BUG: Add PyArrow datelike type support for map()

pandas/core/arrays/arrow/array.py

@@ -1490,10 +1490,9 @@ def to_numpy(
         return result
 
     def map(self, mapper, na_action: Literal["ignore"] | None = None):
-        if is_numeric_dtype(self.dtype):
+        if self.dtype.kind in "mM" or is_numeric_dtype(self.dtype):
             return map_array(self.to_numpy(), mapper, na_action=na_action)
-        else:
-            return super().map(mapper, na_action)
+        return super().map(mapper, na_action)
 
     @doc(ExtensionArray.duplicated)
     def duplicated(

pandas/tests/extension/test_arrow.py

@@ -64,6 +64,7 @@
     is_string_dtype,
     is_unsigned_integer_dtype,
 )
+from pandas.core.algorithms import map_array
 from pandas.tests.extension import base
 
 pa = pytest.importorskip("pyarrow")
@@ -278,8 +279,9 @@ def test_compare_scalar(self, data, comparison_op):
     @pytest.mark.parametrize("na_action", [None, "ignore"])
     def test_map(self, data_missing, na_action):
         if data_missing.dtype.kind in "mM":
-            result = data_missing.map(lambda x: x, na_action=na_action)
-            expected = data_missing.to_numpy(dtype=object)
+            mapper = lambda x: x
+            result = data_missing.map(mapper, na_action=na_action)
+            expected = map_array(data_missing.to_numpy(), mapper, na_action=na_action)
             tm.assert_numpy_array_equal(result, expected)
         else:
             result = data_missing.map(lambda x: x, na_action=na_action)