Merge pull request #6718 from jreback/groupby_agg

BUG: Bug in consistency of groupby aggregation when passing a custom function (GH6715)

Author: jreback

doc/source/release.rst

@@ -139,11 +139,11 @@ API Changes
 - Following keywords are now acceptable for :meth:`DataFrame.plot(kind='bar')` and :meth:`DataFrame.plot(kind='barh')`.
 
   - `width`: Specify the bar width. In previous versions, static value 0.5 was passed to matplotlib and it cannot be overwritten. (:issue:`6604`)
-  
+
   - `align`: Specify the bar alignment. Default is `center` (different from matplotlib). In previous versions, pandas passes `align='edge'` to matplotlib and adjust the location to `center` by itself, and it results `align` keyword is not applied as expected. (:issue:`4525`)
-  
+
   - `position`: Specify relative alignments for bar plot layout. From 0 (left/bottom-end) to 1(right/top-end). Default is 0.5 (center). (:issue:`6604`)
-  
+
 - Define and document the order of column vs index names in query/eval
     (:issue:`6676`)
 
@@ -289,6 +289,7 @@ Bug Fixes
 - Bug in binary operations with a rhs of a Series not aligning (:issue:`6681`)
 - Bug in ``DataFrame.to_stata`` which incorrectly handles nan values and ignores 'with_index' keyword argument (:issue:`6685`)
 - Bug in resample with extra bins when using an evenly divisible frequency (:issue:`4076`)
+- Bug in consistency of groupby aggregation when passing a custom function (:issue:`6715`)
 
 pandas 0.13.1
 -------------

pandas/core/groupby.py

@@ -1081,10 +1081,13 @@ def apply(self, f, data, axis=0):
             try:
                 values, mutated = splitter.fast_apply(f, group_keys)
                 return group_keys, values, mutated
-            except Exception:
+            except (lib.InvalidApply):
                 # we detect a mutation of some kind
                 # so take slow path
                 pass
+            except (Exception) as e:
+                # raise this error to the caller
+                pass
 
         result_values = []
         for key, (i, group) in zip(group_keys, splitter):
@@ -2295,7 +2298,15 @@ def aggregate(self, arg, *args, **kwargs):
             if self.grouper.nkeys > 1:
                 return self._python_agg_general(arg, *args, **kwargs)
             else:
-                result = self._aggregate_generic(arg, *args, **kwargs)
+
+                # try to treat as if we are passing a list
+                try:
+                    assert not args and not kwargs
+                    result = self._aggregate_multiple_funcs([arg])
+                    result.columns = Index(result.columns.levels[0],
+                                           name=self._selected_obj.columns.name)
+                except:
+                    result = self._aggregate_generic(arg, *args, **kwargs)
 
         if not self.as_index:
             if isinstance(result.index, MultiIndex):

pandas/tests/test_groupby.py

@@ -1946,6 +1946,29 @@ def test_grouping_ndarray(self):
         expected = self.df.groupby('A').sum()
         assert_frame_equal(result, expected, check_names=False)  # Note: no names when grouping by value
 
+    def test_agg_consistency(self):
+        # agg with ([]) and () not consistent
+        # GH 6715
+
+        def P1(a):
+            try:
+                return np.percentile(a.dropna(), q=1)
+            except:
+                return np.nan
+
+        import datetime as dt
+        df = DataFrame({'col1':[1,2,3,4],
+                        'col2':[10,25,26,31],
+                        'date':[dt.date(2013,2,10),dt.date(2013,2,10),dt.date(2013,2,11),dt.date(2013,2,11)]})
+
+        g = df.groupby('date')
+
+        expected = g.agg([P1])
+        expected.columns = expected.columns.levels[0]
+
+        result = g.agg(P1)
+        assert_frame_equal(result, expected)
+
     def test_apply_typecast_fail(self):
         df = DataFrame({'d': [1., 1., 1., 2., 2., 2.],
                         'c': np.tile(['a', 'b', 'c'], 2),