From 5c00e790aec84781a7ae9230ac03c5e87584e409 Mon Sep 17 00:00:00 2001
From: Lawrence Mitchell <lmitchell@nvidia.com>
Date: Fri, 22 Mar 2024 12:37:18 +0000
Subject: [PATCH] Add test of #12243

---
 .../tests/dataframe/test_io_serialization.py  | 38 ++++++++++++++++++-
 python/cudf/cudf/tests/test_parquet.py        | 14 +++++--
 2 files changed, 48 insertions(+), 4 deletions(-)

diff --git a/python/cudf/cudf/tests/dataframe/test_io_serialization.py b/python/cudf/cudf/tests/dataframe/test_io_serialization.py
index 06777c8e6af..911a7f9e865 100644
--- a/python/cudf/cudf/tests/dataframe/test_io_serialization.py
+++ b/python/cudf/cudf/tests/dataframe/test_io_serialization.py
@@ -1 +1,37 @@
-# Copyright (c) 2023, NVIDIA CORPORATION.
+# Copyright (c) 2023-2024, NVIDIA CORPORATION.
+from io import BytesIO
+
+import pandas as pd
+import pyarrow.parquet as pq
+import pytest
+
+import cudf
+from cudf.testing._utils import assert_eq
+
+
+@pytest.mark.parametrize(
+    "index",
+    [range(1, 11), list(range(1, 11)), range(1, 11)[::2]],
+    ids=["RangeIndex", "IntIndex", "StridedRange"],
+)
+@pytest.mark.parametrize("write_index", [False, True, None])
+@pytest.mark.parametrize("empty", [False, True], ids=["nonempty", "empty"])
+def test_dataframe_parquet_roundtrip(index, write_index, empty):
+    if empty:
+        data = {}
+    else:
+        data = {"a": [i * 2 for i in index]}
+    df = cudf.DataFrame(data=data, index=index)
+    pf = pd.DataFrame(data=data, index=index)
+    gpu_buf = BytesIO()
+    cpu_buf = BytesIO()
+
+    df.to_parquet(gpu_buf, index=write_index)
+    pf.to_parquet(cpu_buf, index=write_index)
+    gpu_table = pq.read_table(gpu_buf)
+    cpu_table = pq.read_table(cpu_buf)
+    assert gpu_table.schema.pandas_metadata == cpu_table.schema.pandas_metadata
+
+    gpu_read = cudf.read_parquet(gpu_buf)
+    cpu_read = cudf.read_parquet(cpu_buf)
+    assert_eq(gpu_read, cpu_read)
diff --git a/python/cudf/cudf/tests/test_parquet.py b/python/cudf/cudf/tests/test_parquet.py
index 8b72fe84359..9ba71b28637 100644
--- a/python/cudf/cudf/tests/test_parquet.py
+++ b/python/cudf/cudf/tests/test_parquet.py
@@ -2442,9 +2442,17 @@ def test_parquet_index(pdf, index):
     run_parquet_index(pdf, index)
 
 
-@pytest.mark.parametrize("index", [None, True])
-@pytest.mark.xfail(
-    reason="https://github.com/rapidsai/cudf/issues/12243",
+@pytest.mark.parametrize(
+    "index",
+    [
+        pytest.param(
+            None,
+            marks=pytest.mark.xfail(
+                reason="https://github.com/apache/arrow/issues/40743"
+            ),
+        ),
+        True,
+    ],
 )
 def test_parquet_index_empty(index):
     pdf = pd.DataFrame(index=pd.RangeIndex(0, 10, 1))