zhuqi-lucas commented on issue #10381:
URL: https://github.com/apache/arrow-rs/issues/10381#issuecomment-5043914947

   Thanks @vustef! Good catch on the metadata-reuse case — that would be a 
nasty state-dependent bug where the same "read with row numbers" call succeeds 
or fails depending on whether the metadata was decoded fresh or reused from a 
prior no-row-number read.
   
   Agreed, best-effort sequential-fill on read is the right shape:
   
   - If `RowGroup.ordinal` is set in metadata → preserve it (needed for 
encryption chunk key lookup).
   - If missing → sequential-fill unconditionally, regardless of whether row 
numbers are enabled.
   - Never error on mixed/missing ordinals.
   
   This gives us the "lenient on reads, strict on writes" behavior 
consistently, and any downstream consumer (row numbers, encryption, anything 
future) sees complete ordinals no matter how the metadata was obtained.
   
   Will update the PR accordingly.


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to