You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Apache Spark (Jira)" <ji...@apache.org> on 2020/09/25 16:57:00 UTC
[jira] [Commented] (SPARK-32995) CostBasedJoinReorder optimizer
rule should be idempotent
[ https://issues.apache.org/jira/browse/SPARK-32995?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17202300#comment-17202300 ]
Apache Spark commented on SPARK-32995:
--------------------------------------
User 'tanelk' has created a pull request for this issue:
https://github.com/apache/spark/pull/29871
> CostBasedJoinReorder optimizer rule should be idempotent
> --------------------------------------------------------
>
> Key: SPARK-32995
> URL: https://issues.apache.org/jira/browse/SPARK-32995
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Affects Versions: 3.1.0
> Reporter: Tanel Kiis
> Priority: Minor
>
> The CostBasedJoinReorder has currently the following comment:
> {code}
> // Since join costs in AQP can change between multiple runs, there is no reason that we have an
> // idempotence enforcement on this batch. We thus make it FixedPoint(1) instead of Once.
> {code}
> This is incorrect reasoning.
> In the optimizer, we mean that a batch is idempotent, when applying it for the second time in a row (without changing anything) it does not change the plan at all.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org