BUG fixes tuple agg issue 18079

doc/source/whatsnew/v0.21.1.txt

@@ -147,4 +147,4 @@ Other
 ^^^^^
 
 -
--
+-

doc/source/whatsnew/v0.22.0.txt

@@ -168,7 +168,7 @@ Plotting
 Groupby/Resample/Rolling
 ^^^^^^^^^^^^^^^^^^^^^^^^
 
--
+- Bug when grouping by a single column and aggregating with a class like ``list`` or ``tuple`` (:issue:`18079`)
 -
 -
 

grp_test.patch

@@ -0,0 +1,15 @@
+--- a/pandas/tests/groupby/test_groupby.py
++++ b/pandas/tests/groupby/test_groupby.py
+@@ -2725,3 +2725,12 @@ def _check_groupby(df, result, keys, field, f=lambda x: x.sum()):
+     expected = f(df.groupby(tups)[field])
+     for k, v in compat.iteritems(expected):
+         assert (result[k] == v)
++
++
++def test_tuple():
++    df = pd.DataFrame({'A': [1, 1, 1, 3, 3, 3],
++                       'B': [1, 1, 1, 4, 4, 4], 'C': [1, 1, 1, 3, 4, 4]})
++
++    result = df.groupby(['A', 'B']).aggregate(tuple)
++    result2 = df.groupby('A').aggregate(tuple)
++    result2 = df.groupby('A').aggregate([tuple])

pandas/core/groupby.py

@@ -2299,8 +2299,7 @@ def _aggregate_series_pure_python(self, obj, func):
         for label, group in splitter:
             res = func(group)
             if result is None:
-                if (isinstance(res, (Series, Index, np.ndarray)) or
-                        isinstance(res, list)):
+                if (isinstance(res, (Series, Index, np.ndarray))):
                     raise ValueError('Function does not reduce')
                 result = np.empty(ngroups, dtype='O')
 
@@ -3022,7 +3021,9 @@ def aggregate(self, func_or_funcs, *args, **kwargs):
         if isinstance(func_or_funcs, compat.string_types):
             return getattr(self, func_or_funcs)(*args, **kwargs)
 
-        if hasattr(func_or_funcs, '__iter__'):
+        if isinstance(func_or_funcs, collections.Iterable):
+            # Catch instances of lists / tuples
+            # but not the class list / tuple itself.
             ret = self._aggregate_multiple_funcs(func_or_funcs,
                                                  (_level or 0) + 1)
         else:

pandas/tests/groupby/test_aggregate.py

@@ -892,3 +892,36 @@ def test_sum_uint64_overflow(self):
         expected.index.name = 0
         result = df.groupby(0).sum()
         tm.assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("structure, expected", [
+        (tuple, pd.DataFrame({'C': {(1, 1): (1, 1, 1), (3, 4): (3, 4, 4)}})),
+        (list, pd.DataFrame({'C': {(1, 1): [1, 1, 1], (3, 4): [3, 4, 4]}})),
+        (lambda x: tuple(x), pd.DataFrame({'C': {(1, 1): (1, 1, 1),
+                                                 (3, 4): (3, 4, 4)}})),
+        (lambda x: list(x), pd.DataFrame({'C': {(1, 1): [1, 1, 1],
+                                                (3, 4): [3, 4, 4]}}))
+    ])
+    def test_agg_structs_dataframe(self, structure, expected):
+        df = pd.DataFrame({'A': [1, 1, 1, 3, 3, 3],
+                           'B': [1, 1, 1, 4, 4, 4], 'C': [1, 1, 1, 3, 4, 4]})
+
+        result = df.groupby(['A', 'B']).aggregate(structure)
+        expected.index.names = ['A', 'B']
+        assert_frame_equal(result, expected)
+
+    @pytest.mark.parametrize("structure, expected", [
+        (tuple, pd.Series([(1, 1, 1), (3, 4, 4)], index=[1, 3], name='C')),
+        (list, pd.Series([[1, 1, 1], [3, 4, 4]], index=[1, 3], name='C')),
+        (lambda x: tuple(x), pd.Series([(1, 1, 1), (3, 4, 4)],
+                                       index=[1, 3], name='C')),
+        (lambda x: list(x), pd.Series([[1, 1, 1], [3, 4, 4]],
+                                      index=[1, 3], name='C'))
+    ])
+    def test_agg_structs_series(self, structure, expected):
+        # Issue #18079
+        df = pd.DataFrame({'A': [1, 1, 1, 3, 3, 3],
+                           'B': [1, 1, 1, 4, 4, 4], 'C': [1, 1, 1, 3, 4, 4]})
+
+        result = df.groupby('A')['C'].aggregate(structure)
+        expected.index.name = 'A'
+        assert_series_equal(result, expected)