[I] HoodieFileGroupReader should already know the schema that it is reading [hudi]

["hudi-bot (via GitHub)" <gi...@apache.org>]

hudi-bot opened a new issue, #17020:
URL: https://github.com/apache/hudi/issues/17020

   The current implementation in HudiPageSourceProvider requires us to pass a requestedSchema into HoodieFileGroupReader. In theory, all schema (evolution) related handling should be handled by HoodieFileGroupReader and all HudiPageSourceProvider needs to do is pass the required columns that needs to be read (for column pruning/projection)
   
    
   
   A followup cleanup within HoodieFileGroupReader is required to expose this interface and implement this functionality.
   
   ## JIRA info
   
   - Link: https://issues.apache.org/jira/browse/HUDI-9429
   - Type: Task
   - Epic: https://issues.apache.org/jira/browse/HUDI-2687


-- 
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: commits-unsubscribe@hudi.apache.org.apache.org

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

Re: [I] HoodieFileGroupReader should already know the schema that it is reading [hudi]

["rangareddy (via GitHub)" <gi...@apache.org>]

rangareddy commented on issue #17020:
URL: https://github.com/apache/hudi/issues/17020#issuecomment-5391446339

   This issue was reviewed as part of the JIRA-migrated backlog triage (HUDI-9429).
   
   **Findings: not done, and broader than it looks.**
   
   **Where this lives now.** The Trino Hudi connector was migrated into this repository by commit `c3c936790727`, *"feat(trino): Migrate the Trino-Hudi connector into the Hudi repo (RFC-105)"* (#18837, 2026-07-27). It is the `hudi-trino/` module, with its own CI in `.github/workflows/hudi_trino_ci.yml`, `hudi_trino_compat.yml` and `hudi_trino_e2e.yml`. This ticket was filed when the connector lived in `trinodb/trino`, so it reads as out of scope here - it is not, and can now be worked in this repo.
   
   The ask here is that `HoodieFileGroupReader` own all schema-evolution handling, so `HudiPageSourceProvider` only passes the columns it needs projected. That is a change to the file group reader's interface in `hudi-common`, not a connector-local cleanup - so it affects every FG-reader caller, not just Trino.
   
   Worth sequencing against #17075 (HUDI-9561), which documents concrete schema-evolution gaps in the FG reader across Hive, Java and Flink. Moving more schema responsibility into the reader while those gaps are open would push the failures into the reader rather than remove them.
   
   Keeping this open.
   


-- 
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: commits-unsubscribe@hudi.apache.org

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