ENH: Series.str.join for ArrowDtype(pa.string()) (pandas-dev#53646)

* Series.str.join to support ArrowDtype(pa.string())

* gh ref

Author: lukemanley

doc/source/whatsnew/v2.1.0.rst

@@ -102,6 +102,7 @@ Other enhancements
 - :meth:`DataFrame.stack` gained the ``sort`` keyword to dictate whether the resulting :class:`MultiIndex` levels are sorted (:issue:`15105`)
 - :meth:`DataFrame.unstack` gained the ``sort`` keyword to dictate whether the resulting :class:`MultiIndex` levels are sorted (:issue:`15105`)
 - :meth:`DataFrameGroupby.agg` and :meth:`DataFrameGroupby.transform` now support grouping by multiple keys when the index is not a :class:`MultiIndex` for ``engine="numba"`` (:issue:`53486`)
+- :meth:`Series.str.join` now supports ``ArrowDtype(pa.string())`` (:issue:`53646`)
 - :meth:`SeriesGroupby.agg` and :meth:`DataFrameGroupby.agg` now support passing in multiple functions for ``engine="numba"`` (:issue:`53486`)
 - :meth:`SeriesGroupby.transform` and :meth:`DataFrameGroupby.transform` now support passing in a string as the function for ``engine="numba"`` (:issue:`53579`)
 - Added ``engine_kwargs`` parameter to :meth:`DataFrame.to_excel` (:issue:`53220`)

pandas/core/arrays/arrow/array.py

@@ -2073,7 +2073,12 @@ def _str_get(self, i: int):
         return type(self)(result)
 
     def _str_join(self, sep: str):
-        return type(self)(pc.binary_join(self._pa_array, sep))
+        if pa.types.is_string(self._pa_array.type):
+            result = self._apply_elementwise(list)
+            result = pa.chunked_array(result, type=pa.list_(pa.string()))
+        else:
+            result = self._pa_array
+        return type(self)(pc.binary_join(result, sep))
 
     def _str_partition(self, sep: str, expand: bool):
         predicate = lambda val: val.partition(sep)

pandas/tests/extension/test_arrow.py

@@ -2026,6 +2026,13 @@ def test_str_join():
     tm.assert_series_equal(result, expected)
 
 
+def test_str_join_string_type():
+    ser = pd.Series(ArrowExtensionArray(pa.array(["abc", "123", None])))
+    result = ser.str.join("=")
+    expected = pd.Series(["a=b=c", "1=2=3", None], dtype=ArrowDtype(pa.string()))
+    tm.assert_series_equal(result, expected)
+
+
 @pytest.mark.parametrize(
     "start, stop, step, exp",
     [