[
https://issues.apache.org/jira/browse/SPARK-24705?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16531023#comment-16531023
]
cheng commented on SPARK-24705:
-------------------------------
It is currently closed. If I want to use this feature to improve performance,
it is currently not possible. I think this is a bug and it is necessary to fix
it.
> Spark.sql.adaptive.enabled=true is enabled and self-join query
> --------------------------------------------------------------
>
> Key: SPARK-24705
> URL: https://issues.apache.org/jira/browse/SPARK-24705
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 2.2.1, 2.3.1
> Reporter: cheng
> Priority: Minor
> Attachments: Error stack.txt
>
>
> [~smilegator]
> When loading data using jdbc and enabling spark.sql.adaptive.enabled=true ,
> for example loading a tableA table, unexpected results can occur when you use
> the following query.
> For example:
> device_loc table comes from the jdbc data source
> select tv_a.imei
> from ( select a.imei,a.speed from device_loc a) tv_a
> inner join ( select a.imei,a.speed from device_loc a ) tv_b on tv_a.imei =
> tv_b.imei
> group by tv_a.imei
> When the cache tabel device_loc is executed before this query is executed,
> everything is fine,However, if you do not execute cache table, unexpected
> results will occur, resulting in failure to execute.
> Remarks：Attachment records the stack when the error occurred
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
