danny0405 commented on a change in pull request #4446:
URL: https://github.com/apache/hudi/pull/4446#discussion_r780733934
##########
File path:
hudi-client/hudi-java-client/src/main/java/org/apache/hudi/table/action/commit/BaseJavaCommitActionExecutor.java
##########
@@ -90,27 +90,29 @@ public BaseJavaCommitActionExecutor(HoodieEngineContext
context,
public HoodieWriteMetadata<List<WriteStatus>> execute(List<HoodieRecord<T>>
inputRecords) {
HoodieWriteMetadata<List<WriteStatus>> result = new
HoodieWriteMetadata<>();
- WorkloadProfile profile = null;
+ WorkloadProfile inputProfile = null;
if (isWorkloadProfileNeeded()) {
- profile = new WorkloadProfile(buildProfile(inputRecords));
- LOG.info("Workload profile :" + profile);
+ inputProfile = new WorkloadProfile(buildProfile(inputRecords));
+ LOG.info("Input workload profile :" + inputProfile);
+ }
+
+ final Partitioner partitioner = getPartitioner(inputProfile);
+ try {
+ WorkloadProfile executionProfile =
partitioner.getExecutionWorkloadProfile();
+ LOG.info("Execution workload profile :" + inputProfile);
+ saveWorkloadProfileMetadataToInflight(executionProfile, instantTime);
Review comment:
And why we must use the execution profile here ? I know the original
profile also works only for bloomfilter index but we should fix the profile
building instead of fetch it from the partitioner, if we have a way to
distinguish between `INSERT`s and `UPDATE`s before write.
--
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]
