[
https://issues.apache.org/jira/browse/HIVE-26482?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Work on HIVE-26482 started by Zsolt Miskolczi.
----------------------------------------------
> Create a unit test checking compaction output file names on a partitioned
> table
> -------------------------------------------------------------------------------
>
> Key: HIVE-26482
> URL: https://issues.apache.org/jira/browse/HIVE-26482
> Project: Hive
> Issue Type: Test
> Components: Hive
> Reporter: Zsolt Miskolczi
> Assignee: Zsolt Miskolczi
> Priority: Major
>
> Compaction output directories' writeIds only reflect the writeIds of the
> deltas it compacts, and not the max write id of the table
> Example:
> Pre-compaction...
> {code:java}
> Partition p=1 contains:
> delta_1_1
> delta_2_2
> partition p=2 contains
> delta_3_3
> delta_4_4
> {code}
> After minor compaction...
> {code:java}
> Partition p=1 contains:
> delta_1_2
> partition p=2 contains
> delta_3_4
> {code}
> AFAIK there are no unit tests that reflect this.
> TestTxnCommands2#testFullACIDAbortWithManyPartitions is a good template to
> start with.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
