Github user SangeetaGulia commented on a diff in the pull request:
https://github.com/apache/carbondata/pull/1661#discussion_r157135341
--- Diff:
integration/spark-common/src/main/scala/org/apache/spark/util/PartitionUtils.scala
---
@@ -113,7 +113,8 @@ object PartitionUtils {
}
if (partitionId == 0) {
- partitionInfo.addPartition(splitInfo.size)
+ val addedPartitionList =
PartitionUtils.getListInfo(splitInfo.mkString(","))
--- End diff --
Add partition does support String group. The problem in this bug is, when
you are trying to add these two partitions "'SAUDI
ARABIA,(VIETNAM,RUSSIA,UNITED KINGDOM,UNITED STATES)'" then splitInfo variable
takes it as one partition and increment partition count by 1. To get the
correct partitions out of it, we need to extract the list out of the splitList.
As this is a sequence based test case, where the problem is encountered
when we "show partition" after "alter table to add partition", So is it
alright to add its test case to TestAlterPartitionTable.scala?
---
