Fix binning by unsorted array (#7762)

* Fix binning by unsorted array

Closes #7759

* Update xarray/tests/test_groupby.py

Co-authored-by: Illviljan <[email protected]>

---------

Co-authored-by: Illviljan <[email protected]>

Author: dcherian

doc/whats-new.rst

@@ -16,33 +16,17 @@ What's New
 
 
 
-.. _whats-new.2023.05.0:
+.. _whats-new.2023.04.1:
 
-v2023.05.0 (unreleased)
+v2023.04.1 (unreleased)
 -----------------------
 
-New Features
-~~~~~~~~~~~~
-
-
-Breaking changes
-~~~~~~~~~~~~~~~~
-
-
-Deprecations
-~~~~~~~~~~~~
-
+This is a patch release to fix a bug with binning (:issue:`7759`)
 
 Bug fixes
 ~~~~~~~~~
 
-
-Documentation
-~~~~~~~~~~~~~
-
-
-Internal Changes
-~~~~~~~~~~~~~~~~
+- Fix binning by unsorted arrays. (:issue:`7759`)
 
 
 .. _whats-new.2023.04.0:

xarray/core/groupby.py

@@ -338,7 +338,7 @@ def _factorize_bins(
     if (codes == -1).all():
         raise ValueError(f"None of the data falls within bins with edges {bins!r}")
     full_index = binned.categories
-    unique_values = binned.unique().dropna()
+    unique_values = np.sort(binned.unique().dropna())
     group_indices = [g for g in _codes_to_groups(codes, len(full_index)) if g]
 
     if len(group_indices) == 0:

xarray/tests/test_groupby.py

@@ -1371,36 +1371,51 @@ def test_groupby_multidim_map(self):
         )
         assert_identical(expected, actual)
 
-    def test_groupby_bins(self):
-        array = DataArray(np.arange(4), dims="dim_0")
+    @pytest.mark.parametrize("use_flox", [True, False])
+    @pytest.mark.parametrize("coords", [np.arange(4), np.arange(4)[::-1], [2, 0, 3, 1]])
+    def test_groupby_bins(self, coords: np.typing.ArrayLike, use_flox: bool) -> None:
+        array = DataArray(
+            np.arange(4), dims="dim_0", coords={"dim_0": coords}, name="a"
+        )
         # the first value should not be part of any group ("right" binning)
         array[0] = 99
         # bins follow conventions for pandas.cut
         # http://pandas.pydata.org/pandas-docs/stable/generated/pandas.cut.html
         bins = [0, 1.5, 5]
-        bin_coords = pd.cut(array["dim_0"], bins).categories
-        expected = DataArray(
-            [1, 5], dims="dim_0_bins", coords={"dim_0_bins": bin_coords}
+
+        df = array.to_dataframe()
+        df["dim_0_bins"] = pd.cut(array["dim_0"], bins)
+
+        expected_df = df.groupby("dim_0_bins").sum()
+        # TODO: can't convert df with IntervalIndex to Xarray
+
+        expected = (
+            expected_df.reset_index(drop=True)
+            .to_xarray()
+            .assign_coords(index=np.array(expected_df.index))
+            .rename({"index": "dim_0_bins"})["a"]
         )
-        actual = array.groupby_bins("dim_0", bins=bins).sum()
-        assert_identical(expected, actual)
 
-        actual = array.groupby_bins("dim_0", bins=bins, labels=[1.2, 3.5]).sum()
-        assert_identical(expected.assign_coords(dim_0_bins=[1.2, 3.5]), actual)
+        with xr.set_options(use_flox=use_flox):
+            actual = array.groupby_bins("dim_0", bins=bins).sum()
+            assert_identical(expected, actual)
 
-        actual = array.groupby_bins("dim_0", bins=bins).map(lambda x: x.sum())
-        assert_identical(expected, actual)
+            actual = array.groupby_bins("dim_0", bins=bins, labels=[1.2, 3.5]).sum()
+            assert_identical(expected.assign_coords(dim_0_bins=[1.2, 3.5]), actual)
 
-        # make sure original array dims are unchanged
-        assert len(array.dim_0) == 4
+            actual = array.groupby_bins("dim_0", bins=bins).map(lambda x: x.sum())
+            assert_identical(expected, actual)
 
-        da = xr.DataArray(np.ones((2, 3, 4)))
-        bins = [-1, 0, 1, 2]
-        with xr.set_options(use_flox=False):
-            actual = da.groupby_bins("dim_0", bins).mean(...)
-        with xr.set_options(use_flox=True):
-            expected = da.groupby_bins("dim_0", bins).mean(...)
-        assert_allclose(actual, expected)
+            # make sure original array dims are unchanged
+            assert len(array.dim_0) == 4
+
+            da = xr.DataArray(np.ones((2, 3, 4)))
+            bins = [-1, 0, 1, 2]
+            with xr.set_options(use_flox=False):
+                actual = da.groupby_bins("dim_0", bins).mean(...)
+            with xr.set_options(use_flox=True):
+                expected = da.groupby_bins("dim_0", bins).mean(...)
+            assert_allclose(actual, expected)
 
     def test_groupby_bins_empty(self):
         array = DataArray(np.arange(4), [("x", range(4))])