Backport PR #36675 on branch 1.1.x (REGR: Series.loc with a MultiIndex containing Timestamp raises InvalidIndexError) (#36818)

Co-authored-by: Simon Hawkins <[email protected]>

Author: meeseeksmachine

doc/source/whatsnew/v1.1.3.rst

@@ -37,6 +37,7 @@ Fixed regressions
 - Fixed regression in modulo of :class:`Index`, :class:`Series` and :class:`DataFrame` using ``numexpr`` using C not Python semantics (:issue:`36047`, :issue:`36526`)
 - Fixed regression in :meth:`read_excel` with ``engine="odf"`` caused ``UnboundLocalError`` in some cases where cells had nested child nodes (:issue:`36122`, :issue:`35802`)
 - Fixed regression in :meth:`DataFrame.replace` inconsistent replace when using a float in the replace method (:issue:`35376`)
+- Fixed regression in :meth:`Series.loc` on a :class:`Series` with a :class:`MultiIndex` containing :class:`Timestamp` raising ``InvalidIndexError`` (:issue:`35858`)
 - Fixed regression in :class:`DataFrame` and :class:`Series` comparisons between numeric arrays and strings (:issue:`35700`, :issue:`36377`)
 - Fixed regression in :meth:`DataFrame.apply` with ``raw=True`` and user-function returning string (:issue:`35940`)
 - Fixed regression when setting empty :class:`DataFrame` column to a :class:`Series` in preserving name of index in frame (:issue:`36527`)

pandas/core/indexing.py

@@ -1064,7 +1064,7 @@ def _handle_lowerdim_multi_index_axis0(self, tup: Tuple):
         try:
             # fast path for series or for tup devoid of slices
             return self._get_label(tup, axis=axis)
-        except TypeError:
+        except (TypeError, InvalidIndexError):
             # slices are unhashable
             pass
         except KeyError as ek:

pandas/tests/indexing/multiindex/test_loc.py

@@ -493,6 +493,16 @@ def test_loc_datetime_mask_slicing():
     tm.assert_series_equal(result, expected)
 
 
+def test_loc_datetime_series_tuple_slicing():
+    # https://github.com/pandas-dev/pandas/issues/35858
+    date = pd.Timestamp("2000")
+    ser = pd.Series(
+        1, index=pd.MultiIndex.from_tuples([("a", date)], names=["a", "b"]), name="c",
+    )
+    result = ser.loc[:, [date]]
+    tm.assert_series_equal(result, ser)
+
+
 def test_loc_with_mi_indexer():
     # https://github.com/pandas-dev/pandas/issues/35351
     df = DataFrame(