[
https://issues.apache.org/jira/browse/HIVE-21883?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16867756#comment-16867756
]
slim bouguerra commented on HIVE-21883:
---------------------------------------
Hi am part of the authors of this storage handler, TBH we never tested against
Spark, the only target we have had is LLAP and TEZ containers.
Am not very familiar with the mechanics of Spark engine, but looking at your
stack traces seems like you have a class path issue.
> Cannot use CTAS for managed table backed by Druid for Hive on Spark
> -------------------------------------------------------------------
>
> Key: HIVE-21883
> URL: https://issues.apache.org/jira/browse/HIVE-21883
> Project: Hive
> Issue Type: Bug
> Components: Druid integration
> Affects Versions: 2.3.5
> Environment: Hive version: 2.3.5
> Druid version: 0.14.2
> Execution engine Spark: 2.4.1
> Reporter: Peeranut Ngaorungsri
> Priority: Major
> Attachments: error_from_hive_server.txt, error_from_hive_server_2.txt
>
>
> Run the following CTAS:
> "CREATE TABLE district
> STORED BY 'org.apache.hadoop.hive.druid.DruidStorageHandler'
> AS SELECT dt_cre as `__time`, dis_code, dis_desc FROM finwiz_cdc.tu_district"
>
> PS1. Where dt_cre is timestamp, dis_code and dis_desc are string.
> PS2. Please find an attached error from hive
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
