BUG: concat([ints, bools]) incorrect casting (#62889)

Author: jbrockmendel

doc/source/whatsnew/v3.0.0.rst

@@ -1176,6 +1176,7 @@ Groupby/resample/rolling
 
 Reshaping
 ^^^^^^^^^
+- Bug in :func:`concat` with mixed integer and bool dtypes incorrectly casting the bools to integers (:issue:`45101`)
 - Bug in :func:`qcut` where values at the quantile boundaries could be incorrectly assigned (:issue:`59355`)
 - Bug in :meth:`DataFrame.combine_first` not preserving the column order (:issue:`60427`)
 - Bug in :meth:`DataFrame.explode` producing incorrect result for :class:`pyarrow.large_list` type (:issue:`61091`)

pandas/core/dtypes/concat.py

@@ -161,6 +161,10 @@ def _get_result_dtype(
                 # coerce to object
                 target_dtype = np.dtype(object)
                 kinds = {"o"}
+    elif "b" in kinds and len(kinds) > 1:
+        # GH#21108, GH#45101
+        target_dtype = np.dtype(object)
+        kinds = {"o"}
     else:
         # error: Argument 1 to "np_find_common_type" has incompatible type
         # "*Set[Union[ExtensionDtype, Any]]"; expected "dtype[Any]"

pandas/tests/reshape/concat/test_dataframe.py

@@ -12,6 +12,28 @@
 
 
 class TestDataFrameConcat:
+    @pytest.mark.xfail(reason="GH#62888 the `mi[2][1] is 1` check fails")
+    def test_concat_multiindex_level_bool_and_numeric(self):
+        # GH#21108, GH#45101
+        left = DataFrame([123, 456], columns=["data"], index=[True, False])
+        right = DataFrame(
+            [55, 983, 69, 112, 0], columns=["data"], index=[1, 2, 3, 4, 99]
+        )
+        result = concat({"One": left, "Two": right})
+
+        # in particular, the first two entries should not be cast to ints, the
+        #  other 1 should not cast to True
+        mi = pd.MultiIndex.from_arrays(
+            [
+                ["One"] * 2 + ["Two"] * 5,
+                np.array([True, False, 1, 2, 3, 4, 99], dtype=object),
+            ],
+        )
+        assert mi[0][1] is True
+        assert type(mi[2][1]) is int
+        expected = DataFrame({"data": [123, 456, 55, 983, 69, 112, 0]}, index=mi)
+        tm.assert_frame_equal(result, expected)
+
     def test_concat_multiple_frames_dtypes(self):
         # GH#2759
         df1 = DataFrame(data=np.ones((10, 2)), columns=["foo", "bar"], dtype=np.float64)

pandas/tests/reshape/concat/test_series.py

@@ -14,6 +14,18 @@
 
 
 class TestSeriesConcat:
+    @pytest.mark.parametrize("bool_dtype", [bool, "boolean"])
+    @pytest.mark.parametrize("dtype", [np.int64, np.float64, "Int64", "Float64"])
+    def test_concat_bool_and_numeric(self, bool_dtype, dtype):
+        # GH#21108, GH#45101
+        left = Series([True, False], dtype=bool_dtype)
+        right = Series([1, 2], dtype=dtype)
+        result = concat([left, right], ignore_index=True)
+        expected = Series([True, False, 1, 2], dtype=object)
+        assert result.iloc[0] is True
+        assert type(result.iloc[2]) in [int, float]  # i.e. not bool
+        tm.assert_series_equal(result, expected)
+
     def test_concat_series(self):
         ts = Series(
             np.arange(20, dtype=np.float64),