[pull] main from pydata:main

asv_bench/benchmarks/pandas.py

@@ -13,7 +13,7 @@ def setup(self, dtype, subset):
             [
                 list("abcdefhijk"),
                 list("abcdefhijk"),
-                pd.date_range(start="2000-01-01", periods=1000, freq="B"),
+                pd.date_range(start="2000-01-01", periods=1000, freq="D"),
             ]
         )
         series = pd.Series(data, index)

doc/whats-new.rst

@@ -26,7 +26,10 @@ New Features
   different collections of coordinates prior to assign them to a Dataset or
   DataArray (:pull:`8102`) at once.
   By `Benoît Bovy <https://github.com/benbovy>`_.
-- Provide `preferred_chunks` for data read from netcdf files (:issue:`1440`, :pull:`7948`)
+- Provide `preferred_chunks` for data read from netcdf files (:issue:`1440`, :pull:`7948`).
+  By `Martin Raspaud <https://github.com/mraspaud>`_.
+- Improved static typing of reduction methods (:pull:`6746`).
+  By `Richard Kleijn <https://github.com/rhkleijn>`_.
 
 Breaking changes
 ~~~~~~~~~~~~~~~~
@@ -63,6 +66,10 @@ Bug fixes
 - Fix bug where :py:class:`DataArray` instances on the right-hand side
   of :py:meth:`DataArray.__setitem__` lose dimension names.
   (:issue:`7030`, :pull:`8067`) By `Darsh Ranjan <https://github.com/dranjan>`_.
+- Return ``float64`` in presence of ``NaT`` in :py:class:`~core.accessor_dt.DatetimeAccessor` and
+  special case ``NaT`` handling in :py:meth:`~core.accessor_dt.DatetimeAccessor.isocalendar()`
+  (:issue:`7928`, :pull:`8084`).
+  By `Kai Mühlbauer <https://github.com/kmuehlbauer>`_.
 
 Documentation
 ~~~~~~~~~~~~~

xarray/core/_aggregations.py

@@ -8,7 +8,7 @@
 
 from xarray.core import duck_array_ops
 from xarray.core.options import OPTIONS
-from xarray.core.types import Dims
+from xarray.core.types import Dims, Self
 from xarray.core.utils import contains_only_chunked_or_numpy, module_available
 
 if TYPE_CHECKING:
@@ -30,7 +30,7 @@ def reduce(
         keep_attrs: bool | None = None,
         keepdims: bool = False,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         raise NotImplementedError()
 
     def count(
@@ -39,7 +39,7 @@ def count(
         *,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``count`` along some dimension(s).
 
@@ -111,7 +111,7 @@ def all(
         *,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``all`` along some dimension(s).
 
@@ -183,7 +183,7 @@ def any(
         *,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``any`` along some dimension(s).
 
@@ -256,7 +256,7 @@ def max(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``max`` along some dimension(s).
 
@@ -343,7 +343,7 @@ def min(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``min`` along some dimension(s).
 
@@ -430,7 +430,7 @@ def mean(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``mean`` along some dimension(s).
 
@@ -522,7 +522,7 @@ def prod(
         min_count: int | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``prod`` along some dimension(s).
 
@@ -629,7 +629,7 @@ def sum(
         min_count: int | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``sum`` along some dimension(s).
 
@@ -736,7 +736,7 @@ def std(
         ddof: int = 0,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``std`` along some dimension(s).
 
@@ -840,7 +840,7 @@ def var(
         ddof: int = 0,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``var`` along some dimension(s).
 
@@ -943,7 +943,7 @@ def median(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``median`` along some dimension(s).
 
@@ -1034,7 +1034,7 @@ def cumsum(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``cumsum`` along some dimension(s).
 
@@ -1127,7 +1127,7 @@ def cumprod(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> Dataset:
+    ) -> Self:
         """
         Reduce this Dataset's data by applying ``cumprod`` along some dimension(s).
 
@@ -1226,7 +1226,7 @@ def reduce(
         keep_attrs: bool | None = None,
         keepdims: bool = False,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         raise NotImplementedError()
 
     def count(
@@ -1235,7 +1235,7 @@ def count(
         *,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``count`` along some dimension(s).
 
@@ -1301,7 +1301,7 @@ def all(
         *,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``all`` along some dimension(s).
 
@@ -1367,7 +1367,7 @@ def any(
         *,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``any`` along some dimension(s).
 
@@ -1434,7 +1434,7 @@ def max(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``max`` along some dimension(s).
 
@@ -1513,7 +1513,7 @@ def min(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``min`` along some dimension(s).
 
@@ -1592,7 +1592,7 @@ def mean(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``mean`` along some dimension(s).
 
@@ -1676,7 +1676,7 @@ def prod(
         min_count: int | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``prod`` along some dimension(s).
 
@@ -1773,7 +1773,7 @@ def sum(
         min_count: int | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``sum`` along some dimension(s).
 
@@ -1870,7 +1870,7 @@ def std(
         ddof: int = 0,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``std`` along some dimension(s).
 
@@ -1964,7 +1964,7 @@ def var(
         ddof: int = 0,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``var`` along some dimension(s).
 
@@ -2057,7 +2057,7 @@ def median(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``median`` along some dimension(s).
 
@@ -2140,7 +2140,7 @@ def cumsum(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``cumsum`` along some dimension(s).
 
@@ -2229,7 +2229,7 @@ def cumprod(
         skipna: bool | None = None,
         keep_attrs: bool | None = None,
         **kwargs: Any,
-    ) -> DataArray:
+    ) -> Self:
         """
         Reduce this DataArray's data by applying ``cumprod`` along some dimension(s).
 

xarray/core/accessor_dt.py

@@ -74,11 +74,25 @@ def _access_through_series(values, name):
         months = values_as_series.dt.month.values
         field_values = _season_from_months(months)
     elif name == "isocalendar":
+        # special NaT-handling can be removed when
+        # https://github.com/pandas-dev/pandas/issues/54657 is resolved
+        field_values = values_as_series.dt.isocalendar()
+        # test for <NA> and apply needed dtype
+        hasna = any(field_values.year.isnull())
+        if hasna:
+            field_values = np.dstack(
+                [
+                    getattr(field_values, name).astype(np.float64, copy=False).values
+                    for name in ["year", "week", "day"]
+                ]
+            )
+        else:
+            field_values = np.array(field_values, dtype=np.int64)
         # isocalendar returns iso- year, week, and weekday -> reshape
-        field_values = np.array(values_as_series.dt.isocalendar(), dtype=np.int64)
         return field_values.T.reshape(3, *values.shape)
     else:
         field_values = getattr(values_as_series.dt, name).values
+
     return field_values.reshape(values.shape)
 
 
@@ -110,7 +124,7 @@ def _get_date_field(values, name, dtype):
         from dask.array import map_blocks
 
         new_axis = chunks = None
-        # isocalendar adds adds an axis
+        # isocalendar adds an axis
         if name == "isocalendar":
             chunks = (3,) + values.chunksize
             new_axis = 0
@@ -119,7 +133,12 @@ def _get_date_field(values, name, dtype):
             access_method, values, name, dtype=dtype, new_axis=new_axis, chunks=chunks
         )
     else:
-        return access_method(values, name).astype(dtype, copy=False)
+        out = access_method(values, name)
+        # cast only for integer types to keep float64 in presence of NaT
+        # see https://github.com/pydata/xarray/issues/7928
+        if np.issubdtype(out.dtype, np.integer):
+            out = out.astype(dtype, copy=False)
+        return out
 
 
 def _round_through_series_or_index(values, name, freq):

xarray/core/computation.py

@@ -1402,14 +1402,14 @@ def _cov_corr(
     ) / (valid_count)
 
     if method == "cov":
-        return cov  # type: ignore[return-value]
+        return cov
 
     else:
         # compute std + corr
         da_a_std = da_a.std(dim=dim)
         da_b_std = da_b.std(dim=dim)
         corr = cov / (da_a_std * da_b_std)
-        return corr  # type: ignore[return-value]
+        return corr
 
 
 def cross(

xarray/core/rolling.py

@@ -475,9 +475,8 @@ def reduce(
             obj, rolling_dim, keep_attrs=keep_attrs, fill_value=fillna
         )
 
-        result = windows.reduce(
-            func, dim=list(rolling_dim.values()), keep_attrs=keep_attrs, **kwargs
-        )
+        dim = list(rolling_dim.values())
+        result = windows.reduce(func, dim=dim, keep_attrs=keep_attrs, **kwargs)
 
         # Find valid windows based on count.
         counts = self._counts(keep_attrs=False)
@@ -494,14 +493,15 @@ def _counts(self, keep_attrs: bool | None) -> DataArray:
         # array is faster to be reduced than object array.
         # The use of skipna==False is also faster since it does not need to
         # copy the strided array.
+        dim = list(rolling_dim.values())
         counts = (
             self.obj.notnull(keep_attrs=keep_attrs)
             .rolling(
                 {d: w for d, w in zip(self.dim, self.window)},
                 center={d: self.center[i] for i, d in enumerate(self.dim)},
             )
             .construct(rolling_dim, fill_value=False, keep_attrs=keep_attrs)
-            .sum(dim=list(rolling_dim.values()), skipna=False, keep_attrs=keep_attrs)
+            .sum(dim=dim, skipna=False, keep_attrs=keep_attrs)
         )
         return counts
 

xarray/tests/test_dataset.py

@@ -380,7 +380,7 @@ def test_repr_multiindex(self) -> None:
 
     def test_repr_period_index(self) -> None:
         data = create_test_data(seed=456)
-        data.coords["time"] = pd.period_range("2000-01-01", periods=20, freq="B")
+        data.coords["time"] = pd.period_range("2000-01-01", periods=20, freq="D")
 
         # check that creating the repr doesn't raise an error #GH645
         repr(data)