[
https://issues.apache.org/jira/browse/HIVE-12017?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14997797#comment-14997797
]
Ashutosh Chauhan edited comment on HIVE-12017 at 11/10/15 1:34 AM:
-------------------------------------------------------------------
I went through golden file plan changes and found following categories of plan
diffs:
* 1) extra select operator : Many plans now have extra select operator in
plans. e.g., auto_sortmerge_join_*.q
* 2) agg expr lost : In some tests, it seems like we dropped the aggregation
altogether count (*) e.g, auto_smb_mapjoin_14.q,auto_sortmerge_join_10.q
* 3) Shuffle join warning : Some tests now are generating shuffle join warning,
e.g, multiMapJoin2.q,orc_llap.q,parquet_join.q,pcr.q,pointlookup2.q
* 4) extra columns : seems like column pruning issue:
auto_join1.q,auto_join10.q,auto_join11.q
* 5) PTF op missing : This one seems like ptf operator got dropped altogether
ptfgroupbyjoin.q.
* 6) Non-skew-join plan : Seems like skew join optimization is broken and we
drop that optimization. e.g., skewjoin_mapjoin*.q
Among these 1) & 4) are not a big concern. However, 2) & 5) could be
correctness issue and 3) & 6) could be substantial perf losses.
was (Author: ashutoshc):
I went through golden file plan changes and found following categories of plan
diffs:
* 1) extra select operator : Many plans now have extra select operator in
plans. e.g., auto_sortmerge_join_*.q
* 2) agg expr lost : In some tests, it seems like we dropped the aggregation
altogether count (*) e.g, auto_smb_mapjoin_14.q,auto_sortmerge_join_10.q
* 3) Shuffle join warning : Some tests now are generating shuffle join warning,
e.g, multiMapJoin2.q,orc_llap.q,parquet_join.q,pcr.q,pointlookup2.q
* 4) extra columns : seems like column pruning issue:
auto_join1.q,auto_join10.q,auto_join11.q
* 5) PTF op missing : This one seems like ptf operator got dropped altogether
ptfgroupbyjoin.q.
* 6) Non-skew-join plan : Seems like skew join optimization is broken and we
drop that optimization. e.g., skewjoin_mapjoin*.q
Among these 1) & 4) are not a big concern. However, 2) & 5) could be
correctness issue and 3) & 7) could be substantial perf losses.
> Do not disable CBO by default when number of joins in a query is equal or
> less than 1
> -------------------------------------------------------------------------------------
>
> Key: HIVE-12017
> URL: https://issues.apache.org/jira/browse/HIVE-12017
> Project: Hive
> Issue Type: Improvement
> Components: CBO
> Affects Versions: 2.0.0
> Reporter: Jesus Camacho Rodriguez
> Assignee: Jesus Camacho Rodriguez
> Attachments: HIVE-12017.01.patch, HIVE-12017.02.patch,
> HIVE-12017.03.patch, HIVE-12017.04.patch, HIVE-12017.05.patch,
> HIVE-12017.06.patch, HIVE-12017.07.patch, HIVE-12017.08.patch
>
>
> Instead, we could disable some parts of CBO that are not relevant if the
> query contains 1 or 0 joins. Implementation should be able to define easily
> other query patterns for which we might disable some parts of CBO (in case we
> want to do it in the future).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
