Allow pd.ArrowDtype in cudf.from_pandas (rapidsai#13465)

closes rapidsai#13003

Allow pandas objects with `ArrowDtype` to be accepted by `cudf.from_pandas`

Authors:
  - Matthew Roeschke (https://github.com/mroeschke)
  - GALI PREM SAGAR (https://github.com/galipremsagar)

Approvers:
  - Bradley Dice (https://github.com/bdice)
  - GALI PREM SAGAR (https://github.com/galipremsagar)

URL: rapidsai#13465

python/cudf/cudf/core/column/column.py

@@ -1990,10 +1990,13 @@ def as_column(
         return col
 
     elif isinstance(arbitrary, (pd.Series, pd.Categorical)):
-        if isinstance(arbitrary, pd.Series) and isinstance(
-            arbitrary.array, pd.core.arrays.masked.BaseMaskedArray
-        ):
-            return as_column(arbitrary.array)
+        if isinstance(arbitrary, pd.Series):
+            if isinstance(
+                arbitrary.array, pd.core.arrays.masked.BaseMaskedArray
+            ):
+                return as_column(arbitrary.array)
+            elif PANDAS_GE_150 and isinstance(arbitrary.dtype, pd.ArrowDtype):
+                return as_column(pa.array(arbitrary.array, from_pandas=True))
         if is_categorical_dtype(arbitrary):
             data = as_column(pa.array(arbitrary, from_pandas=True))
         elif is_interval_dtype(arbitrary.dtype):

python/cudf/cudf/tests/test_dataframe.py

@@ -1694,18 +1694,30 @@ def test_nonmatching_index_setitem(nrows):
     assert_eq(gdf["c"].to_pandas(), gdf_series.to_pandas())
 
 
-def test_from_pandas():
-    df = pd.DataFrame({"x": [1, 2, 3]}, index=[4.0, 5.0, 6.0])
+@pytest.mark.parametrize(
+    "dtype",
+    [
+        "int",
+        pytest.param(
+            "int64[pyarrow]",
+            marks=pytest.mark.skipif(
+                not PANDAS_GE_150, reason="pyarrow support only in >=1.5"
+            ),
+        ),
+    ],
+)
+def test_from_pandas(dtype):
+    df = pd.DataFrame({"x": [1, 2, 3]}, index=[4.0, 5.0, 6.0], dtype=dtype)
     gdf = cudf.DataFrame.from_pandas(df)
     assert isinstance(gdf, cudf.DataFrame)
 
-    assert_eq(df, gdf)
+    assert_eq(df, gdf, check_dtype="pyarrow" not in dtype)
 
     s = df.x
     gs = cudf.Series.from_pandas(s)
     assert isinstance(gs, cudf.Series)
 
-    assert_eq(s, gs)
+    assert_eq(s, gs, check_dtype="pyarrow" not in dtype)
 
 
 @pytest.mark.parametrize("dtypes", [int, float])