drill-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "ASF GitHub Bot (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (DRILL-4998) Perform parquet filter pushdown in logical planning.
Date Fri, 04 Nov 2016 04:31:58 GMT

    [ https://issues.apache.org/jira/browse/DRILL-4998?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15635198#comment-15635198
] 

ASF GitHub Bot commented on DRILL-4998:
---------------------------------------

Github user jinfengni commented on a diff in the pull request:

    https://github.com/apache/drill/pull/637#discussion_r86487867
  
    --- Diff: exec/java-exec/src/main/java/org/apache/drill/exec/planner/PlannerPhase.java
---
    @@ -346,6 +355,20 @@ static RuleSet getPruneScanRules(OptimizerRulesContext optimizerRulesContext)
{
       }
     
       /**
    +   *   Get an immutable list of partition pruning rules that will be used in logical
planning.
    +   */
    +  static RuleSet getPhysicalPruneScanRules(OptimizerRulesContext optimizerRulesContext)
{
    +    final ImmutableSet<RelOptRule> pruneRules = ImmutableSet.<RelOptRule>builder()
    +        .add(
    +            ParquetPushDownFilter.getFilterOnProject(optimizerRulesContext),
    --- End diff --
    
    Opened DRILL-4998, and put comments to explain why we put this rule in physical planning
phase. 


> Perform parquet filter pushdown in logical planning.
> ----------------------------------------------------
>
>                 Key: DRILL-4998
>                 URL: https://issues.apache.org/jira/browse/DRILL-4998
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Query Planning & Optimization
>            Reporter: Jinfeng Ni
>
> DRILL-1950 put parquet filter pushdown rule in physical planning phase, after all join
order related tasks have been done.  Main reason for doing this is we want to reduce the performance
regression possibility caused by a different join order, as a result of reduced row count
in scan operator. 
> Ideally this should be done in logical planning, before join order planning is done.
Before we can make such change, we have to figure out how to adjust the selectivity estimation
of filter operator, after filter is pushed down to scan.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message