[jira] [Comment Edited] (HIVE-14257) CBO: Push Join through Groupby to trigger shuffle reductions

["Gopal V (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-14257?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15380365#comment-15380365 ] 

Gopal V edited comment on HIVE-14257 at 7/16/16 12:32 AM:
----------------------------------------------------------

The {{SyntheticJoinPredicate}} is actually in place to indicate these relationships. However those are not leveraged for anything other than the DPP runtime.

RedundantDynamicPruningConditionsRemoval removes them for all non-partitioned columns.


was (Author: gopalv):
The {{SyntheticJoinPredicate}} is actually in place to indicate these relationships. However those are not leveraged for anything other than the DPP runtime.

> CBO: Push Join through Groupby to trigger shuffle reductions
> ------------------------------------------------------------
>
>                 Key: HIVE-14257
>                 URL: https://issues.apache.org/jira/browse/HIVE-14257
>             Project: Hive
>          Issue Type: Improvement
>          Components: CBO
>            Reporter: Gopal V
>
> Similar to the optimizations in hive, already which push aggregates through a join (hive.transpose.aggr.join=true).
> {code}
> select count(v) from (select d_year, count(ss_item_sk) as v from store_sales, date_dim where ss_sold_date_sk=d_Date_sk group by d_year) w, date_dim d where d.d_year = w.d_year and d_date_sk = 1;
> {code}
> currently produces an entire aggregate of all years before discarding all of that (because obviously, there's no data for d_date_sk=1;
> This particular example is a simplified version of TPC-DS Query59's join condition, which can have a reduction in scans by applying the d_month_seq between 1185 and 1185 + 11 into the wss alias.



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