[
https://issues.apache.org/jira/browse/KYLIN-4883?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17744425#comment-17744425
]
Shao Feng Shi commented on KYLIN-4883:
--------------------------------------
If you can provide the error log in kylin's backend, that would help. I think
Hive 3.1 might be too new, because Kylin 3.0 is compiled with Hive 1.1; Maybe
you can try to build with your Hive version:
https://github.com/apache/kylin/blob/kylin-3.0.1/pom.xml
> Cannot sync Hive partitioned table - Ooops.. Cannot get Hive TableMeta
> ----------------------------------------------------------------------
>
> Key: KYLIN-4883
> URL: https://issues.apache.org/jira/browse/KYLIN-4883
> Project: Kylin
> Issue Type: Bug
> Environment: Ubuntu 20.4
> hadoop 3.2.1
> hbase 2.2.6
> hive 3.1.2
> kylin 3.1.0
> Reporter: Iosiv Popescu
> Priority: Major
> Fix For: v3.1.0
>
>
> Dear Kylini team,
> I have started working with Kylin in October 2020. All good until I read to
> partition the fact table in Hive to match partition column in Kylin.
> Tried to do that and I fail when I try to sync the new defined table
> containing any kind of partioned column type. I tried int, date, etc.
> I think I've read anything available on internet about this and I probably
> tried all configurations possible.
> I have a working environment in general except this sync in Kylin of a
> partioned table. To avoid this I created a view over the partitioned fact
> table but I understood this is not very helpful as Hive would be forced to
> full scan the underlying table.
> I have a working HCatalog I think and it should be fine in my opinion.
> I do not know what else to do. And the table started to get some volume as
> 400 milion rows and the time to load the data from Hive started to get large
> enough.
>
> Please help.
>
> Thank you,
> Iosiv
>
>
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
