[
https://issues.apache.org/jira/browse/HIVE-19867?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16516053#comment-16516053
]
Steve Yeom edited comment on HIVE-19867 at 6/22/18 5:40 AM:
------------------------------------------------------------
I found the patch 03 does not address the cases of concurrent writes to a
table.
If writeId for transactional stats is saved into TBLS/PARTITIONS, then
it can be used to figure out concurrent writes by comparing a write with its
writeIdList.
Simply by checking it is in the list or not.
I.e., if we assume two concurrent NSERTs are subsequently committed and updated
stats
of, let's say, table1. Then first INSERT's writeId is now saved and so can be
used to see if it is
in second INSERT's writeIdList when second INSERT comes in. Also second
INSERT's writeId
can be checked to see it is in first INSERT's writeIdList that is saved in
TBLS/PARTITIONS.
If each writeId is contained in the other's writeIdList, we can say it is
concurrent INSERT case
and turn the COLUMN_STATS_ACCURATE flag off.
was (Author: steveyeom2017):
I found the patch 03 does not address the cases of concurrent writes to a
table.
If writeId for transactional stats is saved into TBLS/PARTITIONS, then
it can be used to figure out concurrent writes by comparing a write with its
writeIdList.
Simply by checking it is in the list or not.
> Test and verify Concurrent INSERTS
> ------------------------------------
>
> Key: HIVE-19867
> URL: https://issues.apache.org/jira/browse/HIVE-19867
> Project: Hive
> Issue Type: Sub-task
> Components: Transactions
> Affects Versions: 4.0.0
> Reporter: Steve Yeom
> Assignee: Steve Yeom
> Priority: Major
> Fix For: 4.0.0
>
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
