Don't tupleize categories

Author: TomAugspurger

pandas/core/dtypes/dtypes.py

@@ -146,7 +146,7 @@ class CategoricalDtype(ExtensionDtype):
     def __new__(cls, categories=None, ordered=False, fastpath=False):
         from pandas.core.indexes.base import Index
         if categories is not None:
-            categories = Index(categories)
+            categories = Index(categories, tupleize_cols=False)
             # validation
             cls._validate_categories(categories, fastpath=fastpath)
             cls._validate_ordered(ordered)
@@ -211,8 +211,18 @@ def __repr__(self):
 
     @staticmethod
     def _hash_categories(categories, ordered=True):
-        from pandas.core.util.hashing import hash_array, _combine_hash_arrays
-        cat_array = hash_array(np.asarray(categories), categorize=False)
+        from pandas.core.util.hashing import (
+            hash_array, _combine_hash_arrays, hash_tuples
+        )
+
+        categories = np.asarray(categories)
+        if len(categories) and isinstance(categories[0], tuple):
+            # assumes if any individual category is a tuple, then all our. ATM
+            # I don't really want to support just some of the categories being
+            # tuples.
+            cat_array = hash_tuples(categories)
+        else:
+            cat_array = hash_array(np.asarray(categories), categorize=False)
         if ordered:
             cat_array = np.vstack([
                 cat_array, np.arange(len(cat_array), dtype=cat_array.dtype)

pandas/tests/dtypes/test_dtypes.py

@@ -119,6 +119,11 @@ def test_basic(self):
         assert not is_categorical(np.dtype('float64'))
         assert not is_categorical(1.0)
 
+    def test_tuple_categories(self):
+        categories = [(1, 'a'), (2, 'b'), (3, 'c')]
+        result = CategoricalDtype(categories)
+        assert all(result.categories == categories)
+
 
 class TestDatetimeTZDtype(Base):