KanakaKumar commented on a change in pull request #3262: [CARBONDATA-3415]
Merge index is not working for partition table. Merge index for partition table
is taking significantly longer time than normal table.
URL: https://github.com/apache/carbondata/pull/3262#discussion_r291466837
##########
File path:
integration/spark-common/src/main/scala/org/apache/spark/rdd/CarbonMergeFilesRDD.scala
##########
@@ -130,9 +135,28 @@ class CarbonMergeFilesRDD(
extends CarbonRDD[String](ss, Nil) {
override def internalGetPartitions: Array[Partition] = {
- segments.zipWithIndex.map {s =>
- CarbonMergeFilePartition(id, s._2, s._1)
- }.toArray
+ if (carbonTable.isHivePartitionTable) {
+ // in case of partition table make rdd partitions per partition of the
carbon table
+ val partitionPaths: java.util.Map[String, String] = new
java.util.HashMap()
+ segments.foreach(segment => {
+ val partitionSpecs = SegmentFileStore.getPartitionSpecs(segment,
carbonTable.getTablePath)
Review comment:
1) SegmentFileStore.getPartitionSpecs is reading table status file for each
segment. So, read & send the load details from outside to avoid this.
2) partitionLoc may be same for each segment. So, at the end only one
segment per partition will remain in map. Please correct to use list based map
instead
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
