[ 
https://issues.apache.org/jira/browse/IGNITE-25093?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Pavel Pereslegin updated IGNITE-25093:
--------------------------------------
    Ignite Flags:   (was: Docs Required,Release Notes Required)

> Sql. FragmentMappingTest testCorrelated uses random distribution instead of 
> affinity
> ------------------------------------------------------------------------------------
>
>                 Key: IGNITE-25093
>                 URL: https://issues.apache.org/jira/browse/IGNITE-25093
>             Project: Ignite
>          Issue Type: Bug
>          Components: sql
>            Reporter: Maksim Zhuravkov
>            Assignee: Maksim Zhuravkov
>            Priority: Major
>              Labels: ignite-3
>             Fix For: 3.1
>
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Upgrade to calcite 1.39 caused a changed in a plan - distribution of a 
> correlated table changed from affinity to random)
> {noformat}
> N0
> SELECT t.c1 FROM t_n1 t WHERE t.c1 < 5 AND
> EXISTS (SELECT x FROM table(system_range(t.c1, t.c2)) WHERE mod(x, 2) = 0)
> {noformat}
> Before:
> {noformat}
> ---
> Fragment#0 root
>   executionNodes: [N0]
>   remoteFragments: [1]
>   exchangeSourceNodes: {1=[N1]}
>   tree:
>     Project
>       CorrelatedNestedLoopJoin
>         Receiver(sourceFragment=1, exchange=1, distribution=single)
>         ColocatedHashAggregate
>           Project
>             Filter
>               TableFunctionScan(source=2, distribution=single)
> Fragment#1
>   targetNodes: [N0]
>   executionNodes: [N1]
>   tables: [T_N1]
>   partitions: {N1=[0:1]}
>   tree:
>     Sender(targetFragment=0, exchange=1, distribution=single)
> >>>      TableScan(name=PUBLIC.T_N1, source=3, partitions=1, 
> >>> distribution=affinity[table: T_N1, columns: [ID]])
> {noformat}
> After:
> {noformat}
> ---
> Fragment#0 root
>   executionNodes: [N0]
>   remoteFragments: [1]
>   exchangeSourceNodes: {1=[N1]}
>   tree:
>     Project
>       CorrelatedNestedLoopJoin
>         Receiver(sourceFragment=1, exchange=1, distribution=single)
>         ColocatedHashAggregate
>           Project
>             Filter
>               TableFunctionScan(source=2, distribution=single)
> Fragment#1
>   targetNodes: [N0]
>   executionNodes: [N1]
>   tables: [T_N1]
>   partitions: {N1=[0:1]}
>   tree:
>     Sender(targetFragment=0, exchange=1, distribution=single)
> >>>      TableScan(name=PUBLIC.T_N1, source=3, partitions=1, 
> >>> distribution=random)
> {noformat}



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

Reply via email to