Add doc for counting categorical dtype

doc/source/user_guide/categorical.rst

@@ -240,6 +240,8 @@ expects a ``dtype``. For example :func:`pandas.read_csv`,
     array. In other words, ``dtype='category'`` is equivalent to
     ``dtype=CategoricalDtype()``.
 
+.. _categorical.equalitysemantics:
+
 Equality semantics
 ~~~~~~~~~~~~~~~~~~
 
@@ -1178,3 +1180,9 @@ Use ``copy=True`` to prevent such a behaviour or simply don't reuse ``Categorica
     This also happens in some cases when you supply a NumPy array instead of a ``Categorical``:
     using an int array (e.g. ``np.array([1,2,3,4])``) will exhibit the same behavior, while using
     a string array (e.g. ``np.array(["a","b","c","a"])``) will not.
+
+Counting CategoricalDtype
+~~~~~~~~~~~~~~~~~~~~~~~~~
+
+As mentioned in :ref:`Equality Semantics <categorical.equalitysemantics>`, two instances of :class:`~pandas.api.types.CategoricalDtype` compare equal
+whenever they have the same categories and order. Therefore, the multiple instances of :class:`~pandas.api.types.CategoricalDtype` will be counted as one group if they have the same categories and order.

pandas/core/base.py

@@ -1049,6 +1049,34 @@ def value_counts(
         4.0    1
         NaN    1
         Name: count, dtype: int64
+
+        **categorial_dtypes**
+
+        Rows with categorical type will be counted as one group
+        if they have same categories and order.
+        In the example below, even though ``a``, ``c``, and ``d``
+        all have the same data types of ``category``,
+        only ``c`` and ``d`` will be counted as one group
+        since ``a`` doesn't have the same categories.
+
+        >>> df = pd.DataFrame({"a": [1], "b": ["2"], "c": [3], "d": [3]})
+        >>> df = df.astype({"a": "category", "c": "category", "d": "category"})
+        >>> df
+           a  b  c  d
+        0  1  2  3  3
+
+        >>> df.dtypes
+        a    category
+        b      object
+        c    category
+        d    category
+        dtype: object
+
+        >>> df.dtypes.value_counts()
+        category    2
+        category    1
+        object      1
+        Name: count, dtype: int64
         """
         return algorithms.value_counts_internal(
             self,