[jira] [Comment Edited] (IMPALA-7919) Add predicates line in plan output for partition key predicates

["Paul Rogers (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IMPALA-7919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16733717#comment-16733717 ] 

Paul Rogers edited comment on IMPALA-7919 at 1/4/19 1:52 AM:
-------------------------------------------------------------

Turns out that this feature is important to diagnosing issues with plans. Adding it. Example (along with the addition of cardinality):

{noformat}
|  01:SCAN HDFS [functional.alltypessmall b]
|     partitions=2/4 files=2 size=3.17KB row-size=89B cardinality=5
|     predicates: b.string_col = '15'
|     partition predicates: b.month > 2
{noformat}

Added the new line after the existing line to minimize change. (One could make a good argument that the two predicate lines should appear at the top, followed ty the partitions/cardinality lines that show the result of those predicates, but this would change the order of the lines relative to current plans.)


was (Author: paul.rogers):
Turns out that this feature is important to diagnosing issues with plans. Adding it.

> Add predicates line in plan output for partition key predicates
> ---------------------------------------------------------------
>
>                 Key: IMPALA-7919
>                 URL: https://issues.apache.org/jira/browse/IMPALA-7919
>             Project: IMPALA
>          Issue Type: Improvement
>          Components: Frontend
>            Reporter: Greg Rahn
>            Assignee: Paul Rogers
>            Priority: Major
>              Labels: planner, ramp-up
>
> When there is a predicate on a partitioned table's partition key column the SCAN node does not print the "predicates" line as it would if the table was not partitioned. IMO predicates should always be included in the nodes where they are applied irregardless of partitioning or not to make it clear.
> Query:
> {noformat}
> select * from t1 where part_key=42;
> {noformat}
> From a non-partitioned table:
> {noformat}
> 00:SCAN HDFS [default.t1]
>    partitions=1/1 files=2 size=10B
>    predicates: default.t1.part_key = 42
> {noformat}
> From a partitioned table:
> {noformat}
> 00:SCAN HDFS [default.t1]
>    partitions=1/2 files=1 size=2B
> {noformat}



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-all-unsubscribe@impala.apache.org
For additional commands, e-mail: issues-all-help@impala.apache.org