Fix select_dtypes(include='int') for Windows. (#36808)

Author: Oleh Kozynets

doc/source/whatsnew/v1.3.0.rst

@@ -305,6 +305,7 @@ Numeric
 - Bug in :meth:`DataFrame.mode` and :meth:`Series.mode` not keeping consistent integer :class:`Index` for empty input (:issue:`33321`)
 - Bug in :meth:`DataFrame.rank` with ``np.inf`` and mixture of ``np.nan`` and ``np.inf`` (:issue:`32593`)
 - Bug in :meth:`DataFrame.rank` with ``axis=0`` and columns holding incomparable types raising ``IndexError`` (:issue:`38932`)
+- Bug in :func:`select_dtypes` different behavior between Windows and Linux with ``include="int"`` (:issue:`36569`)
 -
 
 Conversion

pandas/core/frame.py

@@ -3777,8 +3777,21 @@ def select_dtypes(self, include=None, exclude=None) -> DataFrame:
             raise ValueError("at least one of include or exclude must be nonempty")
 
         # convert the myriad valid dtypes object to a single representation
-        include = frozenset(infer_dtype_from_object(x) for x in include)
-        exclude = frozenset(infer_dtype_from_object(x) for x in exclude)
+        def check_int_infer_dtype(dtypes):
+            converted_dtypes = []
+            for dtype in dtypes:
+                # Numpy maps int to different types (int32, in64) on Windows and Linux
+                # see https://github.com/numpy/numpy/issues/9464
+                if (isinstance(dtype, str) and dtype == "int") or (dtype is int):
+                    converted_dtypes.append(np.int32)
+                    converted_dtypes.append(np.int64)
+                else:
+                    converted_dtypes.append(infer_dtype_from_object(dtype))
+            return frozenset(converted_dtypes)
+
+        include = check_int_infer_dtype(include)
+        exclude = check_int_infer_dtype(exclude)
+
         for dtypes in (include, exclude):
             invalidate_string_dtypes(dtypes)
 

pandas/tests/frame/methods/test_select_dtypes.py

@@ -110,7 +110,7 @@ def test_select_dtypes_exclude_include_using_list_like(self):
             {
                 "a": list("abc"),
                 "b": list(range(1, 4)),
-                "c": np.arange(3, 6).astype("u1"),
+                "c": np.arange(3, 6, dtype="u1"),
                 "d": np.arange(4.0, 7.0, dtype="float64"),
                 "e": [True, False, True],
                 "f": pd.date_range("now", periods=3).values,
@@ -128,6 +128,26 @@ def test_select_dtypes_exclude_include_using_list_like(self):
         e = df[["b", "e"]]
         tm.assert_frame_equal(r, e)
 
+    @pytest.mark.parametrize(
+        "include", [(np.bool_, "int"), (np.bool_, "integer"), ("bool", int)]
+    )
+    def test_select_dtypes_exclude_include_int(self, include):
+        # Fix select_dtypes(include='int') for Windows, FYI #36596
+        df = DataFrame(
+            {
+                "a": list("abc"),
+                "b": list(range(1, 4)),
+                "c": np.arange(3, 6, dtype="int32"),
+                "d": np.arange(4.0, 7.0, dtype="float64"),
+                "e": [True, False, True],
+                "f": pd.date_range("now", periods=3).values,
+            }
+        )
+        exclude = (np.datetime64,)
+        result = df.select_dtypes(include=include, exclude=exclude)
+        expected = df[["b", "c", "e"]]
+        tm.assert_frame_equal(result, expected)
+
     def test_select_dtypes_include_using_scalars(self):
         df = DataFrame(
             {