Antoine Pitrou commented on ARROW-7102:

Hmm, so, to make sure there's no misunderstanding, we have been developing a 
C++ filesystem layer that's pretty much functional now, with local, mock and S3 

There's also a Python wrapping layer for the C++ filesystem layer. It's in 
{{pyarrow.fs}}, not {{pyarrow.filesystem}} which we now consider as legacy.

Going forward, two things could be useful:
* Make a {{fsspec}} wrapper for {{pyarrow.fs}}
* Make a {{pyarrow.fs}} wrapper for {{fsspec}}

You can find the C++ filesystem API doc here:
The Python wrappers are not documented on the website yet. But there are 
docstrings that you can also find here:

> Make filesystem wrappers compatible with fsspec
> -----------------------------------------------
>                 Key: ARROW-7102
>                 URL: https://issues.apache.org/jira/browse/ARROW-7102
>             Project: Apache Arrow
>          Issue Type: Improvement
>          Components: Python
>            Reporter: Tom Augspurger
>            Priority: Major
>              Labels: FileSystem
> [fsspec|https://filesystem-spec.readthedocs.io/en/latest] defines a common 
> API for a variety filesystem implementations. I'm proposing a FSSpecWrapper, 
> similar to S3FSWrapper, that works with any fsspec implementation.
> Right now, pyarrow has a pyarrow.filesystems.S3FSWrapper, which is specific 
> to s3fs. 
> [https://github.com/apache/arrow/blob/21ad7ac1162eab188a1e15923fb1de5b795337ec/python/pyarrow/filesystem.py#L320].
>  This implementation could be removed entirely once an FSSPecWrapper is done, 
> or kept as an alias if it's part of the public API.
> This is realted to ARROW-3717, which requested a GCSFSWrapper for working 
> with google cloud storage.

This message was sent by Atlassian Jira

Reply via email to