Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Relax inconsistent schema handling in dask_cudf.read_parquet #17554

Open
wants to merge 6 commits into
base: branch-25.02
Choose a base branch
from
Open

Relax inconsistent schema handling in dask_cudf.read_parquet #17554

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
6a8bdfb
relax logic in set_object_dtypes_from_pa_schema
rjzamora Dec 9, 2024
45407c0
Merge branch 'branch-25.02' into relax-pq-schema-translation
rjzamora Dec 9, 2024
7cf2e45
Merge branch 'branch-25.02' into relax-pq-schema-translation
rjzamora Dec 10, 2024
27a59ce
Merge remote-tracking branch 'upstream/branch-25.02' into relax-pq-sc…
rjzamora Dec 11, 2024
7bd106b
Merge remote-tracking branch 'upstream/branch-25.02' into relax-pq-sc…
rjzamora Dec 12, 2024
e2db4d0
remove None block
rjzamora Dec 12, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
20 changes: 7 additions & 13 deletions python/dask_cudf/dask_cudf/_legacy/io/parquet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -42,7 +42,7 @@ def _create_dd_meta(cls, dataset_info, **kwargs):
kwargs = dataset_info.get("kwargs", {})
set_object_dtypes_from_pa_schema(
meta_cudf,
kwargs.get("schema", None),
kwargs.get("dataset", {}).get("schema", None),
)

return meta_cudf
Expand Down Expand Up @@ -434,18 +434,12 @@ def set_object_dtypes_from_pa_schema(df, schema):
# pyarrow schema.
if schema:
for col_name, col in df._data.items():
if col_name is None:
# Pyarrow cannot handle `None` as a field name.
# However, this should be a simple range index that
# we can ignore anyway
continue
typ = cudf_dtype_from_pa_type(schema.field(col_name).type)
if (
col_name in schema.names
and not isinstance(typ, (cudf.ListDtype, cudf.StructDtype))
and isinstance(col, cudf.core.column.StringColumn)
):
df._data[col_name] = col.astype(typ)
if col_name in schema.names:
typ = cudf_dtype_from_pa_type(schema.field(col_name).type)
if not isinstance(
typ, (cudf.ListDtype, cudf.StructDtype)
) and isinstance(col, cudf.core.column.StringColumn):
df._data[col_name] = col.astype(typ)


def read_parquet(path, columns=None, **kwargs):
Expand Down
27 changes: 27 additions & 0 deletions python/dask_cudf/dask_cudf/io/tests/test_parquet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -492,6 +492,33 @@ def test_create_metadata_file_inconsistent_schema(tmpdir):
dd.assert_eq(ddf1.compute(), ddf2.compute())


def test_read_inconsistent_schema(tmpdir):
records = [
{"id": 123, "text": "foo"},
{
"text": "bar",
"meta1": [{"field1": "cat"}],
"id": 456,
},
]
columns = ["text", "id"]
pd.DataFrame(records[:1]).to_parquet(tmpdir / "part.0.parquet")
pd.DataFrame(records[1:]).to_parquet(tmpdir / "part.1.parquet")
# Check that cuDF and Dask cuDF match
dd.assert_eq(
cudf.read_parquet(
tmpdir, columns=columns, allow_mismatched_pq_schemas=True
),
dask_cudf.read_parquet(tmpdir, columns=columns),
check_index=False,
)
# Check that "pandas" and "cudf" backends match
dd.assert_eq(
dd.read_parquet(tmpdir, columns=columns),
dask_cudf.read_parquet(tmpdir, columns=columns),
)


@pytest.mark.parametrize(
"data",
[
Expand Down
Loading