datafusion icon indicating copy to clipboard operation
datafusion copied to clipboard

feat(cli): support external tables on multiple locations

Open fpetkovski opened this issue 3 months ago • 4 comments

Which issue does this PR close?

  • Related to https://github.com/apache/datafusion/issues/16303

Rationale for this change

The commit adds support for creating external tables that can reference multiple file locations. This is useful for querying files from multiple folders, or for scoping down an external table to specific files inside a single folder.

The change is implemented by splitting the location string on the , delimiter as suggested in the original issue.

Are these changes tested?

There is an added unit test for this change. I also did manual testing against the branch.

Are there any user-facing changes?

This feature does make a backwards compatible user-facing change, which should require documentation updates.

fpetkovski avatar Sep 21 '25 23:09 fpetkovski

Thank you for this PR @fpetkovski -- I plan to review it carefully in the next day or two

alamb avatar Sep 29 '25 17:09 alamb

Thank you @alamb. I am aware of the CI issues and plan to address them. I was hoping to first get feedback on the overall direction.

fpetkovski avatar Sep 29 '25 19:09 fpetkovski

@alamb just wanted to ask if you've had a chance to look at the PR?

I was hoping to add this and follow up with additional features like glob and brace expansion syntax.

fpetkovski avatar Oct 15 '25 19:10 fpetkovski

Sorry @fpetkovski -- I have been totally slammed trying to keep up with reviews recently (partly due to how awesome the arrow 57 release is shaping up to be) I will try and check this out soon

alamb avatar Oct 17 '25 19:10 alamb