You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Yuming Wang (Jira)" <ji...@apache.org> on 2022/08/31 01:41:00 UTC
[jira] [Resolved] (SPARK-39971) ANALYZE TABLE makes some queries run forever
[ https://issues.apache.org/jira/browse/SPARK-39971?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Yuming Wang resolved SPARK-39971.
---------------------------------
Resolution: Not A Problem
> ANALYZE TABLE makes some queries run forever
> --------------------------------------------
>
> Key: SPARK-39971
> URL: https://issues.apache.org/jira/browse/SPARK-39971
> Project: Spark
> Issue Type: Bug
> Components: Optimizer, SQL
> Affects Versions: 3.2.2
> Reporter: Felipe
> Priority: Major
> Attachments: 1.1.BeforeAnalyzeTable-joinreorder-disabled.txt, 1.2.BeforeAnalyzeTable-joinreorder-enabled.txt, 2.1.AfterAnalyzeTable WITHOUT ForAllColumns-joinreorder-disabled.txt, 2.2.AfterAnalyzeTable WITHOUT ForAllColumns-joinreorder-enabled.txt, 3.1.AfterAnalyzeTableForAllColumns-joinreorder-disabled.txt, 3.2.AfterAnalyzeTableForAllColumns-joinreorder-enabled.txt, explainMode-cost.zip
>
>
> I'm using TPCDS to run benchmarks, and after running ANALYZE TABLE (without the FOR ALL COLUMNS) some queries became really slow. For example query24 - [https://raw.githubusercontent.com/Agirish/tpcds/master/query24.sql] takes between 10~15min before running the ANALYZE TABLE.
> After running ANALYZE TABLE I waited 24h before cancelling the execution.
> If I disable spark.sql.cbo.joinReorder.enabled or
> spark.sql.cbo.enabled it becomes fast again.
> It seems something in join reordering is not working well when we have table stats, but not column stats.
> Rows Count:
> store_sales - 2879966589
> store_returns - 288009578
> store - 1002
> item - 300000
> customer - 12000000
> customer_address - 6000000
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org