rangareddy commented on issue #17278:
URL: https://github.com/apache/hudi/issues/17278#issuecomment-5365811722

   This issue was reviewed as part of the JIRA-migrated backlog triage 
(HUDI-8283).
   
   **Findings: partially delivered. Two of the paths named have moved onto the 
file group reader; the rest have not.**
   
   **Done:**
   
   - Metadata table - reads through `HoodieFileGroupReader` / 
`HoodieLsmFileGroupReader` for key lookups (`HoodieBackedTableMetadata:617`, 
`:629`), via PR #13544 (HUDI-9592), which also deleted 
`HoodieMetadataLogRecordReader`.
   - CDC - moved onto the FG reader via #13444 (HUDI-9526) and #13694 
(HUDI-9695).
   
   **Not done:**
   
   - Compaction and clustering still do not read through it. The sibling that 
tracks this for the metadata table, #17321 (HUDI-8636), has no code referencing 
it at all.
   - There is no full-file-slice read capability, which compaction needs: 
`readSliceAndFilterByKeysIntoList` returns an `EmptyIterator` when given no 
keys. That is tracked as #17393 (HUDI-9130) and is still open.
   - Global-index partition-change reads were not separately verified.
   
   So the duplicate-code concern in the description is genuinely reduced but 
not resolved. Given the dependency chain (#17393 unblocks #17321, which is part 
of this), it is worth linking those explicitly so the ordering is visible.
   
   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: [email protected]

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

Reply via email to