n3nash commented on a change in pull request #1009: [HUDI-308] Avoid Renames for tracking state transitions of all actions on dataset URL: https://github.com/apache/incubator-hudi/pull/1009#discussion_r356321928
########## File path: hudi-client/src/main/java/org/apache/hudi/io/HoodieCommitArchiveLog.java ########## @@ -169,7 +180,14 @@ public boolean archiveIfRequired(final JavaSparkContext jsc) throws IOException }).limit(commitTimeline.countInstants() - minCommitsToKeep)); } - return instants; + // For archiving and cleaning instants, we need to include intermediate state files if they exist + HoodieActiveTimeline rawActiveTimeline = new HoodieActiveTimeline(metaClient, false); + Map<Pair<String, String>, List<HoodieInstant>> groupByTsAction = rawActiveTimeline.getInstants() + .collect(Collectors.groupingBy(x -> Pair.of(x.getTimestamp(), + x.getAction().equals(HoodieTimeline.COMPACTION_ACTION) ? HoodieTimeline.COMMIT_ACTION : x.getAction()))); + + return instants.flatMap(hoodieInstant -> Review comment: What I'm saying is group by only the instants chosen to archive, all types of actions is fine.. ---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services