(fix): remove unnecessary casting + nbytes

Author: ilan-gold

xarray/core/indexing.py

@@ -1926,6 +1926,10 @@ def copy(self, deep: bool = True) -> Self:
         array = self.array.copy(deep=True) if deep else self.array
         return type(self)(array, self._dtype)
 
+    @property
+    def nbytes(self) -> int:
+        return self.array.nbytes
+
 
 class PandasMultiIndexingAdapter(PandasIndexingAdapter):
     """Handles explicit indexing for a pandas.MultiIndex.

xarray/core/variable.py

@@ -192,8 +192,6 @@ def _maybe_wrap_data(data):
     if isinstance(data, pd.Index):
         return PandasIndexingAdapter(data)
     if isinstance(data, pd.api.extensions.ExtensionArray):
-        if isinstance(data.dtype, pd.Int64Dtype | pd.Float64Dtype | pd.StringDtype):
-            return np.array(data)
         return PandasExtensionArray(data)
     return data
 

xarray/tests/test_dataset.py

@@ -4380,13 +4380,7 @@ def test_setitem_pandas(self) -> None:
         ds = self.make_example_math_dataset()
         ds["x"] = np.arange(3)
         ds_copy = ds.copy()
-        series = ds["bar"].to_pandas()
-        # to_pandas will actually give the result where the internal array of the series is a NumpyExtensionArray
-        # but ds["bar"] is a numpy array.
-        # TODO: should assert_equal be updated to handle?
-        assert (ds["bar"] == series).all()
-        del ds["bar"]
-        del ds_copy["bar"]
+        ds_copy["bar"] = ds["bar"].to_pandas()
         assert_equal(ds, ds_copy)
 
     def test_setitem_auto_align(self) -> None:
@@ -4977,19 +4971,6 @@ def test_to_and_from_dataframe(self) -> None:
         expected = pd.DataFrame([[]], index=idx)
         assert expected.equals(actual), (expected, actual)
 
-    def test_from_dataframe_int_float_str_pandas_dtype(self) -> None:
-        df = pd.DataFrame([1, 2, 3], dtype=pd.Int64Dtype())
-        ds = xr.Dataset.from_dataframe(df)
-        assert isinstance(ds[0].data, np.ndarray)
-
-        df = pd.DataFrame([1, 2, 3], dtype=pd.Float64Dtype())
-        ds = xr.Dataset.from_dataframe(df)
-        assert isinstance(ds[0].data, np.ndarray)
-
-        df = pd.DataFrame(["1", "2", "3"], dtype=pd.StringDtype())
-        ds = xr.Dataset.from_dataframe(df)
-        assert isinstance(ds[0].data, np.ndarray)
-
     def test_from_dataframe_categorical_dtype_index(self) -> None:
         cat = pd.CategoricalIndex(list("abcd"))
         df = pd.DataFrame({"f": [0, 1, 2, 3]}, index=cat)

xarray/tests/test_pandas_to_xarray.py

@@ -104,8 +104,6 @@ def index_flat(request):
     index fixture, but excluding MultiIndex cases.
     """
     key = request.param
-    if key in ["bool-object", "bool-dtype", "nullable_bool", "repeats"]:
-        pytest.xfail(reason="doesn't work")
     return indices_dict[key].copy()