ENH: add a metaclass to CategoricalDtype to provide issubclass support (for select_dtypes)

Author: jreback

pandas/core/common.py

@@ -106,15 +106,23 @@ class to receive bound method
     else:
         setattr(cls, name, func)
 
+class CategoricalDtypeType(type):
+    """
+    the type of CategoricalDtype, this metaclass determines subclass ability
+    """
+    def __init__(cls, name, bases, attrs):
+        pass
+
 class CategoricalDtype(object):
+    __meta__ = CategoricalDtypeType
     """
     A np.dtype duck-typed class, suitable for holding a custom categorical dtype.
 
     THIS IS NOT A REAL NUMPY DTYPE, but essentially a sub-class of np.object
     """
     name = 'category'
     names = None
-    type = np.object_
+    type = CategoricalDtypeType
     subdtype = None
     kind = 'O'
     str = '|O08'
@@ -1702,6 +1710,8 @@ def _get_dtype_from_object(dtype):
     elif isinstance(dtype, compat.string_types):
         if dtype == 'datetime' or dtype == 'timedelta':
             dtype += '64'
+        elif dtype == 'category':
+            return CategoricalDtypeType
         try:
             return _get_dtype_from_object(getattr(np, dtype))
         except AttributeError:
@@ -1712,10 +1722,6 @@ def _get_dtype_from_object(dtype):
     return _get_dtype_from_object(np.dtype(dtype))
 
 
-_string_dtypes = frozenset(map(_get_dtype_from_object, (compat.binary_type,
-                                                        compat.text_type)))
-
-
 def _get_info_slice(obj, indexer):
     """Slice the info axis of `obj` with `indexer`."""
     if not hasattr(obj, '_info_axis_number'):
@@ -2383,6 +2389,7 @@ def is_bool_dtype(arr_or_dtype):
 def is_categorical_dtype(arr_or_dtype):
     if hasattr(arr_or_dtype,'dtype'):
         arr_or_dtype = arr_or_dtype.dtype
+
     if isinstance(arr_or_dtype, CategoricalDtype):
         return True
     try:
@@ -2427,6 +2434,10 @@ def _is_sequence(x):
         return False
 
 
+_string_dtypes = frozenset(map(_get_dtype_from_object, (compat.binary_type,
+                                                        compat.text_type)))
+
+
 _ensure_float64 = algos.ensure_float64
 _ensure_float32 = algos.ensure_float32
 _ensure_int64 = algos.ensure_int64

pandas/tests/test_frame.py

@@ -13020,11 +13020,16 @@ def test_select_dtypes_include(self):
                         'b': list(range(1, 4)),
                         'c': np.arange(3, 6).astype('u1'),
                         'd': np.arange(4.0, 7.0, dtype='float64'),
-                        'e': [True, False, True]})
+                        'e': [True, False, True],
+                        'f': pd.Categorical(list('abc'))})
         ri = df.select_dtypes(include=[np.number])
         ei = df[['b', 'c', 'd']]
         tm.assert_frame_equal(ri, ei)
 
+        ri = df.select_dtypes(include=[np.number,'category'])
+        ei = df[['b', 'c', 'd', 'f']]
+        tm.assert_frame_equal(ri, ei)
+
     def test_select_dtypes_exclude(self):
         df = DataFrame({'a': list('abc'),
                         'b': list(range(1, 4)),