matrixsparse opened a new pull request, #3296:
URL: https://github.com/apache/fluss/pull/3296

   ## Summary
   
   This is a follow-up fix for PR #3208.
   
   For primary-key tables in batch mode, when no lake snapshot exists, the 
previous fallback logic called `initPartitionedSplits()` / 
`initNonPartitionedSplits()`, which internally invokes 
`getSnapshotAndLogSplits()`. This method may generate mixed split types 
(`HybridSnapshotLogSplit` for buckets with KV snapshots + `LogSplit` for 
buckets without), and the Flink connector does not support merging these two 
split types.
   
   This fix replaces the fallback path with `initLogTablePartitionSplits()` / 
`getLogSplit()` to generate uniform `LogSplit` types for all buckets, aligning 
with Spark's `planFallbackPartitions()` behavior.
   
   ### Changes
   - Partitioned tables: `initPartitionedSplits(partitions)` → 
`initLogTablePartitionSplits(partitions)`
   - Non-partitioned tables: `initNonPartitionedSplits()` → `getLogSplit(null, 
null)`


-- 
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