From ffec56615d7e1170b13ecc16e54ccbb04d7731ae Mon Sep 17 00:00:00 2001
From: samukweku <samueloranyeli@gmail.com>
Date: Wed, 26 Aug 2026 09:34:28 +1000
Subject: [PATCH] perf: update min reverse starts-match boundary

---
 AGENTS.md                                             | 11 +++++++++++
 janitor/functions/_conditional_join/_agg_functions.py |  2 --
 janitor/functions/_conditional_join/_get_join_aggs.py |  6 ++++--
 3 files changed, 15 insertions(+), 4 deletions(-)

diff --git a/AGENTS.md b/AGENTS.md
index 34a7f1921..fdbed46f8 100644
--- a/AGENTS.md
+++ b/AGENTS.md
@@ -548,6 +548,17 @@ CLI.
 include MkDocs. The documentation task is available in the `docs` environment.
 **Recommendation**: Run `pixi run -e docs build-docs` to build documentation.
 
+### [2026-08-26] Include performance evidence in optimization PRs and issues
+
+**Context**: Performance changes are coordinated with implementation changes
+in janitor-rs.
+**Learning**: Benchmark results are part of the performance change's review
+record, not merely local investigation notes.
+**Recommendation**: Every performance PR and its tracking issue must include a
+properly formatted comparison with the old implementation, covering runtime
+and memory for the agreed tiny, large, very-large, and super-large cases and
+duplicate/unique label distributions. State benchmark limitations explicitly.
+
 ---
 
 ## Version History
diff --git a/janitor/functions/_conditional_join/_agg_functions.py b/janitor/functions/_conditional_join/_agg_functions.py
index 96fa0b7a2..5687c6cee 100644
--- a/janitor/functions/_conditional_join/_agg_functions.py
+++ b/janitor/functions/_conditional_join/_agg_functions.py
@@ -1356,7 +1356,6 @@ def _min_rev_starts_matches(
     index: np.ndarray,
     matches: np.ndarray,
     booleans: np.ndarray,
-    length: int,
 ) -> tuple:
     """
     Compute min
@@ -1385,7 +1384,6 @@ def _min_rev_starts_matches(
         index=index,
         matches=matches,
         booleans=booleans,
-        length=length,
     )
 
 
diff --git a/janitor/functions/_conditional_join/_get_join_aggs.py b/janitor/functions/_conditional_join/_get_join_aggs.py
index d757c6f41..c97b5487d 100644
--- a/janitor/functions/_conditional_join/_get_join_aggs.py
+++ b/janitor/functions/_conditional_join/_get_join_aggs.py
@@ -180,15 +180,17 @@ def _agg_join_left(df: pd.DataFrame, aggfunc: list, indices: dict) -> pd.DataFra
                 booleans = pd.isna(arr)
                 arr = _helpers._convert_array_to_numpy(array=arr)
                 func = mapping[agg]
-                _index, out = func(
+                kwargs = dict(
                     arr=arr,
                     starts=indices["starts"],
                     index=indices["right_index"],
                     matches=indices["matches"],
                     counts=indices["counts_array"],
                     booleans=booleans,
-                    length=indices["right_index"].size,
                 )
+                if agg != "min":
+                    kwargs["length"] = indices["right_index"].size
+                _index, out = func(**kwargs)
                 if agg in {
                     "sum",
                     "prod",