move tests, adress feedback

Author: chris-b1

pandas/tests/test_categorical.py

@@ -3943,38 +3943,6 @@ def f():
                                   'category', categories=list('cab'))})
         tm.assert_frame_equal(result, expected)
 
-    def test_union(self):
-        from pandas.types.concat import union_categoricals
-
-        s = Categorical(list('abc'))
-        s2 = Categorical(list('abd'))
-        result = union_categoricals([s, s2])
-        expected = Categorical(list('abcabd'))
-        tm.assert_categorical_equal(result, expected, ignore_order=True)
-
-        s = Categorical([0,1,2])
-        s2 = Categorical([2,3,4])
-        result = union_categoricals([s, s2])
-        expected = Categorical([0,1,2,2,3,4])
-        tm.assert_categorical_equal(result, expected, ignore_order=True)
-
-        s = Categorical([0,1.2,2])
-        s2 = Categorical([2,3.4,4])
-        result = union_categoricals([s, s2])
-        expected = Categorical([0,1.2,2,2,3.4,4])
-        tm.assert_categorical_equal(result, expected, ignore_order=True)
-
-        # can't be ordered
-        s = Categorical([0,1.2,2], ordered=True)
-        with tm.assertRaises(TypeError):
-            union_categoricals([s, s2])
-
-        # must exactly match types
-        s = Categorical([0,1.2,2])
-        s2 = Categorical([2,3,4])
-        with tm.assertRaises(TypeError):
-            union_categoricals([s, s2])
-
     def test_categorical_index_preserver(self):
 
         a = Series(np.arange(6, dtype='int64'))

pandas/tools/tests/test_concat.py

@@ -9,7 +9,8 @@
 from pandas import (DataFrame, concat,
                     read_csv, isnull, Series, date_range,
                     Index, Panel, MultiIndex, Timestamp,
-                    DatetimeIndex)
+                    DatetimeIndex, Categorical)
+from pandas.types.concat import union_categoricals
 from pandas.util import testing as tm
 from pandas.util.testing import (assert_frame_equal,
                                  makeCustomDataframe as mkdf,
@@ -919,6 +920,37 @@ def test_concat_keys_with_none(self):
                           keys=['b', 'c', 'd', 'e'])
         tm.assert_frame_equal(result, expected)
 
+    def test_union_categorical(self):
+        # GH 13361
+        s = Categorical(list('abc'))
+        s2 = Categorical(list('abd'))
+        result = union_categoricals([s, s2])
+        expected = Categorical(list('abcabd'))
+        tm.assert_categorical_equal(result, expected, ignore_order=True)
+
+        s = Categorical([0, 1, 2])
+        s2 = Categorical([2, 3, 4])
+        result = union_categoricals([s, s2])
+        expected = Categorical([0, 1, 2, 2, 3, 4])
+        tm.assert_categorical_equal(result, expected, ignore_order=True)
+
+        s = Categorical([0, 1.2, 2])
+        s2 = Categorical([2, 3.4, 4])
+        result = union_categoricals([s, s2])
+        expected = Categorical([0, 1.2, 2, 2, 3.4, 4])
+        tm.assert_categorical_equal(result, expected, ignore_order=True)
+
+        # can't be ordered
+        s = Categorical([0, 1.2, 2], ordered=True)
+        with tm.assertRaises(TypeError):
+            union_categoricals([s, s2])
+
+        # must exactly match types
+        s = Categorical([0, 1.2, 2])
+        s2 = Categorical([2, 3, 4])
+        with tm.assertRaises(TypeError):
+            union_categoricals([s, s2])
+
     def test_concat_bug_1719(self):
         ts1 = tm.makeTimeSeries()
         ts2 = tm.makeTimeSeries()[::2]

pandas/types/concat.py

@@ -216,7 +216,7 @@ def union_categoricals(to_union):
        A single array, categories will be ordered as they
        appear in the list
     """
-    from pandas import Index, Categorical
+    from pandas import Index, Categorical, unique
 
     if any(c.ordered for c in to_union):
         raise TypeError("Can only combine unordered Categoricals")
@@ -226,19 +226,16 @@ def union_categoricals(to_union):
                for c in to_union):
         raise TypeError("dtype of categories must be the same")
 
-    for i, c in enumerate(to_union):
-        if i == 0:
-            cats = c.categories.tolist()
-        else:
-            cats = cats + c.categories.difference(Index(cats)).tolist()
+    unique_cats = unique(np.concatenate([c.categories for c in to_union]))
+    categories = Index(unique_cats)
 
-    cats = Index(cats)
     new_codes = []
     for c in to_union:
-        indexer = cats.get_indexer(c.categories)
+        indexer = categories.get_indexer(c.categories)
         new_codes.append(indexer.take(c.codes))
     codes = np.concatenate(new_codes)
-    return Categorical.from_codes(codes, cats)
+    return Categorical(codes, categories=categories, ordered=False,
+                       fastpath=True)
 
 
 def _concat_datetime(to_concat, axis=0, typs=None):