change groupby-behaviour (duplicates) & tests

Only duplicates get suffixed by their corresponding enumeration value:
['name', None, 'name'] gets transformed into ['name_0', None, 'name_1']

Superfluous test cases have been deleted and some additonal test
statements have been added.

Author: guenteru

pandas/core/groupby/groupby.py

@@ -2298,18 +2298,28 @@ def levels(self):
 
     @property
     def names(self):
-        # GH 19029
         # add suffix to level name in case they contain duplicates (GH 19029):
         orig_names = [ping.name for ping in self.groupings]
         # if no names were assigned return the original names
         if all(x is None for x in orig_names):
             return orig_names
-        # in case duplicates are contained rename all of them
-        if len(set(orig_names)) < len(orig_names):
-            orig_names = [''.join([str(x), str(i)])
-                          for i, x in enumerate(orig_names)]
 
-        return orig_names
+        suffixes = collections.defaultdict(int)
+        dups = {n: count for n, count in
+                collections.Counter(orig_names).items() if count > 1}
+        new_names = []
+        for name in orig_names:
+            if name not in dups:
+                new_names.append(name)
+            else:
+                if name is not None:
+                    new_name = '{0}_{1}'.format(name, suffixes[name])
+                else:
+                    new_name = '{0}'.format(suffixes[name])
+                suffixes[name] += 1
+                new_names.append(new_name)
+
+        return new_names
 
     def size(self):
         """

pandas/tests/groupby/test_categorical.py

@@ -558,14 +558,6 @@ def test_as_index():
     result = df.groupby(['cat', s], as_index=False, observed=True).sum()
     tm.assert_frame_equal(result, expected)
 
-    # GH 19029: conflicitng names should not raise a value error anymore
-    raised = False
-    try:
-        df.groupby(['cat', s.rename('cat')], observed=True).sum()
-    except ValueError:
-        raised = True
-    assert raised is False
-
     # is original index dropped?
     group_columns = ['cat', 'A']
     expected = DataFrame(

pandas/tests/groupby/test_groupby.py

@@ -1678,38 +1678,40 @@ def test_tuple_correct_keyerror():
 
 def test_dup_index_names():
     # dup. index names in groupby operations should be renamed (GH 19029):
-    df = pd.DataFrame({'date': list(pd.date_range('5.1.2018', '5.3.2018')),
+    df = pd.DataFrame({'date': pd.date_range('5.1.2018', '5.3.2018'),
                        'vals': list(range(3))})
 
-    mi = pd.MultiIndex.from_product([[5], [1, 2, 3]], names=['date0', 'date1'])
+    # duplicates get suffixed by integer position
+    mi = pd.MultiIndex.from_product([[5], [1, 2, 3]],
+                                    names=['date_0', 'date_1'])
     expected = pd.Series(data=list(range(3)), index=mi, name='vals')
-
-    failed = False
-    try:
-        result = df.groupby([df.date.dt.month, df.date.dt.day])['vals'].sum()
-    except ValueError:
-        failed = True
-
-    assert failed is False
+    result = df.groupby([df.date.dt.month, df.date.dt.day])['vals'].sum()
 
     tm.assert_series_equal(result, expected)
 
-
-def test_empty_index_names():
-    # don't rename frames in case no names were assigned (GH 19029)
-    df = pd.DataFrame({'date': list(pd.date_range('5.1.2018', '5.3.2018')),
-                       'vals': list(range(3))})
-
-    mi = pd.MultiIndex.from_product([[5], [1, 2, 3]])
+    # 2 out of 3 are duplicates and None
+    mi = pd.MultiIndex.from_product([[2018], [5], [1, 2, 3]],
+                                    names=['0', '1', 'date'])
     expected = pd.Series(data=list(range(3)), index=mi, name='vals')
+    result = df.groupby([df.date.dt.year.rename(None),
+                         df.date.dt.month.rename(None),
+                         df.date.dt.day])['vals'].sum()
+    tm.assert_series_equal(result, expected)
 
-    failed = False
-    try:
-        result = df.groupby([df.date.dt.month.rename(None),
-                             df.date.dt.day.rename(None)])['vals'].sum()
-    except ValueError:
-        failed = True
-
-    assert failed is False
+    # 2 out of 3 names (not None) are duplicates, the remaining is None
+    mi = pd.MultiIndex.from_product([[2018], [5], [1, 2, 3]],
+                                    names=['date_0', None, 'date_1'])
+    expected = pd.Series(data=list(range(3)), index=mi, name='vals')
+    result = df.groupby([df.date.dt.year,
+                         df.date.dt.month.rename(None),
+                         df.date.dt.day])['vals'].sum()
+    tm.assert_series_equal(result, expected)
 
+    # all are None
+    mi = pd.MultiIndex.from_product([[2018], [5], [1, 2, 3]],
+                                    names=[None, None, None])
+    expected = pd.Series(data=list(range(3)), index=mi, name='vals')
+    result = df.groupby([df.date.dt.year.rename(None),
+                         df.date.dt.month.rename(None),
+                         df.date.dt.day.rename(None)])['vals'].sum()
     tm.assert_series_equal(result, expected)