Check if the name already exists in the output_dict
-        if name in output_dict:
-            for key, value in entry_values.items():
-                # Append values to the existing lists
-                output_dict[name][key].append(value)
-        else:
-            # Initialize the entry in the output_dict with lists for each value
+        if name not in output_dict:
             output_dict[name] = {k: [v] for k, v in entry_values.items()}
+            continue
+
+        grouped_values = output_dict[name]
+        existing_row_count = len(next(iter(grouped_values.values()), []))
+
+        # Backfill columns that first appear in a later row so every column remains aligned.
+        for key in entry_values:
+            if key not in grouped_values:
+                grouped_values[key] = [None] * existing_row_count
+
+        # Append one value per known column, using None for fields omitted by this row.
+        for key in grouped_values:
+            grouped_values[key].append(entry_values.get(key))
 
     return output_dict

From f0ac38816e429484aad94c44ceb4760e20e3c046 Mon Sep 17 00:00:00 2001
From: Florian Pfaff <6773539+FlorianPfaff@users.noreply.github.com>
Date: Mon, 13 Jul 2026 11:47:07 +0200
Subject: [PATCH 2/2] Test heterogeneous evaluation result grouping

---
 tests/test_group_results_by_filter.py | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/tests/test_group_results_by_filter.py b/tests/test_group_results_by_filter.py
index efcfee7873..cc6cc2e511 100644
--- a/tests/test_group_results_by_filter.py
+++ b/tests/test_group_results_by_filter.py
@@ -16,6 +16,19 @@ def test_mixed_parameters_do_not_crash_sorting(self):
         self.assertEqual(grouped["pf"]["parameter"], [None, 1, "b"])
         self.assertEqual(grouped["pf"]["score"], [0.0, 1.0, 3.0])
 
+    def test_heterogeneous_metric_keys_stay_aligned(self):
+        rows = [
+            {"name": "pf", "parameter": 3, "std": 0.3},
+            {"name": "pf", "parameter": 1, "score": 1.0},
+            {"name": "pf", "parameter": 2, "score": 2.0, "std": 0.2},
+        ]
+
+        grouped = group_results_by_filter(rows)
+
+        self.assertEqual(grouped["pf"]["parameter"], [1, 2, 3])
+        self.assertEqual(grouped["pf"]["score"], [1.0, 2.0, None])
+        self.assertEqual(grouped["pf"]["std"], [None, 0.2, 0.3])
+
 
 if __name__ == "__main__":
     unittest.main()