Fix binning by unsorted array

Closes pydata#7759
dcherian · Apr 17, 2023 · f22e214 · f22e214
1 parent 68bae08
commit f22e214
Show file tree

Hide file tree

Showing 3 changed files with 23 additions and 26 deletions.
diff --git a/doc/whats-new.rst b/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:

diff --git a/xarray/core/groupby.py b/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:

diff --git a/xarray/tests/test_groupby.py b/xarray/tests/test_groupby.py
@@ -1371,17 +1371,30 @@ 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("coords", [np.arange(4), np.arange(4)[::-1], [2, 0, 3, 1]])
+    def test_groupby_bins(self, coords):
+        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)