BUG: fixed Series.dt methods in ArrowDtype class that were returning incorrect time values. (pandas-dev#57355)

Author: St0rmie

doc/source/whatsnew/v3.0.0.rst

@@ -324,6 +324,7 @@ Bug fixes
 - Fixed bug in :meth:`DataFrame.update` bool dtype being converted to object (:issue:`55509`)
 - Fixed bug in :meth:`DataFrameGroupBy.apply` that was returning a completely empty DataFrame when all return values of ``func`` were ``None`` instead of returning an empty DataFrame with the original columns and dtypes. (:issue:`57775`)
 - Fixed bug in :meth:`Series.diff` allowing non-integer values for the ``periods`` argument. (:issue:`56607`)
+- Fixed bug in :meth:`Series.dt` methods in :class:`ArrowDtype` that were returning incorrect values. (:issue:`57355`)
 - Fixed bug in :meth:`Series.rank` that doesn't preserve missing values for nullable integers when ``na_option='keep'``. (:issue:`56976`)
 - Fixed bug in :meth:`Series.replace` and :meth:`DataFrame.replace` inconsistently replacing matching instances when ``regex=True`` and missing values are present. (:issue:`56599`)
 

pandas/core/arrays/arrow/array.py

@@ -2642,7 +2642,11 @@ def _dt_minutes(self) -> Self:
     def _dt_seconds(self) -> Self:
         return type(self)(
             pa.array(
-                self._to_timedeltaarray().seconds, from_pandas=True, type=pa.int32()
+                [
+                    td.components.seconds if td is not NaT else None
+                    for td in self._to_timedeltaarray()
+                ],
+                type=pa.int32(),
             )
         )
 
@@ -2662,8 +2666,10 @@ def _dt_milliseconds(self) -> Self:
     def _dt_microseconds(self) -> Self:
         return type(self)(
             pa.array(
-                self._to_timedeltaarray().microseconds,
-                from_pandas=True,
+                [
+                    td.components.microseconds if td is not NaT else None
+                    for td in self._to_timedeltaarray()
+                ],
                 type=pa.int32(),
             )
         )
@@ -2672,7 +2678,11 @@ def _dt_microseconds(self) -> Self:
     def _dt_nanoseconds(self) -> Self:
         return type(self)(
             pa.array(
-                self._to_timedeltaarray().nanoseconds, from_pandas=True, type=pa.int32()
+                [
+                    td.components.nanoseconds if td is not NaT else None
+                    for td in self._to_timedeltaarray()
+                ],
+                type=pa.int32(),
             )
         )
 

pandas/tests/extension/test_arrow.py

@@ -2901,6 +2901,37 @@ def test_dt_components():
     tm.assert_frame_equal(result, expected)
 
 
+def test_duration_dt_components_overflow():
+    start_dates = pd.Series(
+        [
+            "2015-03-08",
+        ],
+        dtype="datetime64[ns]",
+    ).astype("timestamp[ns][pyarrow]")
+
+    end_dates = pd.Series(
+        [
+            "2020-05-28 23:59:59.999",
+        ],
+        dtype="datetime64[ns]",
+    ).astype("timestamp[ns][pyarrow]")
+
+    result = end_dates - start_dates
+    expected_days = pd.Series([1908]).astype("int32[pyarrow]").rename("days")
+    expected_hours = pd.Series([23]).astype("int32[pyarrow]").rename("hours")
+    expected_minutes = pd.Series([59]).astype("int32[pyarrow]").rename("minutes")
+    expected_seconds = pd.Series([59]).astype("int32[pyarrow]").rename("seconds")
+    expected_milliseconds = (
+        pd.Series([999]).astype("int32[pyarrow]").rename("milliseconds")
+    )
+
+    tm.assert_series_equal(result.dt.components.days, expected_days)
+    tm.assert_series_equal(result.dt.components.hours, expected_hours)
+    tm.assert_series_equal(result.dt.components.minutes, expected_minutes)
+    tm.assert_series_equal(result.dt.components.seconds, expected_seconds)
+    tm.assert_series_equal(result.dt.components.milliseconds, expected_milliseconds)
+
+
 @pytest.mark.parametrize("skipna", [True, False])
 def test_boolean_reduce_series_all_null(all_boolean_reductions, skipna):
     # GH51624