Merge pull request #3017 from y-p/GH3011

segmentation fault on groupby with categorical grouper of mismatched len

doc/source/v0.11.0.txt

@@ -321,6 +321,9 @@ Bug Fixes
   - Fix pretty-printing of infinite data structures (closes GH2978_)
   - Fixed exception when plotting timeseries bearing a timezone (closes GH2877_)
   - str.contains ignored na argument (GH2806_)
+  - Substitute warning for segfault when grouping with categorical grouper
+    of mismatched length (GH3011_)
+
 
 See the `full release notes
 <https://github.com/pydata/pandas/blob/master/RELEASE.rst>`__ or issue tracker
@@ -337,3 +340,4 @@ on GitHub for a complete list.
 .. _GH2806: https://github.com/pydata/pandas/issues/2806
 .. _GH2807: https://github.com/pydata/pandas/issues/2807
 .. _GH2918: https://github.com/pydata/pandas/issues/2918
+.. _GH3011: https://github.com/pydata/pandas/issues/3011

pandas/core/groupby.py

@@ -1310,6 +1310,11 @@ def _get_grouper(obj, key=None, axis=0, level=None, sort=True):
             exclusions.append(gpr)
             name = gpr
             gpr = obj[gpr]
+
+        if (isinstance(gpr,Categorical) and len(gpr) != len(obj)):
+            errmsg = "Categorical grouper must have len(grouper) == len(data)"
+            raise AssertionError(errmsg)
+
         ping = Grouping(group_axis, gpr, name=name, level=level, sort=sort)
         groupings.append(ping)
 

pandas/tests/test_groupby.py

@@ -2237,6 +2237,15 @@ def test_groupby_first_datetime64(self):
         got_dt = result.dtype
         self.assert_(issubclass(got_dt.type, np.datetime64))
 
+    def test_groupby_categorical_unequal_len(self):
+        import pandas as pd
+        #GH3011
+        series = Series([np.nan, np.nan, 1, 1, 2, 2, 3, 3, 4, 4])
+        bins =  pd.cut(series.dropna(), 4)
+
+        # len(bins) != len(series) here
+        self.assertRaises(AssertionError,lambda : series.groupby(bins).mean())
+
 def assert_fp_equal(a, b):
     assert((np.abs(a - b) < 1e-12).all())