BUG: DataFrame.drop() does nothing for non-unique MultiIndex

closes #12701
closes #12783

Author: jonaslb

doc/source/whatsnew/v0.18.1.txt

@@ -188,6 +188,9 @@ Bug Fixes
 - Bug in ``GroupBy.filter`` when ``dropna=False`` and no groups fulfilled the criteria (:issue:`12768`)
 
 
+- Bug in ``.drop()`` with a non-unique ``MultiIndex``. (:issue:`12701`)
+
+
 
 - Bug in ``Timestamp.__repr__`` that caused ``pprint`` to fail in nested structures (:issue:`12622`)
 - Bug in ``Timedelta.min`` and ``Timedelta.max``, the properties now report the true minimum/maximum ``timedeltas`` as recognized by Pandas. See :ref:`documentation <timedeltas.limitations>`. (:issue:`12727`)

pandas/core/generic.py

@@ -1881,8 +1881,7 @@ def drop(self, labels, axis=0, level=None, inplace=False, errors='raise'):
             if level is not None:
                 if not isinstance(axis, MultiIndex):
                     raise AssertionError('axis must be a MultiIndex')
-                indexer = ~lib.ismember(
-                    axis.get_level_values(level).values, set(labels))
+                indexer = ~axis.get_level_values(level).isin(labels)
             else:
                 indexer = ~axis.isin(labels)
 

pandas/tests/test_multilevel.py

@@ -1948,6 +1948,25 @@ def test_drop_level(self):
         expected = self.frame.ix[[0, 2, 3, 6, 7, 9]].T
         assert_frame_equal(result, expected)
 
+    def test_drop_level_nonunique_datetime(self):
+        # GH 12701
+        idx = pd.Index([2, 3, 4, 4, 5], name='id')
+        idxdt = pd.to_datetime(['201603231400',
+                                '201603231500',
+                                '201603231600',
+                                '201603231600',
+                                '201603231700'])
+        df = DataFrame(np.arange(10).reshape(5, 2),
+                       columns=list('ab'), index=idx)
+        df['tstamp'] = idxdt
+        df = df.set_index('tstamp', append=True)
+        ts = pd.Timestamp('201603231600')
+        self.assertFalse(df.index.is_unique)
+
+        result = df.drop(ts, level='tstamp')
+        expected = df.loc[idx != 4]
+        assert_frame_equal(result, expected)
+
     def test_drop_preserve_names(self):
         index = MultiIndex.from_arrays([[0, 0, 0, 1, 1, 1],
                                         [1, 2, 3, 1, 2, 3]],