[jira] [Commented] (SPARK-35379) Improve InferFiltersFromConstraints rule performance when parsing spark sql

["Apache Spark (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-35379?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17342992#comment-17342992 ] 

Apache Spark commented on SPARK-35379:
--------------------------------------

User 'wankunde' has created a pull request for this issue:
https://github.com/apache/spark/pull/32514

> Improve InferFiltersFromConstraints rule performance when parsing spark sql
> ---------------------------------------------------------------------------
>
>                 Key: SPARK-35379
>                 URL: https://issues.apache.org/jira/browse/SPARK-35379
>             Project: Spark
>          Issue Type: Improvement
>          Components: SQL
>    Affects Versions: 3.1.1
>            Reporter: Wan Kun
>            Priority: Major
>
> *InferFiltersFromConstraints* rule will generates too many constraints when there are many aliases in Project, . For example:
>  
> {code:java}
> test("Expression explosion when analyze test") {
>     RuleExecutor.resetMetrics()
>     Seq((1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14))
>       .toDF("a", "b", "c", "d", "e", "f", "g", "h", "i", "j",
>         "k", "l", "m", "n")
>       .write.saveAsTable("test")
>     val df = spark.table("test")
>     val df2 = df.filter("a+b+c+d+e+f+g+h+i+j+k+l+m+n > 100")
>     val df3 = df2.select('a as 'a1, 'b as 'b1,
>       'c as 'c1, 'd as 'd1, 'e as 'e1, 'f as 'f1,
>       'g as 'g1, 'h as 'h1, 'i as 'i1, 'j as 'j1,
>       'k as 'k1, 'l as 'l1, 'm as 'm1, 'n as 'n1)
>     val df4 = df3.join(df2, df3("a1") === df2("a"))
>     df4.explain(true)
>     logWarning(RuleExecutor.dumpTimeSpent())
>   }
> {code}
> So we need to optimize *InferFiltersFromConstraints* rule to improve performance.



--
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