[ 
https://issues.apache.org/jira/browse/CALCITE-6887?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17935699#comment-17935699
 ] 

Yu Xu commented on CALCITE-6887:
--------------------------------

[~julianhyde] Thanks for the suggestion!  had update the Jira title.

> ReduceExpressionsRule applied to 'IN subquery' should make the values 
> distinct if the subquery is a constant Values
> -------------------------------------------------------------------------------------------------------------------
>
>                 Key: CALCITE-6887
>                 URL: https://issues.apache.org/jira/browse/CALCITE-6887
>             Project: Calcite
>          Issue Type: Bug
>          Components: core
>    Affects Versions: 1.38.0
>            Reporter: Yu Xu
>            Assignee: Yu Xu
>            Priority: Major
>              Labels: pull-request-available
>             Fix For: 1.40.0
>
>
> Currently IN operator in ReduceExpressionsRule would not distinct values, so 
> need optimize it with distinct values in IN.
> for example  *in (1,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,3,1)*  transform to 
> *in (1,2,3)* is better, but currently would be  *in 
> (1,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,3,1)* without distinct.
> Like case follow:
> *test case:*
> @Test void testReduceExpressionsWithIn()
> { final String sql = "select deptno, sal " + "from emp " + "where deptno in 
> (1,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,3,1) "; 
> sql(sql).withRule(CoreRules.PROJECT_REDUCE_EXPRESSIONS) .check(); }
>  
> *before plan:*
> LogicalProject(DEPTNO=[$7], SAL=[$5])
> LogicalFilter(condition=[IN($7,
> { LogicalValues(tuples=[[
> { 1 }, \{ 1 }, \{ 2 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 
> }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 
> 3 }, \{ 1 }]])
> })])
> LogicalTableScan(table=[[CATALOG, SALES, EMP]])
>  
> *after plan:* 
> LogicalProject(DEPTNO=[$7], SAL=[$5])
> LogicalFilter(condition=[IN($7,
> { LogicalValues(tuples=[[
> { 1 }
> , \{ 1 }, \{ 2 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 
> 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 1 }, \{ 3 }, 
> \{ 1 }]])
> })])
> LogicalTableScan(table=[[CATALOG, SALES, EMP]])



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to