[ARROW-5501] [R] Reorganize read/write file/stream functions

• Type: Improvement
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 0.17.0
• Component/s: R
• Labels: pull-request-available
• External issue URL: https://github.com/apache/arrow/issues/21949
• Assignee: Neal Richardson
• Reporter: Neal Richardson
• Created: 03/Jun/19 22:37
• Updated: 11/Jan/23 07:40
• Resolved: 02/Apr/20 21:42

Description

read_feather and write_feather exist, and there is also write_arrow. But no read_arrow.

Some questions (which go beyond just R): There's talk of a "feather 2.0", i.e. "just" serializing the IPC format (which IIUC is what write_arrow does). Are we going to continue to call the file format "Feather", and possibly continue supporting the "feather 1.0" format as a subset/special case? Or will "feather" mean this limited format and "arrow" be the name of the full-featured file?

In terms of this issue, should write_arrow be folded into write_feather and there be an argument for indicating which version to write? Or should the distinction be maintained, and we need to add a read_arrow() function?

Issue Links

• is related to: ARROW-7520 [R] Writing many batches causes a crash
• links to: GitHub Pull Request #6771