You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hive.apache.org by Lawulu <ya...@foxmail.com> on 2022/01/21 15:16:14 UTC

Which Hive version should I choose？

HI!




I have used hive2 for several years, and i worked well.

Recently I setup a new Hadoop cluster using Hdp 3.x（Hadoop 3.1，Hive 3.1, Tez 0.9）




But there are some bugs, cbo/left join/vectorized execution..



Bugs are reported by some other people to the community, but most of&nbsp; resolved bugs are fixed on Hive 4.0 instead of Hive 3.x.


Also there are bugs which can reproduce on my cluster are not resolved. such like :


[HIVE-24907] Wrong results with LEFT JOIN and subqueries with UNION and GROUP BY - ASF JIRA
[HIVE-24239] Reduce side chain of Map join produce wrong result - ASF JIRA


It is very painful for me to maintain such a cluster.





I've tried to downgrade Hive3 to Hive2, but got some issues when&nbsp;using Tez engine.

Is it possible to run hive2 on Hadoop3 using Tez engine?

Another choice is upgrading Hive3 to Hive4,&nbsp;although it hasn't been released yet.





Does anyone have a similar situation? Could you give me some advice?















&nbsp;