REF/API: Stricter extension checking.

Removes is_extension_array_dtype's handling of both arrays and dtypes.
Now it handles just arrays, and we provide `is_extension_dtype` for checking
whether a dtype is an extension dtype. It's the caller's responsibility to know
whether they have an array or dtype.

Closes pandas-dev#22021

Author: TomAugspurger

doc/source/api.rst

@@ -2501,6 +2501,8 @@ Dtype introspection
     api.types.is_datetime64_ns_dtype
     api.types.is_datetime64tz_dtype
     api.types.is_extension_type
+    api.types.is_extension_array_dtype
+    api.types.is_extension_dtype
     api.types.is_float_dtype
     api.types.is_int64_dtype
     api.types.is_integer_dtype

doc/source/whatsnew/v0.24.0.txt

@@ -323,7 +323,8 @@ ExtensionType Changes
 - Bug in :meth:`Series.get` for ``Series`` using ``ExtensionArray`` and integer index (:issue:`21257`)
 - :meth:`Series.combine()` works correctly with :class:`~pandas.api.extensions.ExtensionArray` inside of :class:`Series` (:issue:`20825`)
 - :meth:`Series.combine()` with scalar argument now works for any function type (:issue:`21248`)
--
+- Added :func:`pandas.api.types.is_extension_array_dtype` for testing whether an array is an ExtensionArray and :func:`pandas.api.types.is_extension_dtype` for testing whether a dtype is an ExtensionDtype (:issue:`22021`)
+
 
 .. _whatsnew_0240.api.incompatibilities:
 

pandas/core/algorithms.py

@@ -19,6 +19,7 @@
     is_integer_dtype, is_complex_dtype,
     is_object_dtype,
     is_extension_array_dtype,
+    is_extension_dtype,
     is_categorical_dtype, is_sparse,
     is_period_dtype,
     is_numeric_dtype, is_float_dtype,
@@ -153,7 +154,7 @@ def _reconstruct_data(values, dtype, original):
     Index for extension types, otherwise ndarray casted to dtype
     """
     from pandas import Index
-    if is_extension_array_dtype(dtype):
+    if is_extension_dtype(dtype):
         values = dtype.construct_array_type()._from_sequence(values)
     elif is_datetime64tz_dtype(dtype) or is_period_dtype(dtype):
         values = Index(original)._shallow_copy(values, name=None)

pandas/core/dtypes/api.py

@@ -4,6 +4,8 @@
 
 from .common import (pandas_dtype,
                      is_dtype_equal,
+                     is_extension_dtype,
+                     is_extension_array_dtype,
                      is_extension_type,
 
                      # categorical

pandas/core/dtypes/cast.py

@@ -12,7 +12,7 @@
                      is_complex, is_datetimetz, is_categorical_dtype,
                      is_datetimelike,
                      is_extension_type,
-                     is_extension_array_dtype,
+                     is_extension_dtype,
                      is_object_dtype,
                      is_datetime64tz_dtype, is_datetime64_dtype,
                      is_datetime64_ns_dtype,
@@ -294,7 +294,7 @@ def maybe_promote(dtype, fill_value=np.nan):
     elif is_datetimetz(dtype):
         if isna(fill_value):
             fill_value = iNaT
-    elif is_extension_array_dtype(dtype) and isna(fill_value):
+    elif is_extension_dtype(dtype) and isna(fill_value):
         fill_value = dtype.na_value
     elif is_float(fill_value):
         if issubclass(dtype.type, np.bool_):
@@ -332,7 +332,7 @@ def maybe_promote(dtype, fill_value=np.nan):
         dtype = np.object_
 
     # in case we have a string that looked like a number
-    if is_extension_array_dtype(dtype):
+    if is_extension_dtype(dtype):
         pass
     elif is_datetimetz(dtype):
         pass
@@ -650,7 +650,7 @@ def astype_nansafe(arr, dtype, copy=True):
         need to be very careful as the result shape could change! """
 
     # dispatch on extension dtype if needed
-    if is_extension_array_dtype(dtype):
+    if is_extension_dtype(dtype):
         return dtype.construct_array_type()._from_sequence(
             arr, dtype=dtype, copy=copy)
 

pandas/core/dtypes/common.py

@@ -1688,38 +1688,69 @@ def is_extension_type(arr):
     return False
 
 
-def is_extension_array_dtype(arr_or_dtype):
-    """Check if an object is a pandas extension array type.
+def is_extension_array_dtype(arr):
+    """Check if an array object is a pandas extension array type.
 
     Parameters
     ----------
-    arr_or_dtype : object
+    arr : object
 
     Returns
     -------
     bool
 
     Notes
     -----
-    This checks whether an object implements the pandas extension
+    This checks whether an array object implements the pandas extension
     array interface. In pandas, this includes:
 
     * Categorical
+    * Interval
 
-    Third-party libraries may implement arrays or types satisfying
+    Third-party libraries may implement arrays satisfying
     this interface as well.
-    """
-    from pandas.core.arrays import ExtensionArray
 
-    if isinstance(arr_or_dtype, (ABCIndexClass, ABCSeries)):
-        arr_or_dtype = arr_or_dtype._values
+    See Also
+    --------
+    is_extension_dtype : Similar method for dtypes.
+    """
+    from pandas.core.dtypes.base import ExtensionDtype
 
     try:
-        arr_or_dtype = pandas_dtype(arr_or_dtype)
-    except TypeError:
-        pass
+        dtype = getattr(arr, 'dtype')
+    except AttributeError:
+        return False
+
+    return isinstance(dtype, ExtensionDtype)
 
-    return isinstance(arr_or_dtype, (ExtensionDtype, ExtensionArray))
+
+def is_extension_dtype(dtype):
+    """Check if a dtype object is a pandas extension dtype.
+
+    Parameters
+    ----------
+    arr : object
+
+    Returns
+    -------
+    bool
+
+    Notes
+    -----
+    This checks whether a dtype object implements the pandas extension
+    array interface. In pandas, this includes:
+
+    * CategoricalDtype
+    * IntervalDtype
+
+    Third-party libraries may implement dtypes satisfying
+    this interface as well.
+
+    See Also
+    --------
+    is_extension_array_dtype : Similar method for arrays.
+    """
+    return isinstance(dtype, ExtensionDtype)
 
 
 def is_complex_dtype(arr_or_dtype):

pandas/core/indexes/base.py

@@ -45,6 +45,7 @@
     is_datetime64tz_dtype,
     is_timedelta64_dtype,
     is_extension_array_dtype,
+    is_extension_dtype,
     is_hashable,
     is_iterator, is_list_like,
     is_scalar)
@@ -275,7 +276,7 @@ def __new__(cls, data=None, dtype=None, copy=False, name=None,
                                  closed=closed)
 
         # extension dtype
-        elif is_extension_array_dtype(data) or is_extension_array_dtype(dtype):
+        elif is_extension_array_dtype(data) or is_extension_dtype(dtype):
             data = np.asarray(data)
             if not (dtype is None or is_object_dtype(dtype)):
 
@@ -1191,7 +1192,7 @@ def astype(self, dtype, copy=True):
             return CategoricalIndex(self.values, name=self.name, dtype=dtype,
                                     copy=copy)
 
-        elif is_extension_array_dtype(dtype):
+        elif is_extension_dtype(dtype):
             return Index(np.asarray(self), dtype=dtype, copy=copy)
 
         try:

pandas/core/series.py

@@ -22,6 +22,7 @@
     is_float_dtype,
     is_extension_type,
     is_extension_array_dtype,
+    is_extension_dtype,
     is_datetimelike,
     is_datetime64tz_dtype,
     is_timedelta64_dtype,
@@ -4088,7 +4089,7 @@ def _try_cast(arr, take_fast_path):
                 # that Categorical is the only array type for 'category'.
                 subarr = Categorical(arr, dtype.categories,
                                      ordered=dtype.ordered)
-            elif is_extension_array_dtype(dtype):
+            elif is_extension_dtype(dtype):
                 # create an extension array from its dtype
                 array_type = dtype.construct_array_type()
                 subarr = array_type(subarr, dtype=dtype, copy=copy)

pandas/tests/extension/base/interface.py

@@ -2,8 +2,10 @@
 
 import pandas as pd
 from pandas.compat import StringIO
-from pandas.core.dtypes.common import is_extension_array_dtype
-from pandas.core.dtypes.dtypes import ExtensionDtype
+from pandas.api.types import (
+    is_extension_array_dtype, is_extension_dtype
+)
+from pandas.api.extensions import ExtensionDtype
 
 from .base import BaseExtensionTests
 
@@ -58,10 +60,12 @@ def test_dtype_name_in_info(self, data):
 
     def test_is_extension_array_dtype(self, data):
         assert is_extension_array_dtype(data)
-        assert is_extension_array_dtype(data.dtype)
         assert is_extension_array_dtype(pd.Series(data))
         assert isinstance(data.dtype, ExtensionDtype)
 
+    def test_is_extension_dtype(self, data):
+        assert is_extension_dtype(data.dtype)
+
     def test_no_values_attribute(self, data):
         # GH-20735: EA's with .values attribute give problems with internal
         # code, disallowing this for now until solved

pandas/tests/extension/test_common.py

@@ -4,7 +4,9 @@
 import pandas as pd
 import pandas.util.testing as tm
 from pandas.core.arrays import ExtensionArray
-from pandas.core.dtypes.common import is_extension_array_dtype
+from pandas.core.dtypes.common import (
+    is_extension_array_dtype, is_extension_dtype
+)
 from pandas.core.dtypes import dtypes
 
 
@@ -38,14 +40,20 @@ class TestExtensionArrayDtype(object):
 
     @pytest.mark.parametrize('values', [
         pd.Categorical([]),
-        pd.Categorical([]).dtype,
         pd.Series(pd.Categorical([])),
-        DummyDtype(),
+
         DummyArray(np.array([1, 2])),
     ])
     def test_is_extension_array_dtype(self, values):
         assert is_extension_array_dtype(values)
 
+    @pytest.mark.parametrize('dtype', [
+        pd.Categorical([]).dtype,
+        DummyDtype(),
+    ])
+    def test_is_extension_dtype(self, dtype):
+        assert is_extension_dtype(dtype)
+
     @pytest.mark.parametrize('values', [
         np.array([]),
         pd.Series(np.array([])),
@@ -91,4 +99,4 @@ def test_is_not_extension_array_dtype(dtype):
 ])
 def test_is_extension_array_dtype(dtype):
     assert isinstance(dtype, dtypes.ExtensionDtype)
-    assert is_extension_array_dtype(dtype)
+    assert is_extension_dtype(dtype)