JingsongLi commented on PR #11285:
URL: https://github.com/apache/paimon/pull/11285#issuecomment-7029930053

   **[P1] Resolve the historical end snapshot before using a global index.** In 
a Parquet Data Evolution table, snapshot 1 contained `(100,10)` and snapshot 2 
appended `(200,20)`. Snapshot 3 updated that row to `(200,30)`, and I built a 
complete B-tree on `v` over the current snapshot. DIFF 1→2 with `v=20` returned 
zero rows with default index settings and returned `(200,20)` with 
`global-index.enabled=false`. The current index successfully found `(200,30)` 
in an ordinary latest-snapshot read.
   
   `DataEvolutionBatchScan.planEager` evaluates the index before the DIFF 
scanner selects the historical end snapshot. 
`DataEvolutionGlobalIndexScanner.create` resolves latest in this case, allowing 
newer index data to discard rows from an older snapshot. This persists 
independently of the earlier row-range finding. Please bypass scalar index 
pruning for DIFF, or resolve the end snapshot before evaluating its index and 
coverage.
   


-- 
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