Cleanups

Author: dcherian

doc/whats-new.rst

@@ -278,9 +278,6 @@ New Features
   (:issue:`9914`, :pull:`10336`)
   By `Joseph Nowak <https://github.com/josephnowak>`_.
 
-Breaking changes
-~~~~~~~~~~~~~~~~
-
 Documentation
 ~~~~~~~~~~~~~
 - HTML reprs! By `Scott Henderson <https://github.com/scottyhq>`_.
@@ -8429,8 +8426,6 @@ Backwards incompatible changes
 
   Now, the default always concatenates data variables:
 
-  .. ipython:: python
-    :verbatim:
   .. code:: python
 
     In [1]: ds = xray.Dataset({"x": 0})

xarray/core/parallel.py

@@ -481,6 +481,7 @@ def _wrapper(
 
         coordinates = merge(
             (preserved_coords, template.coords.to_dataset()[new_coord_vars]),
+            # FIXME: this should be join="exact", but breaks a test
             join="outer",
             compat="override",
         ).coords

xarray/tests/arrays.py

@@ -3,7 +3,7 @@
 """
 
 from collections.abc import Callable, Iterable
-from typing import Any
+from typing import Any, Self
 
 import numpy as np
 
@@ -126,6 +126,23 @@ def broadcast_to(
     return ConcatenatableArray(result)
 
 
+@implements(np.full_like)
+def full_like(
+    x: "ConcatenatableArray", /, fill_value, **kwargs
+) -> "ConcatenatableArray":
+    """
+    Broadcasts an array to a specified shape, by either manipulating chunk keys or copying chunk manifest entries.
+    """
+    if not isinstance(x, ConcatenatableArray):
+        raise TypeError
+    return ConcatenatableArray(np.full(x.shape, fill_value=fill_value, **kwargs))
+
+
+@implements(np.all)
+def numpy_all(x: "ConcatenatableArray", **kwargs) -> "ConcatenatableArray":
+    return type(x)(np.all(x._array, **kwargs))
+
+
 class ConcatenatableArray:
     """Disallows loading or coercing to an index but does support concatenation / stacking."""
 
@@ -169,6 +186,9 @@ def __getitem__(self, key) -> "ConcatenatableArray":
                 raise UnexpectedDataAccess("Tried accessing data.")
         return ConcatenatableArray(arr)
 
+    def __eq__(self, other: "ConcatenatableArray") -> "ConcatenatableArray":
+        return ConcatenatableArray(self._array == other._array)
+
     def __array_function__(self, func, types, args, kwargs) -> Any:
         if func not in CONCATENATABLEARRAY_HANDLED_ARRAY_FUNCTIONS:
             return NotImplemented
@@ -190,3 +210,9 @@ def astype(self, dtype: np.dtype, /, *, copy: bool = True) -> "ConcatenatableArr
             raise NotImplementedError()
         else:
             return self
+
+    def __and__(self, other: Self) -> "ConcatenatableArray":
+        return type(self)(self._array & other._array)
+
+    def __or__(self, other: Self) -> "ConcatenatableArray":
+        return type(self)(self._array | other._array)

xarray/tests/test_concat.py

@@ -273,10 +273,13 @@ def test_concat_multiple_datasets_missing_vars(include_day: bool) -> None:
         "cloud_cover",
     ]
 
+    # must specify if concat_dim='day' is not part of the vars
+    kwargs = {"data_vars": "all"} if not include_day else {}
+
     datasets = create_concat_datasets(
         len(vars_to_drop), seed=123, include_day=include_day
     )
-    expected = concat(datasets, dim="day", data_vars="all")
+    expected = concat(datasets, dim="day", **kwargs)
 
     for i, name in enumerate(vars_to_drop):
         if include_day:
@@ -290,7 +293,7 @@ def test_concat_multiple_datasets_missing_vars(include_day: bool) -> None:
         for ds, varname in zip(datasets, vars_to_drop, strict=True)
     ]
 
-    actual = concat(datasets, dim="day", data_vars="all")
+    actual = concat(datasets, dim="day", **kwargs)
 
     assert list(actual.data_vars.keys()) == [
         "pressure",
@@ -500,7 +503,7 @@ def rectify_dim_order(self, data: Dataset, dataset) -> Dataset:
         "dim,data", [["dim1", True], ["dim2", False]], indirect=["data"]
     )
     def test_concat_simple(self, data: Dataset, dim, coords) -> None:
-        datasets = [g for _, g in data.groupby(dim, squeeze=False)]
+        datasets = [g for _, g in data.groupby(dim)]
         assert_identical(data, concat(datasets, dim, coords=coords, compat="equals"))
 
     def test_concat_merge_variables_present_in_some_datasets(
@@ -544,7 +547,7 @@ def test_concat_coords_kwarg(
         data = data.copy(deep=True)
         # make sure the coords argument behaves as expected
         data.coords["extra"] = ("dim4", np.arange(3))
-        datasets = [g.squeeze() for _, g in data.groupby(dim, squeeze=False)]
+        datasets = [g.squeeze() for _, g in data.groupby(dim)]
 
         actual = concat(
             datasets, data[dim], coords=coords, data_vars="all", compat="equals"
@@ -900,6 +903,8 @@ def test_concat_promote_shape_for_scalars_with_mixed_lengths_along_concat_dim(
         actual = concat(objs, "x", coords="different", compat="equals")
         expected = Dataset({"x": [0, 1, 2]}, {"y": ("x", [-1, -2, -2])})
         assert_identical(actual, expected)
+        actual = concat(objs, "x", coords="all")
+        assert_identical(actual, expected)
 
     def test_concat_promote_shape_broadcast_1d_x_1d_goes_to_2d(self) -> None:
         objs = [
@@ -995,8 +1000,8 @@ def test_concat_along_new_dim_multiindex(self) -> None:
     @pytest.mark.parametrize("fill_value", [dtypes.NA, 2, 2.0, {"a": 2, "b": 1}])
     def test_concat_fill_value(self, fill_value) -> None:
         datasets = [
-            Dataset({"a": ("x", [2, 3]), "b": ("x", [-2, 1]), "x": [1, 2]}),
-            Dataset({"a": ("x", [1, 2]), "b": ("x", [3, -1]), "x": [0, 1]}),
+            Dataset({"a": ("x", [2, 3]), "b": ("x", [-2, 1])}, {"x": [1, 2]}),
+            Dataset({"a": ("x", [1, 2]), "b": ("x", [3, -1])}, {"x": [0, 1]}),
         ]
         if fill_value == dtypes.NA:
             # if we supply the default, we expect the missing value for a
@@ -1128,13 +1133,15 @@ def test_concat(self) -> None:
         stacked = concat(grouped, pd.Index(ds["x"], name="x"))
         assert_identical(foo, stacked)
 
-        actual2 = concat([foo[0], foo[1]], pd.Index([0, 1]), coords="all").reset_coords(
-            drop=True
-        )
+        actual2 = concat(
+            [foo.isel(x=0), foo.isel(x=1)], pd.Index([0, 1]), coords="all"
+        ).reset_coords(drop=True)
         expected = foo[:2].rename({"x": "concat_dim"})
         assert_identical(expected, actual2)
 
-        actual3 = concat([foo[0], foo[1]], [0, 1], coords="all").reset_coords(drop=True)
+        actual3 = concat(
+            [foo.isel(x=0), foo.isel(x=1)], [0, 1], coords="all"
+        ).reset_coords(drop=True)
         expected = foo[:2].rename({"x": "concat_dim"})
         assert_identical(expected, actual3)
 
@@ -1196,7 +1203,7 @@ def test_concat_avoids_index_auto_creation(self) -> None:
         assert combined.indexes == {}
 
         # should not raise on stack
-        combined = concat(arrays, dim="z", coords="different", compat="equals")
+        combined = concat(arrays, dim="z")
         assert combined.shape == (2, 3, 3)
         assert combined.dims == ("z", "x", "y")
 

xarray/tests/test_dataarray.py

@@ -1453,10 +1453,11 @@ def test_concat_with_default_coords_warns(self) -> None:
 
         with pytest.warns(FutureWarning):
             original = xr.concat([da, db], dim="x")
+            assert original.y.size == 4
         with set_options(use_new_combine_kwarg_defaults=True):
+            # default compat="minimal" will pick the first one
             new = xr.concat([da, db], dim="x")
-
-        assert original.y.shape != new.y.shape
+            assert new.y.size == 1
 
     def test_virtual_default_coords(self) -> None:
         array = DataArray(np.zeros((5,)), dims="x")

xarray/util/deprecation_helpers.py

@@ -206,9 +206,9 @@ def warning_message(self, message: str, recommend_set_options: bool = True) -> s
 
     def error_message(self) -> str:
         return (
-            f" Error might be related to new default ({self._name}={self._new!r}). "
-            f"Previously the default was {self._name}={self._old!r}. "
-            f"The recommendation is to set {self._name} explicitly for this case."
+            f" Error might be related to new default (`{self._name}={self._new!r}`). "
+            f"Previously the default was `{self._name}={self._old!r}`. "
+            f"The recommendation is to set {self._name!r} explicitly for this case."
         )