Skip to content

ARROW-10462: [Python] Fix usage of fsspec in ParquetDataset causing path issue on Windows #8539

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

Closed
jorisvandenbossche wants to merge 3 commits into from
Closed

ARROW-10462: [Python] Fix usage of fsspec in ParquetDataset causing path issue on Windows #8539

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
1ac75f4
ARROW-xxxx: [CI][Python] Run fsspec tests on Windows
jorisvandenbossche Oct 27, 2020
647c855
Add test for fsspec piece path issue on Windows
jorisvandenbossche Nov 2, 2020
1b4dce0
don't fall back to pyarrow local filesystem for fsspec local filesystem
jorisvandenbossche Nov 2, 2020
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
1 change: 1 addition & 0 deletions ci/appveyor-cpp-setup.bat
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -65,6 +65,7 @@ if "%JOB%" NEQ "Build_Debug" (
"ninja" ^
"nomkl" ^
"pandas" ^
"fsspec" ^
"python=%PYTHON%" ^
|| exit /B
)
Expand Down
4 changes: 0 additions & 4 deletions python/pyarrow/filesystem.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -446,10 +446,6 @@ def _ensure_filesystem(fs):
for mro in inspect.getmro(fs_type):
if mro.__name__ == 'S3FileSystem':
return S3FSWrapper(fs)
# In case its a simple LocalFileSystem (e.g. dask) use native arrow
# FS
elif mro.__name__ == 'LocalFileSystem':
return LocalFileSystem._get_instance()
Copy link
Member Author

@jorisvandenbossche jorisvandenbossche Nov 2, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The reason this now gives troubles (while this code was already there for more than 2 years), is because with pyarrow >= 2.0, fsspec will no longer inherit from pyarrow.filesystem.FileSystem, and so this path was only taken if the passed filesystem was not a pyarrow subclass (so before, in pyarrow.fs, we simply didn't take this code path for fsspec filesystems)

Copy link
Member Author

@jorisvandenbossche jorisvandenbossche Nov 2, 2020

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think it is fine to simply remove this code, as in practice this was also never run anymore with previous pyarrow releases. I suppose this was only ever reached with old fsspec versions that did not yet subclass pyarrow.


if "fsspec" in sys.modules:
fsspec = sys.modules["fsspec"]
Expand Down
16 changes: 16 additions & 0 deletions python/pyarrow/tests/test_parquet.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -4339,3 +4339,19 @@ def test_parquet_dataset_new_filesystem(tempdir):
dataset = pq.ParquetDataset('.', filesystem=filesystem)
result = dataset.read()
assert result.equals(table)


def test_parquet_dataset_partitions_piece_path_with_fsspec(tempdir):
# ARROW-10462 ensure that on Windows we properly use posix-style paths
# as used by fsspec
fsspec = pytest.importorskip("fsspec")
filesystem = fsspec.filesystem('file')
table = pa.table({'a': [1, 2, 3]})
pq.write_table(table, tempdir / 'data.parquet')

# pass a posix-style path (using "/" also on Windows)
path = str(tempdir).replace("\\", "/")
dataset = pq.ParquetDataset(path, filesystem=filesystem)
# ensure the piece path is also posix-style
expected = path + "/data.parquet"
assert dataset.pieces[0].path == expected