BUG: groupby ffill adds labels as extra column (pandas-dev#21521)

doc/source/whatsnew/v0.25.0.rst

@@ -389,6 +389,7 @@ Groupby/Resample/Rolling
 - Bug in :meth:`pandas.core.window.Rolling.min` and :meth:`pandas.core.window.Rolling.max` that caused a memory leak (:issue:`25893`)
 - Bug in :meth:`pandas.core.groupby.GroupBy.idxmax` and :meth:`pandas.core.groupby.GroupBy.idxmin` with datetime column would return incorrect dtype (:issue:`25444`, :issue:`15306`)
 - Bug in :meth:`pandas.core.groupby.GroupBy.cumsum`, :meth:`pandas.core.groupby.GroupBy.cumprod`, :meth:`pandas.core.groupby.GroupBy.cummin` and :meth:`pandas.core.groupby.GroupBy.cummax` with categorical column having absent categories, would return incorrect result or segfault (:issue:`16771`)
+- Bug in :meth:`pandas.core.groupby.DataFrameGroupBy.ffill` and :meth:`pandas.core.groupby.DataFrameGroupBy.bfill` when group labels are not in frame, would concat them with the return value. (:issue:`21521`)
 
 Reshaping
 ^^^^^^^^^

pandas/core/groupby/generic.py

@@ -1473,7 +1473,8 @@ def _fill(self, direction, limit=None):
         """Overridden method to join grouped columns in output"""
         res = super(DataFrameGroupBy, self)._fill(direction, limit=limit)
         output = OrderedDict(
-            (grp.name, grp.grouper) for grp in self.grouper.groupings)
+            (grp.name, grp.grouper) for grp in self.grouper.groupings
+            if grp.in_axis)
 
         from pandas import concat
         return concat((self._wrap_transformed_output(output), res), axis=1)

pandas/tests/groupby/test_transform.py

@@ -880,3 +880,18 @@ def test_transform_absent_categories(func):
     result = getattr(df.y.groupby(df.x), func)()
     expected = df.y
     assert_series_equal(result, expected)
+
+
+@pytest.mark.parametrize('func', ['ffill', 'bfill'])
+@pytest.mark.parametrize('arg', ['level', 'labels'])
+def test_groupby_ffill_index(func, arg):
+    # GH 21521
+    df = pd.DataFrame([[0]])
+    if arg == 'level':
+        groups = df.groupby(level=0)
+    elif arg == 'labels':
+        groups = df.groupby(Series([0]))
+    result = getattr(groups, func)()
+    expected = df
+
+    assert_frame_equal(result, expected)