BUG: DataFrame.drop raising TypeError for string label with non-unique MultiIndex and no level provided

doc/source/whatsnew/v1.3.0.rst

@@ -140,7 +140,7 @@ Missing
 MultiIndex
 ^^^^^^^^^^
 
--
+- Bug in :meth:`DataFrame.drop` raising ``TypeError`` when :class:`MultiIndex` is non-unique and no level is provided (:issue:`36293`)
 -
 
 I/O

pandas/core/generic.py

@@ -4183,11 +4183,16 @@ def _drop_axis(
                 if errors == "raise" and indexer.all():
                     raise KeyError(f"{labels} not found in axis")
             else:
-                indexer = ~axis.isin(labels)
-                # Check if label doesn't exist along axis
-                labels_missing = (axis.get_indexer_for(labels) == -1).any()
-                if errors == "raise" and labels_missing:
-                    raise KeyError(f"{labels} not found in axis")
+                if isinstance(axis, MultiIndex) and labels.dtype == "object":
+                    # Set level to zero in case of MultiIndex and label is string,
+                    #  because isin can't handle strings for MultiIndexes GH#36293
+                    indexer = ~axis.get_level_values(0).isin(labels)
+                else:
+                    indexer = ~axis.isin(labels)
+                    # Check if label doesn't exist along axis
+                    labels_missing = (axis.get_indexer_for(labels) == -1).any()
+                    if errors == "raise" and labels_missing:
+                        raise KeyError(f"{labels} not found in axis")
 
             slicer = [slice(None)] * self.ndim
             slicer[self._get_axis_number(axis_name)] = indexer

pandas/tests/frame/methods/test_drop.py

@@ -441,3 +441,11 @@ def test_inplace_drop_and_operation(self, operation, inplace):
             # Perform operation and check result
             getattr(y, operation)(1)
             tm.assert_frame_equal(df, expected)
+
+    def test_drop_with_non_unique_multiindex(self):
+        # GH#36293
+        mi = MultiIndex.from_arrays([["x", "y", "x"], ["i", "j", "i"]])
+        df = DataFrame([1, 2, 3], index=mi)
+        result = df.drop(index="x")
+        expected = DataFrame([2], index=MultiIndex.from_arrays([["y"], ["j"]]))
+        tm.assert_frame_equal(result, expected)