[jira] [Commented] (HIVE-25698) Hive column update performance too low when table partition over 700

["JungHyun An (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-25698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17465304#comment-17465304 ] 

JungHyun An commented on HIVE-25698:
------------------------------------

[~kgyrtkirk] 

Currently, our HIVE has tables containing hundreds of TB of data. The "alter table 'table_name' change 'cloumn' " statement was executed on the corresponding table, and the execution time was much slower than the execution time of the existing HIVE version 1. I would like to know if there are any architectural changes in HIVE 1 and HIVE 3 related to that feature.

> Hive column update performance too low when table partition over 700
> --------------------------------------------------------------------
>
>                 Key: HIVE-25698
>                 URL: https://issues.apache.org/jira/browse/HIVE-25698
>             Project: Hive
>          Issue Type: Bug
>          Components: Clients, Server Infrastructure
>    Affects Versions: 3.1.1
>         Environment: CentOS 7.8 
> Hadoop 3.1.1
> Impala 3.4.0
>            Reporter: JungHyun An
>            Priority: Minor
>             Fix For: All Versions
>
>
> Now we using hive 3.1.1
>  
> Currently in our hive we have tables with hundreds of partitions and hundreds of gigabytes of data.
>  
> When updating the column information of the corresponding table, it was confirmed that the performance was several tens of times slower than the Hive 1.1 version of the existing CDH.
>  
> I would like to ask if there is any architectural change that makes column updates slower than Hive 1 in Hive 3 and later versions.
>  
> Thank you.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)