Raise MixedTypeError when a mix of bools and None are stored in…

… pandas columns (#3) Fixes: rapidsai/xdf#322 This PR raises an error when a pandas column with a mix of bools & None are detected i.e., when a boolean column is of type object rather than bool/boolean.
rapidsai · Sep 30, 2023 · 952f2bc · 952f2bc
1 parent abb95fe
commit 952f2bc
Show file tree

Hide file tree

Showing 3 changed files with 19 additions and 15 deletions.
diff --git a/python/cudf/cudf/core/column/column.py b/python/cudf/cudf/core/column/column.py
@@ -2061,14 +2061,9 @@ def as_column(
             )
         else:
             pyarrow_array = pa.array(arbitrary, from_pandas=nan_as_null)
-            if (
-                arbitrary.dtype == cudf.dtype("object")
-                and cudf.dtype(pyarrow_array.type.to_pandas_dtype())
-                != cudf.dtype(arbitrary.dtype)
-                and not is_bool_dtype(
-                    cudf.dtype(pyarrow_array.type.to_pandas_dtype())
-                )
-            ):
+            if arbitrary.dtype == cudf.dtype("object") and cudf.dtype(
+                pyarrow_array.type.to_pandas_dtype()
+            ) != cudf.dtype(arbitrary.dtype):
                 raise MixedTypeError("Cannot create column with mixed types")
             if isinstance(pyarrow_array.type, pa.Decimal128Type):
                 pyarrow_type = cudf.Decimal128Dtype.from_arrow(
@@ -2458,9 +2453,6 @@ def as_column(
                     and (
                         cudf.dtype(pyarrow_array.type.to_pandas_dtype())
                         != cudf.dtype(arbitrary.dtype)
-                        and not is_bool_dtype(
-                            cudf.dtype(pyarrow_array.type.to_pandas_dtype())
-                        )
                     )
                 ):
                     raise MixedTypeError(

diff --git a/python/cudf/cudf/tests/test_dataframe.py b/python/cudf/cudf/tests/test_dataframe.py
@@ -4029,7 +4029,7 @@ def test_any(data, axis):
     dtype = None if data else float
     if np.array(data).ndim <= 1:
         pdata = pd.Series(data=data, dtype=dtype)
-        gdata = cudf.Series.from_pandas(pdata)
+        gdata = cudf.Series(data=data, dtype=dtype)
 
         if axis == 1:
             with pytest.raises(NotImplementedError):
@@ -8949,8 +8949,10 @@ def test_agg_for_dataframe_with_string_columns(aggs):
             "c": pd.Series([2, np.nan, 5.0], index=[2, 3, 4]),
         },
         {
-            "a": [True, np.nan, True],
-            "d": pd.Series([False, True, np.nan], index=[0, 1, 3]),
+            "a": pd.Series([True, None, True], dtype=pd.BooleanDtype()),
+            "d": pd.Series(
+                [False, True, None], index=[0, 1, 3], dtype=pd.BooleanDtype()
+            ),
         },
     ],
 )

diff --git a/python/cudf/cudf/tests/test_series.py b/python/cudf/cudf/tests/test_series.py
@@ -2205,7 +2205,7 @@ def test_series_mixed_dtype_error(dtype):
 @pytest.mark.parametrize("index", [None, [10, 20, 30]])
 def test_series_contains(data, index):
     ps = pd.Series(data, index=index)
-    gs = cudf.from_pandas(ps)
+    gs = cudf.Series(data, index=index)
 
     assert_eq(1 in ps, 1 in gs)
     assert_eq(10 in ps, 10 in gs)
@@ -2326,3 +2326,13 @@ def test_series_count_invalid_param():
     s = cudf.Series([], dtype="float64")
     with pytest.raises(TypeError):
         s.count(skipna=True)
+
+
+def test_bool_series_mixed_dtype_error():
+    ps = pd.Series([True, False, None])
+    # ps now has `object` dtype, which
+    # isn't supported by `cudf`.
+    with pytest.raises(TypeError):
+        cudf.Series(ps)
+    with pytest.raises(TypeError):
+        cudf.from_pandas(ps)