BUG: DataFrame(ndarray, dtype=categoricaldtype)

doc/source/whatsnew/v1.3.0.rst

@@ -184,8 +184,8 @@ Categorical
 ^^^^^^^^^^^
 - Bug in :class:`CategoricalIndex` incorrectly failing to raise ``TypeError`` when scalar data is passed (:issue:`38614`)
 - Bug in ``CategoricalIndex.reindex`` failed when ``Index`` passed with elements all in category (:issue:`28690`)
-- Bug where construcing a :class:`Categorical` from an object-dtype array of ``date`` objects did not round-trip correctly with ``astype`` (:issue:`38552`)
-
+- Bug where constructing a :class:`Categorical` from an object-dtype array of ``date`` objects did not round-trip correctly with ``astype`` (:issue:`38552`)
+- Bug in constructing a :class:`DataFrame` from an ``ndarray`` and a :class:`CategoricalDtype` (:issue:`38857`)
 
 Datetimelike
 ^^^^^^^^^^^^

pandas/core/internals/construction.py

@@ -21,7 +21,6 @@
     maybe_upcast,
 )
 from pandas.core.dtypes.common import (
-    is_categorical_dtype,
     is_datetime64tz_dtype,
     is_dtype_equal,
     is_extension_array_dtype,
@@ -160,21 +159,7 @@ def init_ndarray(values, index, columns, dtype: Optional[DtypeObj], copy: bool):
         if not len(values) and columns is not None and len(columns):
             values = np.empty((0, 1), dtype=object)
 
-    # we could have a categorical type passed or coerced to 'category'
-    # recast this to an arrays_to_mgr
-    if is_categorical_dtype(getattr(values, "dtype", None)) or is_categorical_dtype(
-        dtype
-    ):
-
-        if not hasattr(values, "dtype"):
-            values = _prep_ndarray(values, copy=copy)
-            values = values.ravel()
-        elif copy:
-            values = values.copy()
-
-        index, columns = _get_axes(len(values), 1, index, columns)
-        return arrays_to_mgr([values], columns, index, columns, dtype=dtype)
-    elif is_extension_array_dtype(values) or is_extension_array_dtype(dtype):
+    if is_extension_array_dtype(values) or is_extension_array_dtype(dtype):
         # GH#19157
 
         if isinstance(values, np.ndarray) and values.ndim > 1:
@@ -308,6 +293,7 @@ def nested_data_to_arrays(
         if isinstance(data[0], ABCSeries):
             index = _get_names_from_index(data)
         elif isinstance(data[0], Categorical):
+            # GH#38845 hit in test_constructor_categorical
             index = ibase.default_index(len(data[0]))
         else:
             index = ibase.default_index(len(data))
@@ -486,7 +472,9 @@ def _get_names_from_index(data):
     return index
 
 
-def _get_axes(N, K, index, columns) -> Tuple[Index, Index]:
+def _get_axes(
+    N: int, K: int, index: Optional[Index], columns: Optional[Index]
+) -> Tuple[Index, Index]:
     # helper to create the axes as indexes
     # return axes or defaults
 

pandas/tests/frame/test_constructors.py

@@ -1890,6 +1890,16 @@ def test_constructor_lists_to_object_dtype(self):
         assert d["a"].dtype == np.object_
         assert not d["a"][1]
 
+    def test_constructor_ndarray_categorical_dtype(self):
+        cat = Categorical(["A", "B", "C"])
+        arr = np.array(cat).reshape(-1, 1)
+        arr = np.broadcast_to(arr, (3, 4))
+
+        result = DataFrame(arr, dtype=cat.dtype)
+
+        expected = DataFrame({0: cat, 1: cat, 2: cat, 3: cat})
+        tm.assert_frame_equal(result, expected)
+
     def test_constructor_categorical(self):
 
         # GH8626
@@ -1913,11 +1923,13 @@ def test_constructor_categorical(self):
         expected = Series(list("abc"), dtype="category", name=0)
         tm.assert_series_equal(df[0], expected)
 
+    def test_construct_from_1item_list_of_categorical(self):
         # ndim != 1
         df = DataFrame([Categorical(list("abc"))])
         expected = DataFrame({0: Series(list("abc"), dtype="category")})
         tm.assert_frame_equal(df, expected)
 
+    def test_construct_from_list_of_categoricals(self):
         df = DataFrame([Categorical(list("abc")), Categorical(list("abd"))])
         expected = DataFrame(
             {
@@ -1928,18 +1940,22 @@ def test_constructor_categorical(self):
         )
         tm.assert_frame_equal(df, expected)
 
+    def test_from_nested_listlike_mixed_types(self):
         # mixed
         df = DataFrame([Categorical(list("abc")), list("def")])
         expected = DataFrame(
             {0: Series(list("abc"), dtype="category"), 1: list("def")}, columns=[0, 1]
         )
         tm.assert_frame_equal(df, expected)
 
+    def test_construct_from_listlikes_mismatched_lengths(self):
         # invalid (shape)
         msg = r"Shape of passed values is \(6, 2\), indices imply \(3, 2\)"
         with pytest.raises(ValueError, match=msg):
             DataFrame([Categorical(list("abc")), Categorical(list("abdefg"))])
 
+    def test_categorical_1d_only(self):
+        # TODO: belongs in Categorical tests
         # ndim > 1
         msg = "> 1 ndim Categorical are not supported at this time"
         with pytest.raises(NotImplementedError, match=msg):

pandas/tests/series/test_constructors.py

@@ -326,13 +326,16 @@ def test_constructor_categorical(self):
         expected = Series([1, 2, 3], dtype="int64")
         tm.assert_series_equal(result, expected)
 
+    def test_construct_from_categorical_with_dtype(self):
         # GH12574
         cat = Series(Categorical([1, 2, 3]), dtype="category")
         assert is_categorical_dtype(cat)
         assert is_categorical_dtype(cat.dtype)
-        s = Series([1, 2, 3], dtype="category")
-        assert is_categorical_dtype(s)
-        assert is_categorical_dtype(s.dtype)
+
+    def test_construct_intlist_values_category_dtype(self):
+        ser = Series([1, 2, 3], dtype="category")
+        assert is_categorical_dtype(ser)
+        assert is_categorical_dtype(ser.dtype)
 
     def test_constructor_categorical_with_coercion(self):
         factor = Categorical(["a", "b", "b", "a", "a", "c", "c", "c"])