You are viewing a plain text version of this content. The canonical link for it is here.

Posted to github@arrow.apache.org by GitBox <gi...@apache.org> on 2020/10/09 12:59:59 UTC

[GitHub] [arrow] jorisvandenbossche commented on pull request #8305: ARROW-9782: [C++][Dataset] More configurable Dataset writing

jorisvandenbossche commented on pull request #8305:
URL: https://github.com/apache/arrow/pull/8305#issuecomment-706166335


   > FileSystemDataset::Write now parallelizes across scan tasks rather than fragments so there will be no difference in performance/written files even if we create a single in-memory fragment, so I changed this to create a single fragment since it's simpler
   
   There might be no difference, but I think the user should still be able to control how many files are created. Because now whathever you pass, it's always consolidated into a single file? (or one file per partitioning) 
   
   Also, it seems that reading/writing a dataset does not preserve the files? (so if we discover a dataset with multiple files, we write it as a single file?)
   
   


----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org