Backport PR #55348 on branch 2.1.x (REGR: join segfaulting for arrow string with nulls) (#55357)

Backport PR #55348: REGR: join segfaulting for arrow string with nulls

Co-authored-by: Patrick Hoefler <[email protected]>

Author: meeseeksmachine

doc/source/whatsnew/v2.1.2.rst

@@ -14,7 +14,7 @@ including other versions of pandas.
 Fixed regressions
 ~~~~~~~~~~~~~~~~~
 - Fixed bug where PDEP-6 warning about setting an item of an incompatible dtype was being shown when creating a new conditional column (:issue:`55025`)
--
+- Fixed regression in :meth:`DataFrame.join` where result has missing values and dtype is arrow backed string (:issue:`55348`)
 
 .. ---------------------------------------------------------------------------
 .. _whatsnew_212.bug_fixes:

pandas/core/reshape/merge.py

@@ -2442,6 +2442,8 @@ def _factorize_keys(
                 .astype(np.intp, copy=False),
                 len(dc.dictionary),
             )
+            if dc.null_count > 0:
+                count += 1
             if how == "right":
                 return rlab, llab, count
             return llab, rlab, count

pandas/tests/frame/methods/test_join.py

@@ -158,9 +158,14 @@ def test_join_invalid_validate(left_no_dup, right_no_dup):
         left_no_dup.merge(right_no_dup, on="a", validate="invalid")
 
 
-def test_join_on_single_col_dup_on_right(left_no_dup, right_w_dups):
+@pytest.mark.parametrize("dtype", ["object", "string[pyarrow]"])
+def test_join_on_single_col_dup_on_right(left_no_dup, right_w_dups, dtype):
     # GH 46622
     # Dups on right allowed by one_to_many constraint
+    if dtype == "string[pyarrow]":
+        pytest.importorskip("pyarrow")
+    left_no_dup = left_no_dup.astype(dtype)
+    right_w_dups.index = right_w_dups.index.astype(dtype)
     left_no_dup.join(
         right_w_dups,
         on="a",