[ArrowStringArray] PERF: str.extract object fallback

asv_bench/benchmarks/strings.py

@@ -248,6 +248,24 @@ def time_rsplit(self, dtype, expand):
         self.s.str.rsplit("--", expand=expand)
 
 
+class Extract:
+
+    params = (["str", "string", "arrow_string"], [True, False])
+    param_names = ["dtype", "expand"]
+
+    def setup(self, dtype, expand):
+        from pandas.core.arrays.string_arrow import ArrowStringDtype  # noqa: F401
+
+        try:
+            self.s = Series(tm.makeStringIndex(10 ** 5), dtype=dtype)
+        except ImportError:
+            raise NotImplementedError
+
+    def time_extract_single_group(self, dtype, expand):
+        with warnings.catch_warnings(record=True):
+            self.s.str.extract("(\\w*)A", expand=expand)
+
+
 class Dummies:
     params = ["str", "string", "arrow_string"]
     param_names = ["dtype"]
@@ -279,3 +297,20 @@ def setup(self):
     def time_vector_slice(self):
         # GH 2602
         self.s.str[:5]
+
+
+class Iter:
+    params = ["str", "string", "arrow_string"]
+    param_names = ["dtype"]
+
+    def setup(self, dtype):
+        from pandas.core.arrays.string_arrow import ArrowStringDtype  # noqa: F401
+
+        try:
+            self.s = Series(["abcdefg", np.nan] * 500000, dtype=dtype)
+        except ImportError:
+            raise NotImplementedError
+
+    def time_iter(self, dtype):
+        for i in self.s:
+            pass

pandas/core/strings/accessor.py

@@ -3102,7 +3102,9 @@ def _str_extract_noexpand(arr, pat, flags=0):
     result_dtype = _result_dtype(arr)
 
     if regex.groups == 1:
-        result = np.array([groups_or_na(val)[0] for val in arr], dtype=object)
+        result = np.array(
+            [groups_or_na(val)[0] for val in np.asarray(arr)], dtype=object
+        )
         name = _get_single_group_name(regex)
         # not dispatching, so we have to reconstruct here.
         result = pd_array(result, dtype=result_dtype)
@@ -3119,7 +3121,7 @@ def _str_extract_noexpand(arr, pat, flags=0):
             # error: Incompatible types in assignment (expression has type
             # "DataFrame", variable has type "ndarray")
             result = DataFrame(  # type:ignore[assignment]
-                [groups_or_na(val) for val in arr],
+                [groups_or_na(val) for val in np.asarray(arr)],
                 columns=columns,
                 index=arr.index,
                 dtype=result_dtype,
@@ -3148,7 +3150,7 @@ def _str_extract_frame(arr, pat, flags=0):
     except AttributeError:
         result_index = None
     return DataFrame(
-        [groups_or_na(val) for val in arr],
+        [groups_or_na(val) for val in np.asarray(arr)],
         columns=columns,
         index=result_index,
         dtype=result_dtype,

pandas/tests/strings/test_extract.py

@@ -38,25 +38,23 @@ def test_extract_expand_kwarg(any_string_dtype):
 
 
 def test_extract_expand_False_mixed_object():
-    mixed = Series(
-        [
-            "aBAD_BAD",
-            np.nan,
-            "BAD_b_BAD",
-            True,
-            datetime.today(),
-            "foo",
-            None,
-            1,
-            2.0,
-        ]
+    ser = Series(
+        ["aBAD_BAD", np.nan, "BAD_b_BAD", True, datetime.today(), "foo", None, 1, 2.0]
     )
 
-    result = Series(mixed).str.extract(".*(BAD[_]+).*(BAD)", expand=False)
+    # two groups
+    result = ser.str.extract(".*(BAD[_]+).*(BAD)", expand=False)
     er = [np.nan, np.nan]  # empty row
     expected = DataFrame([["BAD_", "BAD"], er, ["BAD_", "BAD"], er, er, er, er, er, er])
     tm.assert_frame_equal(result, expected)
 
+    # single group
+    result = ser.str.extract(".*(BAD[_]+).*BAD", expand=False)
+    expected = Series(
+        ["BAD_", np.nan, "BAD_", np.nan, np.nan, np.nan, np.nan, np.nan, np.nan]
+    )
+    tm.assert_series_equal(result, expected)
+
 
 def test_extract_expand_index_raises():
     # GH9980