scxwhite commented on code in PR #6670:
URL: https://github.com/apache/hudi/pull/6670#discussion_r973539801
##########
hudi-client/hudi-client-common/src/main/java/org/apache/hudi/config/HoodieCompactionConfig.java:
##########
@@ -106,6 +106,12 @@ public class HoodieCompactionConfig extends HoodieConfig {
.withDocumentation("Only if the log file size is greater than the
threshold in bytes,"
+ " the file group will be compacted.");
+ public static final ConfigProperty<Long>
COMPACTION_LOG_FILE_LENGTH_THRESHOLD = ConfigProperty
+ .key("hoodie.compaction.logfile.length.threshold")
+ .defaultValue(0L)
Review Comment:
> Got it. That makes sense. Then do we even need this config? The new
compaction strategy prioritizes the compaction of file groups with more log
files, and should include all file groups for compaction nevertheless.
The default value (0) can meet most of the user's business scenarios.
However, for some deep users, they may feel that if the number of small files
in delta log (such as three) is within the acceptable range of reading
performance, it is unnecessary to trigger the compression operation that has a
great impact on throughput performance. Therefore, I think it is necessary to
retain this configuration.
--
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]
