BUG: resample with ArrowDtype (#56371)

* BUG: resample with ArrowDtype

* Typing

* xfail for windows

* Fix again?

* Avoid tuple

* Add gh numbers

doc/source/whatsnew/v2.2.0.rst

@@ -648,6 +648,7 @@ Groupby/resample/rolling
 - Bug in :meth:`DataFrame.asfreq` and :meth:`Series.asfreq` with a :class:`DatetimeIndex` with non-nanosecond resolution incorrectly converting to nanosecond resolution (:issue:`55958`)
 - Bug in :meth:`DataFrame.ewm` when passed ``times`` with non-nanosecond ``datetime64`` or :class:`DatetimeTZDtype` dtype (:issue:`56262`)
 - Bug in :meth:`DataFrame.resample` not respecting ``closed`` and ``label`` arguments for :class:`~pandas.tseries.offsets.BusinessDay` (:issue:`55282`)
+- Bug in :meth:`DataFrame.resample` when resampling on a :class:`ArrowDtype` of ``pyarrow.timestamp`` or ``pyarrow.duration`` type (:issue:`55989`)
 - Bug in :meth:`DataFrame.resample` where bin edges were not correct for :class:`~pandas.tseries.offsets.BusinessDay` (:issue:`55281`)
 - Bug in :meth:`DataFrame.resample` where bin edges were not correct for :class:`~pandas.tseries.offsets.MonthBegin` (:issue:`55271`)
 -

pandas/core/groupby/grouper.py

@@ -330,7 +330,6 @@ def _get_grouper(
 
         return grouper, obj
 
-    @final
     def _set_grouper(
         self, obj: NDFrameT, sort: bool = False, *, gpr_index: Index | None = None
     ) -> tuple[NDFrameT, Index, npt.NDArray[np.intp] | None]:

pandas/core/resample.py

@@ -38,6 +38,7 @@
     rewrite_warning,
 )
 
+from pandas.core.dtypes.dtypes import ArrowDtype
 from pandas.core.dtypes.generic import (
     ABCDataFrame,
     ABCSeries,
@@ -48,6 +49,7 @@
     ResamplerWindowApply,
     warn_alias_replacement,
 )
+from pandas.core.arrays import ArrowExtensionArray
 from pandas.core.base import (
     PandasObject,
     SelectionMixin,
@@ -68,6 +70,7 @@
 from pandas.core.groupby.grouper import Grouper
 from pandas.core.groupby.ops import BinGrouper
 from pandas.core.indexes.api import MultiIndex
+from pandas.core.indexes.base import Index
 from pandas.core.indexes.datetimes import (
     DatetimeIndex,
     date_range,
@@ -109,7 +112,6 @@
 
     from pandas import (
         DataFrame,
-        Index,
         Series,
     )
 
@@ -511,6 +513,9 @@ def _wrap_result(self, result):
             result.index = _asfreq_compat(obj.index[:0], freq=self.freq)
             result.name = getattr(obj, "name", None)
 
+        if self._timegrouper._arrow_dtype is not None:
+            result.index = result.index.astype(self._timegrouper._arrow_dtype)
+
         return result
 
     @final
@@ -2163,6 +2168,7 @@ def __init__(
         self.fill_method = fill_method
         self.limit = limit
         self.group_keys = group_keys
+        self._arrow_dtype: ArrowDtype | None = None
 
         if origin in ("epoch", "start", "start_day", "end", "end_day"):
             # error: Incompatible types in assignment (expression has type "Union[Union[
@@ -2213,7 +2219,7 @@ def _get_resampler(self, obj: NDFrame, kind=None) -> Resampler:
         TypeError if incompatible axis
 
         """
-        _, ax, indexer = self._set_grouper(obj, gpr_index=None)
+        _, ax, _ = self._set_grouper(obj, gpr_index=None)
         if isinstance(ax, DatetimeIndex):
             return DatetimeIndexResampler(
                 obj,
@@ -2495,6 +2501,17 @@ def _get_period_bins(self, ax: PeriodIndex):
 
         return binner, bins, labels
 
+    def _set_grouper(
+        self, obj: NDFrameT, sort: bool = False, *, gpr_index: Index | None = None
+    ) -> tuple[NDFrameT, Index, npt.NDArray[np.intp] | None]:
+        obj, ax, indexer = super()._set_grouper(obj, sort, gpr_index=gpr_index)
+        if isinstance(ax.dtype, ArrowDtype) and ax.dtype.kind in "Mm":
+            self._arrow_dtype = ax.dtype
+            ax = Index(
+                cast(ArrowExtensionArray, ax.array)._maybe_convert_datelike_array()
+            )
+        return obj, ax, indexer
+
 
 def _take_new_index(
     obj: NDFrameT, indexer: npt.NDArray[np.intp], new_index: Index, axis: AxisInt = 0

pandas/tests/resample/test_datetime_index.py

@@ -7,6 +7,8 @@
 
 from pandas._libs import lib
 from pandas._typing import DatetimeNaTType
+from pandas.compat import is_platform_windows
+import pandas.util._test_decorators as td
 
 import pandas as pd
 from pandas import (
@@ -2195,3 +2197,27 @@ def test_resample_b_55282(unit):
         index=exp_dti,
     )
     tm.assert_series_equal(result, expected)
+
+
+@td.skip_if_no("pyarrow")
+@pytest.mark.parametrize(
+    "tz",
+    [
+        None,
+        pytest.param(
+            "UTC",
+            marks=pytest.mark.xfail(
+                condition=is_platform_windows(),
+                reason="TODO: Set ARROW_TIMEZONE_DATABASE env var in CI",
+            ),
+        ),
+    ],
+)
+def test_arrow_timestamp_resample(tz):
+    # GH 56371
+    idx = Series(date_range("2020-01-01", periods=5), dtype="timestamp[ns][pyarrow]")
+    if tz is not None:
+        idx = idx.dt.tz_localize(tz)
+    expected = Series(np.arange(5, dtype=np.float64), index=idx)
+    result = expected.resample("1D").mean()
+    tm.assert_series_equal(result, expected)

pandas/tests/resample/test_timedelta.py

@@ -3,6 +3,8 @@
 import numpy as np
 import pytest
 
+import pandas.util._test_decorators as td
+
 import pandas as pd
 from pandas import (
     DataFrame,
@@ -207,3 +209,12 @@ def test_resample_closed_right():
         ),
     )
     tm.assert_series_equal(result, expected)
+
+
+@td.skip_if_no("pyarrow")
+def test_arrow_duration_resample():
+    # GH 56371
+    idx = pd.Index(timedelta_range("1 day", periods=5), dtype="duration[ns][pyarrow]")
+    expected = Series(np.arange(5, dtype=np.float64), index=idx)
+    result = expected.resample("1D").mean()
+    tm.assert_series_equal(result, expected)