Merge branch 'main' into netcdf4-memory

Author: shoyer

CLAUDE.md

@@ -19,3 +19,10 @@ uv run pytest xarray/tests/test_dataarray.py  # Specific file
 pre-commit run --all-files  # Includes ruff and other checks
 uv run dmypy run  # Type checking with mypy
 ```
+
+## GitHub Interaction Guidelines
+
+- **NEVER impersonate the user on GitHub** - Do not post comments, create issues, or interact with the xarray GitHub repository unless explicitly instructed
+- Never create GitHub issues or PRs unless explicitly requested by the user
+- Never post "update" messages, progress reports, or explanatory comments on GitHub issues/PRs unless specifically asked
+- Always require explicit user direction before creating pull requests or pushing to the xarray GitHub repository

doc/whats-new.rst

@@ -17,6 +17,10 @@ New Features
 Breaking changes
 ~~~~~~~~~~~~~~~~
 
+- :py:meth:`Dataset.update` now returns ``None``, instead of the updated dataset. This
+  completes the deprecation cycle started in version 0.17. The method still updates the
+  dataset in-place. (:issue:`10167`)
+  By `Maximilian Roos <https://github.com/max-sixty>`_.
 
 Deprecations
 ~~~~~~~~~~~~

pyproject.toml

@@ -65,6 +65,7 @@ types = [
   "types-openpyxl",
   "types-python-dateutil",
   "types-pytz",
+  "types-requests",
   "types-setuptools",
 ]
 

xarray/core/coordinates.py

@@ -180,11 +180,13 @@ def to_index(self, ordered_dims: Sequence[Hashable] | None = None) -> pd.Index:
                     np.tile(np.repeat(code, repeat_counts[i]), tile_counts[i])
                     for code in codes
                 ]
-                level_list += [list(level) for level in levels]
+                level_list += levels
                 names += index.names
 
         return pd.MultiIndex(
-            levels=level_list, codes=[list(c) for c in code_list], names=names
+            levels=level_list,  # type: ignore[arg-type,unused-ignore]
+            codes=[list(c) for c in code_list],
+            names=names,
         )
 
 

xarray/core/dataset.py

@@ -5592,7 +5592,7 @@ def unstack(
                 result = result._unstack_once(d, stacked_indexes[d], fill_value, sparse)
         return result
 
-    def update(self, other: CoercibleMapping) -> Self:
+    def update(self, other: CoercibleMapping) -> None:
         """Update this dataset's variables with those from another dataset.
 
         Just like :py:meth:`dict.update` this is a in-place operation.
@@ -5609,14 +5609,6 @@ def update(self, other: CoercibleMapping) -> Self:
             - mapping {var name: (dimension name, array-like)}
             - mapping {var name: (tuple of dimension names, array-like)}
 
-        Returns
-        -------
-        updated : Dataset
-            Updated dataset. Note that since the update is in-place this is the input
-            dataset.
-
-            It is deprecated since version 0.17 and scheduled to be removed in 0.21.
-
         Raises
         ------
         ValueError
@@ -5629,7 +5621,7 @@ def update(self, other: CoercibleMapping) -> Self:
         Dataset.merge
         """
         merge_result = dataset_update_method(self, other)
-        return self._replace(inplace=True, **merge_result._asdict())
+        self._replace(inplace=True, **merge_result._asdict())
 
     def merge(
         self,

xarray/tests/test_backends.py

@@ -5222,9 +5222,7 @@ def test_memoryview_write_netcdf4_read_h5netcdf() -> None:
 @requires_h5netcdf_ros3
 class TestH5NetCDFDataRos3Driver(TestCommon):
     engine: T_NetcdfEngine = "h5netcdf"
-    test_remote_dataset: str = (
-        "https://www.unidata.ucar.edu/software/netcdf/examples/OMI-Aura_L2-example.nc"
-    )
+    test_remote_dataset: str = "https://archive.unidata.ucar.edu/software/netcdf/examples/OMI-Aura_L2-example.nc"
 
     @pytest.mark.filterwarnings("ignore:Duplicate dimension names")
     def test_get_variable_list(self) -> None:

xarray/tests/test_dataarray.py

@@ -3528,6 +3528,34 @@ def test_to_dataframe_0length(self) -> None:
         assert len(actual) == 0
         assert_array_equal(actual.index.names, list("ABC"))
 
+    @pytest.mark.parametrize(
+        "x_dtype,y_dtype,v_dtype",
+        [
+            (np.uint32, np.float32, np.uint32),
+            (np.int16, np.float64, np.int64),
+            (np.uint8, np.float32, np.uint16),
+            (np.int32, np.float32, np.int8),
+        ],
+    )
+    def test_to_dataframe_coord_dtypes_2d(self, x_dtype, y_dtype, v_dtype) -> None:
+        x = np.array([1], dtype=x_dtype)
+        y = np.array([1.0], dtype=y_dtype)
+        v = np.array([[42]], dtype=v_dtype)
+
+        da = DataArray(v, dims=["x", "y"], coords={"x": x, "y": y})
+        df = da.to_dataframe(name="v").reset_index()
+
+        # Check that coordinate dtypes are preserved
+        assert df["x"].dtype == np.dtype(x_dtype), (
+            f"x coord: expected {x_dtype}, got {df['x'].dtype}"
+        )
+        assert df["y"].dtype == np.dtype(y_dtype), (
+            f"y coord: expected {y_dtype}, got {df['y'].dtype}"
+        )
+        assert df["v"].dtype == np.dtype(v_dtype), (
+            f"v data: expected {v_dtype}, got {df['v'].dtype}"
+        )
+
     @requires_dask_expr
     @requires_dask
     @pytest.mark.xfail(not has_dask_ge_2025_1_0, reason="dask-expr is broken")