Make MultiIndex.get_loc raise for unhashable type (pandas-dev#35914)

Co-authored-by: Jeff Reback <[email protected]>

Author: 2 people

doc/source/whatsnew/v1.1.2.rst

@@ -17,6 +17,7 @@ Fixed regressions
 - Regression in :meth:`DatetimeIndex.intersection` incorrectly raising ``AssertionError`` when intersecting against a list (:issue:`35876`)
 - Fix regression in updating a column inplace (e.g. using ``df['col'].fillna(.., inplace=True)``) (:issue:`35731`)
 - Performance regression for :meth:`RangeIndex.format` (:issue:`35712`)
+- Regression where :meth:`MultiIndex.get_loc` would return a slice spanning the full index when passed an empty list (:issue:`35878`)
 - Fix regression in invalid cache after an indexing operation; this can manifest when setting which does not update the data (:issue:`35521`)
 - Regression in :meth:`DataFrame.replace` where a ``TypeError`` would be raised when attempting to replace elements of type :class:`Interval` (:issue:`35931`)
 - Fix regression in pickle roundtrip of the ``closed`` attribute of :class:`IntervalIndex` (:issue:`35658`)

pandas/core/indexes/multi.py

@@ -2725,6 +2725,8 @@ def get_loc(self, key, method=None):
                 "currently supported for MultiIndex"
             )
 
+        hash(key)
+
         def _maybe_to_slice(loc):
             """convert integer indexer to boolean mask or slice if possible"""
             if not isinstance(loc, np.ndarray) or loc.dtype != "int64":
@@ -2739,8 +2741,7 @@ def _maybe_to_slice(loc):
             mask[loc] = True
             return mask
 
-        if not isinstance(key, (tuple, list)):
-            # not including list here breaks some indexing, xref #30892
+        if not isinstance(key, tuple):
             loc = self._get_level_indexer(key, level=0)
             return _maybe_to_slice(loc)
 

pandas/tests/frame/indexing/test_indexing.py

@@ -2111,7 +2111,7 @@ def test_type_error_multiindex(self):
         )
         dg = df.pivot_table(index="i", columns="c", values=["x", "y"])
 
-        with pytest.raises(TypeError, match="is an invalid key"):
+        with pytest.raises(TypeError, match="unhashable type"):
             dg[:, 0]
 
         index = Index(range(2), name="i")

pandas/tests/indexing/multiindex/test_multiindex.py

@@ -1,4 +1,5 @@
 import numpy as np
+import pytest
 
 import pandas._libs.index as _index
 from pandas.errors import PerformanceWarning
@@ -83,3 +84,10 @@ def test_nested_tuples_duplicates(self):
         df3 = df.copy(deep=True)
         df3.loc[[(dti[0], "a")], "c2"] = 1.0
         tm.assert_frame_equal(df3, expected)
+
+    def test_multiindex_get_loc_list_raises(self):
+        # https://github.com/pandas-dev/pandas/issues/35878
+        idx = pd.MultiIndex.from_tuples([("a", 1), ("b", 2)])
+        msg = "unhashable type"
+        with pytest.raises(TypeError, match=msg):
+            idx.get_loc([])

pandas/tests/series/indexing/test_setitem.py

@@ -1,6 +1,7 @@
 import numpy as np
 
-from pandas import NaT, Series, date_range
+from pandas import MultiIndex, NaT, Series, date_range
+import pandas.testing as tm
 
 
 class TestSetitemDT64Values:
@@ -17,3 +18,11 @@ def test_setitem_none_nan(self):
 
         series[5:7] = np.nan
         assert series[6] is NaT
+
+    def test_setitem_multiindex_empty_slice(self):
+        # https://github.com/pandas-dev/pandas/issues/35878
+        idx = MultiIndex.from_tuples([("a", 1), ("b", 2)])
+        result = Series([1, 2], index=idx)
+        expected = result.copy()
+        result.loc[[]] = 0
+        tm.assert_series_equal(result, expected)