DEPR: Deprecate ordered=None for CategoricalDtype

doc/source/whatsnew/v0.25.0.rst

@@ -261,7 +261,6 @@ Deprecations
 - The deprecated ``.ix[]`` indexer now raises a more visible FutureWarning instead of DeprecationWarning (:issue:`26438`).
 - Deprecated the ``units=M`` (months) and ``units=Y`` (year) parameters for ``units`` of :func:`pandas.to_timedelta`, :func:`pandas.Timedelta` and :func:`pandas.TimedeltaIndex` (:issue:`16344`)
 - The functions :func:`pandas.to_datetime` and :func:`pandas.to_timedelta` have deprecated the ``box`` keyword. Instead, use :meth:`to_numpy` or :meth:`Timestamp.to_datetime64` or :meth:`Timedelta.to_timedelta64`. (:issue:`24416`)
-- The :meth:`DataFrame.compound` and :meth:`Series.compound` methods are deprecated and will be removed in a future version.
 
 
 .. _whatsnew_0250.prior_deprecations:

pandas/core/dtypes/dtypes.py

@@ -538,8 +538,15 @@ def update_dtype(self, dtype):
             new_categories = self.categories
 
         new_ordered = dtype.ordered
+
+        # TODO(GH26336): remove this if block when ordered=None is removed
         if new_ordered is None:
             new_ordered = self.ordered
+            if self.ordered:
+                msg = ("ordered=None is deprecated and will default to False "
+                       "in a future version; ordered=True must be explicitly "
+                       "passed in order to be retained")
+                warnings.warn(msg, FutureWarning, stacklevel=3)
 
         return CategoricalDtype(new_categories, new_ordered)
 

pandas/core/generic.py

@@ -175,7 +175,10 @@ def is_copy(self, msg):
     def _validate_dtype(self, dtype):
         """ validate the passed dtype """
 
-        if dtype is not None:
+        # GH 26336: don't convert 'category' to CategoricalDtype
+        if isinstance(dtype, str) and dtype == 'category':
+            pass
+        elif dtype is not None:
             dtype = pandas_dtype(dtype)
 
             # a compound dtype

pandas/core/internals/construction.py

@@ -538,7 +538,10 @@ def sanitize_array(data, index, dtype=None, copy=False,
     Sanitize input data to an ndarray, copy if specified, coerce to the
     dtype if specified.
     """
-    if dtype is not None:
+    # GH 26336: don't convert 'category' to CategoricalDtype
+    if isinstance(dtype, str) and dtype == 'category':
+        pass
+    elif dtype is not None:
         dtype = pandas_dtype(dtype)
 
     if isinstance(data, ma.MaskedArray):

pandas/io/packers.py

@@ -623,7 +623,12 @@ def decode(obj):
         return Interval(obj['left'], obj['right'], obj['closed'])
     elif typ == 'series':
         dtype = dtype_for(obj['dtype'])
-        pd_dtype = pandas_dtype(dtype)
+
+        # GH 26336: don't convert 'category' to CategoricalDtype
+        if isinstance(dtype, str) and dtype == 'category':
+            pd_dtype = dtype
+        else:
+            pd_dtype = pandas_dtype(dtype)
 
         index = obj['index']
         result = Series(unconvert(obj['data'], dtype, obj['compress']),

pandas/tests/arrays/categorical/test_dtypes.py

@@ -160,6 +160,14 @@ def test_astype_category(self, dtype_ordered, cat_ordered):
             expected = cat
             tm.assert_categorical_equal(result, expected)
 
+    def test_astype_category_ordered_none_deprecated(self):
+        # GH 26336
+        cdt1 = CategoricalDtype(categories=list('cdab'), ordered=True)
+        cdt2 = CategoricalDtype(categories=list('cedafb'))
+        cat = Categorical(list('abcdaba'), dtype=cdt1)
+        with tm.assert_produces_warning(FutureWarning):
+            cat.astype(cdt2)
+
     def test_iter_python_types(self):
         # GH-19909
         cat = Categorical([1, 2])

pandas/tests/dtypes/test_dtypes.py

@@ -817,7 +817,14 @@ def test_update_dtype(self, ordered_fixture, new_categories, new_ordered):
         if expected_ordered is None:
             expected_ordered = dtype.ordered
 
-        result = dtype.update_dtype(new_dtype)
+        # GH 26336
+        if new_ordered is None and ordered_fixture is True:
+            with tm.assert_produces_warning(FutureWarning,
+                                            check_stacklevel=False):
+                result = dtype.update_dtype(new_dtype)
+        else:
+            result = dtype.update_dtype(new_dtype)
+
         tm.assert_index_equal(result.categories, expected_categories)
         assert result.ordered is expected_ordered
 

pandas/tests/indexes/test_category.py

@@ -490,6 +490,14 @@ def test_astype_category(self, name, dtype_ordered, index_ordered):
             expected = index
             tm.assert_index_equal(result, expected)
 
+    def test_astype_category_ordered_none_deprecated(self):
+        # GH 26336
+        cdt1 = CategoricalDtype(categories=list('cdab'), ordered=True)
+        cdt2 = CategoricalDtype(categories=list('cedafb'))
+        idx = CategoricalIndex(list('abcdaba'), dtype=cdt1)
+        with tm.assert_produces_warning(FutureWarning):
+            idx.astype(cdt2)
+
     def test_reindex_base(self):
         # Determined by cat ordering.
         idx = CategoricalIndex(list("cab"), categories=list("cab"))

pandas/tests/series/test_constructors.py

@@ -365,6 +365,33 @@ def test_constructor_categorical_dtype(self):
                           dtype=CategoricalDtype(['a', 'b'], ordered=True))
         tm.assert_series_equal(result, expected, check_categorical=True)
 
+    def test_constructor_categorical_string(self):
+        # GH 26336: the string 'category' maintains existing CategoricalDtype
+        cdt = CategoricalDtype(categories=list('dabc'), ordered=True)
+        expected = Series(list('abcabc'), dtype=cdt)
+
+        # Series(Categorical, dtype='category') keeps existing dtype
+        cat = Categorical(list('abcabc'), dtype=cdt)
+        result = Series(cat, dtype='category')
+        tm.assert_series_equal(result, expected)
+
+        # Series(Series[Categorical], dtype='category') keeps existing dtype
+        result = Series(result, dtype='category')
+        tm.assert_series_equal(result, expected)
+
+    def test_categorical_ordered_none_deprecated(self):
+        # GH 26336
+        cdt1 = CategoricalDtype(categories=list('cdab'), ordered=True)
+        cdt2 = CategoricalDtype(categories=list('cedafb'))
+
+        cat = Categorical(list('abcdaba'), dtype=cdt1)
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            Series(cat, dtype=cdt2)
+
+        s = Series(cat)
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            Series(s, dtype=cdt2)
+
     def test_categorical_sideeffects_free(self):
         # Passing a categorical to a Series and then changing values in either
         # the series or the categorical should not change the values in the

pandas/tests/series/test_dtypes.py

@@ -227,6 +227,14 @@ def test_astype_categories_deprecation(self):
             result = s.astype('category', categories=['a', 'b'], ordered=True)
         tm.assert_series_equal(result, expected)
 
+    def test_astype_category_ordered_none_deprecated(self):
+        # GH 26336
+        cdt1 = CategoricalDtype(categories=list('cdab'), ordered=True)
+        cdt2 = CategoricalDtype(categories=list('cedafb'))
+        s = Series(list('abcdaba'), dtype=cdt1)
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            s.astype(cdt2)
+
     def test_astype_from_categorical(self):
         items = ["a", "b", "c", "a"]
         s = Series(items)