[REVIEW] First pass of pd.read_orc changes in tests

python/cudf/cudf/testing/_utils.py

@@ -10,11 +10,14 @@
 import cupy
 import numpy as np
 import pandas as pd
+import pyarrow as pa
+import pyarrow.orc
 import pytest
 from pandas import testing as tm
 
 import cudf
 from cudf._lib.null_mask import bitmask_allocation_size_bytes
+from cudf.core._compat import PANDAS_GE_100
 from cudf.core.column.timedelta import _unit_to_nanoseconds_conversion
 from cudf.utils import dtypes as dtypeutils
 
@@ -379,6 +382,13 @@ def _create_pandas_series(data=None, index=None, dtype=None, *args, **kwargs):
     return pd.Series(data=data, index=index, dtype=dtype, *args, **kwargs)
 
 
+def _pandas_read_orc(fname, columns=None):
+    if PANDAS_GE_100:
+        return pd.read_orc(fname, columns=columns)
+    else:
+        return pa.orc.ORCFile(fname).read(columns=columns).to_pandas()
+
+
 parametrize_numeric_dtypes_pairwise = pytest.mark.parametrize(
     "left_dtype,right_dtype",
     list(itertools.combinations_with_replacement(NUMERIC_TYPES, 2)),

python/cudf/cudf/tests/test_gcs.py

@@ -5,12 +5,10 @@
 
 import numpy as np
 import pandas as pd
-import pyarrow as pa
-import pyarrow.orc
 import pytest
 
 import cudf
-from cudf.testing._utils import assert_eq
+from cudf.testing._utils import _pandas_read_orc, assert_eq
 
 gcsfs = pytest.importorskip("gcsfs")
 
@@ -71,5 +69,5 @@ def mock_open(*args, **kwargs):
     monkeypatch.setattr(gcsfs.core.GCSFileSystem, "open", mock_open)
     gdf.to_orc(f"gcs://{gcs_fname}")
 
-    got = pa.orc.ORCFile(local_filepath).read().to_pandas()
+    got = _pandas_read_orc(local_filepath)
     assert_eq(pdf, got)

python/cudf/cudf/tests/test_hdfs.py

@@ -8,10 +8,9 @@
 import pandas as pd
 import pyarrow as pa
 import pytest
-from pyarrow import orc
 
 import cudf
-from cudf.testing._utils import assert_eq
+from cudf.testing._utils import _pandas_read_orc, assert_eq
 
 if not os.environ.get("RUN_HDFS_TESTS"):
     pytestmark = pytest.mark.skip("Env not configured to run HDFS tests")
@@ -212,7 +211,7 @@ def test_read_orc(datadir, hdfs, test_url):
         hd_fpath = f"hdfs://{basedir}/file.orc"
 
     got = cudf.read_orc(hd_fpath)
-    expect = orc.ORCFile(buffer).read().to_pandas()
+    expect = _pandas_read_orc(buffer)
     assert_eq(expect, got)
 
 
@@ -232,7 +231,7 @@ def test_write_orc(pdf, hdfs, test_url):
 
     assert hdfs.exists(f"{basedir}/test_orc_writer.orc")
     with hdfs.open(f"{basedir}/test_orc_writer.orc", mode="rb") as f:
-        got = orc.ORCFile(f).read().to_pandas()
+        got = _pandas_read_orc(f)
 
     assert_eq(pdf, got)