API: Change str for CategoricalDtype to category

Better compatibility with older versions

doc/source/whatsnew/v0.21.0.txt

@@ -157,11 +157,10 @@ The values have been correctly interpreted as integers.
 
 The ``.dtype`` property of a ``Categorical``, ``CategoricalIndex`` or a
 ``Series`` with categorical type will now return an instance of
-``CategoricalDtype``. For the most part, this is backwards compatible, though
-the string repr has changed. If you were previously using ``str(s.dtype) ==
-'category'`` to detect categorical data, switch to
-:func:`pandas.api.types.is_categorical_dtype`, which is compatible with the old
-and new ``CategoricalDtype``.
+``CategoricalDtype``. This change should be backwards compatible, though the
+repr has changed. ``str(CategoricalDtype())`` is still the string
+``'category'``, but the preferred way to detect categorical data is to use
+:func:`pandas.api.types.is_categorical_dtype`.
 
 See the :ref:`CategoricalDtype docs <categorical.categoricaldtype>` for more.
 

pandas/core/dtypes/dtypes.py

@@ -220,7 +220,7 @@ def __eq__(self, other):
  # both unordered; this could probably be optimized / cached
  return hash(self) == hash(other)
 
- def __unicode__(self):
+ def __repr__(self):
  tpl = u'CategoricalDtype(categories={}ordered={})'
  if self.categories is None:
  data = u"None, "

pandas/tests/dtypes/test_dtypes.py

@@ -649,3 +649,9 @@ def test_from_categorical_dtype_both(self):
  result = CategoricalDtype._from_categorical_dtype(
  c1, categories=[1, 2], ordered=False)
  assert result == CategoricalDtype([1, 2], ordered=False)
+
+ def test_str_vs_repr(self):
+ c1 = CategoricalDtype(['a', 'b'])
+ assert str(c1) == 'category'
+ expected = "CategoricalDtype(categories=['a', 'b'], ordered=False)"
+ assert repr(c1) == expected