Backport PR #56332: BUG: str.split for ArrowDtype with pat=None

doc/source/whatsnew/v2.1.4.rst

@@ -25,10 +25,12 @@ Bug fixes
 - Bug in :meth:`DataFrame.apply` where passing ``raw=True`` ignored ``args`` passed to the applied function (:issue:`55753`)
 - Bug in :meth:`Index.__getitem__` returning wrong result for Arrow dtypes and negative stepsize (:issue:`55832`)
 - Fixed bug in :func:`to_numeric` converting to extension dtype for ``string[pyarrow_numpy]`` dtype (:issue:`56179`)
-- Fixed bug in :meth:`.DataFrameGroupBy.min()` and :meth:`.DataFrameGroupBy.max()` not preserving extension dtype for empty object (:issue:`55619`)
+- Fixed bug in :meth:`.DataFrameGroupBy.min` and :meth:`.DataFrameGroupBy.max` not preserving extension dtype for empty object (:issue:`55619`)
 - Fixed bug in :meth:`DataFrame.__setitem__` casting :class:`Index` with object-dtype to PyArrow backed strings when ``infer_string`` option is set (:issue:`55638`)
 - Fixed bug in :meth:`DataFrame.to_hdf` raising when columns have ``StringDtype`` (:issue:`55088`)
 - Fixed bug in :meth:`Index.insert` casting object-dtype to PyArrow backed strings when ``infer_string`` option is set (:issue:`55638`)
+- Fixed bug in :meth:`Series.mode` not keeping object dtype when ``infer_string`` is set (:issue:`56183`)
+- Fixed bug in :meth:`Series.str.split` and :meth:`Series.str.rsplit` when ``pat=None`` for :class:`ArrowDtype` with ``pyarrow.string`` (:issue:`56271`)
 - Fixed bug in :meth:`Series.str.translate` losing object dtype when string option is set (:issue:`56152`)
 -
 

pandas/core/arrays/arrow/array.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import functools
 import operator
 import re
 import textwrap
@@ -2320,18 +2321,25 @@ def _str_split(
     ):
         if n in {-1, 0}:
             n = None
-        if regex:
-            split_func = pc.split_pattern_regex
+        if pat is None:
+            split_func = pc.utf8_split_whitespace
+        elif regex:
+            split_func = functools.partial(pc.split_pattern_regex, pattern=pat)
         else:
-            split_func = pc.split_pattern
-        return type(self)(split_func(self._pa_array, pat, max_splits=n))
+            split_func = functools.partial(pc.split_pattern, pattern=pat)
+        return type(self)(split_func(self._pa_array, max_splits=n))
 
     def _str_rsplit(self, pat: str | None = None, n: int | None = -1):
         if n in {-1, 0}:
             n = None
-        return type(self)(
-            pc.split_pattern(self._pa_array, pat, max_splits=n, reverse=True)
-        )
+        if pat is None:
+            return type(self)(
+                pc.utf8_split_whitespace(self._pa_array, max_splits=n, reverse=True)
+            )
+        else:
+            return type(self)(
+                pc.split_pattern(self._pa_array, pat, max_splits=n, reverse=True)
+            )
 
     def _str_translate(self, table: dict[int, str]):
         predicate = lambda val: val.translate(table)

pandas/tests/extension/test_arrow.py

@@ -2202,6 +2202,15 @@ def test_str_partition():
     tm.assert_series_equal(result, expected)
 
 
+@pytest.mark.parametrize("method", ["rsplit", "split"])
+def test_str_split_pat_none(method):
+    # GH 56271
+    ser = pd.Series(["a1 cbc\nb", None], dtype=ArrowDtype(pa.string()))
+    result = getattr(ser.str, method)()
+    expected = pd.Series(ArrowExtensionArray(pa.array([["a1", "cbc", "b"], None])))
+    tm.assert_series_equal(result, expected)
+
+
 def test_str_split():
     # GH 52401
     ser = pd.Series(["a1cbcb", "a2cbcb", None], dtype=ArrowDtype(pa.string()))