rangareddy commented on issue #17298:
URL: https://github.com/apache/hudi/issues/17298#issuecomment-5365817725

   This issue was reviewed as part of the JIRA-migrated backlog triage 
(HUDI-8423).
   
   **Findings: partially delivered. The read side now supports the record 
merger; the write side does not.**
   
   **Read side - done.** 
`hudi-spark-datasource/hudi-spark-common/src/main/scala/org/apache/hudi/cdc/CDCFileGroupIterator.scala`
 resolves the merger from the reader context and threads it through the merge:
   
   - `:130` - `private lazy val recordMerger: HoodieRecordMerger = 
readerContext.getRecordMerger().get()`
   - `:96` - branches on `recordMerger.getMergingStrategy == 
PAYLOAD_BASED_MERGE_STRATEGY_UUID`
   - `:105-106` - passes `readerContext.getMergeMode`, `isPartialMergeEnabled`, 
`Option.of(recordMerger)`, `payloadClass` and `partialUpdateModeOpt` into the 
buffered merge
   
   That came in via #13444 (HUDI-9526, *"Use HoodieFileGroupReader throughout 
the CDC flow"*, `fffa3efaab8d`) and #13694 (HUDI-9695, *"Use 
BufferedRecordMerger for spark cdc reader"*, `7aa92cb32038`) - neither keyed to 
this ticket, which is why it still reads as untouched.
   
   **Write side - not done.** 
`hudi-client/hudi-client-common/src/main/java/org/apache/hudi/io/cdc/HoodieCDCLogger.java`,
 and its `HoodieNativeCDCLogger` / `HoodieAvroNativeCDCLogger` implementations, 
contain no reference to `RecordMergeMode`, `HoodieRecordMerger` or a payload 
class. CDC log-block generation is still merger-unaware.
   
   So the remaining scope for this ticket is the write half only. Worth 
narrowing the description to that, so it does not get re-triaged as wholly 
outstanding.
   
   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