Bug in loc returning multiindex in wrong oder with duplicated indexer (#40987)

Author: phofl

doc/source/whatsnew/v1.3.0.rst

@@ -919,6 +919,7 @@ Indexing
 - Bug in :meth:`DataFrame.loc.__getitem__` with :class:`MultiIndex` casting to float when at least one column is from has float dtype and we retrieve a scalar (:issue:`41369`)
 - Bug in :meth:`DataFrame.loc` incorrectly matching non-boolean index elements (:issue:`20432`)
 - Bug in :meth:`Series.__delitem__` with ``ExtensionDtype`` incorrectly casting to ``ndarray`` (:issue:`40386`)
+- Bug in :meth:`DataFrame.loc` returning :class:`MultiIndex` in wrong order if indexer has duplicates (:issue:`40978`)
 - Bug in :meth:`DataFrame.__setitem__` raising ``TypeError`` when using a str subclass as the column name with a :class:`DatetimeIndex` (:issue:`37366`)
 
 Missing

pandas/core/indexes/multi.py

@@ -3429,6 +3429,7 @@ def _reorder_indexer(
                 new_order = np.arange(n)[indexer]
             elif is_list_like(k):
                 # Generate a map with all level codes as sorted initially
+                k = algos.unique(k)
                 key_order_map = np.ones(len(self.levels[i]), dtype=np.uint64) * len(
                     self.levels[i]
                 )

pandas/tests/indexing/multiindex/test_loc.py

@@ -764,6 +764,28 @@ def test_loc_getitem_index_differently_ordered_slice_none():
     tm.assert_frame_equal(result, expected)
 
 
+@pytest.mark.parametrize("indexer", [[1, 2, 7, 6, 2, 3, 8, 7], [1, 2, 7, 6, 3, 8]])
+def test_loc_getitem_index_differently_ordered_slice_none_duplicates(indexer):
+    # GH#40978
+    df = DataFrame(
+        [1] * 8,
+        index=MultiIndex.from_tuples(
+            [(1, 1), (1, 2), (1, 7), (1, 6), (2, 2), (2, 3), (2, 8), (2, 7)]
+        ),
+        columns=["a"],
+    )
+    result = df.loc[(slice(None), indexer), :]
+    expected = DataFrame(
+        [1] * 8,
+        index=[[1, 1, 2, 1, 2, 1, 2, 2], [1, 2, 2, 7, 7, 6, 3, 8]],
+        columns=["a"],
+    )
+    tm.assert_frame_equal(result, expected)
+
+    result = df.loc[df.index.isin(indexer, level=1), :]
+    tm.assert_frame_equal(result, df)
+
+
 def test_loc_getitem_drops_levels_for_one_row_dataframe():
     # GH#10521
     mi = MultiIndex.from_arrays([["x"], ["y"], ["z"]], names=["a", "b", "c"])