Github user manishgupta88 commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/1812#discussion_r184403180
--- Diff:
integration/spark2/src/main/scala/org/apache/spark/sql/execution/command/preaaggregate/PreAggregateListeners.scala
---
@@ -566,6 +567,11 @@ object AlterPreAggregateTableCompactionPostListener
extends OperationEventListen
val compactionType = compactionEvent.carbonMergerMapping.campactionType
val carbonLoadModel = compactionEvent.carbonLoadModel
val sparkSession = compactionEvent.sparkSession
+ val segmentIds = if (compactionType == CompactionType.CUSTOM) {
+ Some(compactionEvent.carbonMergerMapping.validSegments.map(x =>
x.getSegmentNo).toList)
--- End diff --
Custom compaction should be done only for the given table in the Alter SQL.
Custom compaction executed on main table should not be applied on child
tables/datamaps as one to one mapping of segments might not be there between
main table and its child tables.
Custom compaction can be done for child tables/datamaps directly like main
table by specifying the child table/datamap name in the alter SQL.
---
