[
https://issues.apache.org/jira/browse/HIVE-10833?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eugene Koifman updated HIVE-10833:
----------------------------------
Description:
While working on HIVE-10828 I noticed that internal state of RowResolver looks
odd when CBO is enabled.
Consider the script below.
{noformat}
set hive.enforce.bucketing=true;
set hive.exec.dynamic.partition.mode=nonstrict;
set hive.cbo.enable=false;
drop table if exists acid_partitioned;
create table acid_partitioned (a int, c string)
partitioned by (p int)
clustered by (a) into 1 buckets;
insert into acid_partitioned partition (p) (a,p) values(1,1);
{noformat}
(this test is part of
ql/src/test/results/clientpositive/insert_into_with_schema2.q)
With CBO on,
if you put a break point in {noformat}SemanticAnalyzer.genSelectPlan(String
dest, ASTNode selExprList, QB qb, Operator<?> input,
Operator<?> inputForSelectStar, boolean outerLV){noformat} at line
_selectStar = selectStar && exprList.getChildCount() == posn + 1;_
(currently 3865) and examine _out_rwsch.rslvMap_ variable looks like
{noformat}{null={values__tmp__table__1.tmp_values_col1=_col0: string,
values__tmp__table__1.tmp_values_col2=_col1: string}}{noformat}
with CBO disabled, the same _out_rwsch.rslvMap_ looks like
{noformat}{values__tmp__table__1={tmp_values_col1=_col0: string,
tmp_values_col2=_col1: string}}{noformat}
The _out_rwsch.invRslvMap_ also differs in the same way.
It seems that the version you get with CBO off is the correct one since
_insert into acid_partitioned partition (p) (a,p) values(1,1)_ is rewritten to
_insert into acid_partitioned partition (p) (a,p) select * from
values__tmp__table__1_
CC [~ashutoshc]
was:
While working on HIVE-10828 I noticed that internal state of RowResolver looks
odd when CBO is enabled.
Consider the script below.
{noformat}
set hive.enforce.bucketing=true;
set hive.exec.dynamic.partition.mode=nonstrict;
set hive.cbo.enable=false;
drop table if exists acid_partitioned;
create table acid_partitioned (a int, c string)
partitioned by (p int)
clustered by (a) into 1 buckets;
insert into acid_partitioned partition (p) (a,p) values(1,1);
{noformat}
With CBO on,
if you put a break point in {noformat}SemanticAnalyzer.genSelectPlan(String
dest, ASTNode selExprList, QB qb, Operator<?> input,
Operator<?> inputForSelectStar, boolean outerLV){noformat} at line
_selectStar = selectStar && exprList.getChildCount() == posn + 1;_
(currently 3865) and examine _out_rwsch.rslvMap_ variable looks like
{noformat}{null={values__tmp__table__1.tmp_values_col1=_col0: string,
values__tmp__table__1.tmp_values_col2=_col1: string}}{noformat}
with CBO disabled, the same _out_rwsch.rslvMap_ looks like
{noformat}{values__tmp__table__1={tmp_values_col1=_col0: string,
tmp_values_col2=_col1: string}}{noformat}
The _out_rwsch.invRslvMap_ also differs in the same way.
It seems that the version you get with CBO off is the correct one since
_insert into acid_partitioned partition (p) (a,p) values(1,1)_ is rewritten to
_insert into acid_partitioned partition (p) (a,p) select * from
values__tmp__table__1_
CC [~ashutoshc]
> RowResolver looks mangled with CBO
> -----------------------------------
>
> Key: HIVE-10833
> URL: https://issues.apache.org/jira/browse/HIVE-10833
> Project: Hive
> Issue Type: Bug
> Components: CBO, Query Planning
> Affects Versions: 1.3.0
> Reporter: Eugene Koifman
> Assignee: Laljo John Pullokkaran
>
> While working on HIVE-10828 I noticed that internal state of RowResolver
> looks odd when CBO is enabled.
> Consider the script below.
> {noformat}
> set hive.enforce.bucketing=true;
> set hive.exec.dynamic.partition.mode=nonstrict;
> set hive.cbo.enable=false;
> drop table if exists acid_partitioned;
> create table acid_partitioned (a int, c string)
> partitioned by (p int)
> clustered by (a) into 1 buckets;
>
> insert into acid_partitioned partition (p) (a,p) values(1,1);
> {noformat}
> (this test is part of
> ql/src/test/results/clientpositive/insert_into_with_schema2.q)
> With CBO on,
> if you put a break point in {noformat}SemanticAnalyzer.genSelectPlan(String
> dest, ASTNode selExprList, QB qb, Operator<?> input,
> Operator<?> inputForSelectStar, boolean outerLV){noformat} at line
> _selectStar = selectStar && exprList.getChildCount() == posn + 1;_
> (currently 3865) and examine _out_rwsch.rslvMap_ variable looks like
> {noformat}{null={values__tmp__table__1.tmp_values_col1=_col0: string,
> values__tmp__table__1.tmp_values_col2=_col1: string}}{noformat}
> with CBO disabled, the same _out_rwsch.rslvMap_ looks like
> {noformat}{values__tmp__table__1={tmp_values_col1=_col0: string,
> tmp_values_col2=_col1: string}}{noformat}
> The _out_rwsch.invRslvMap_ also differs in the same way.
> It seems that the version you get with CBO off is the correct one since
> _insert into acid_partitioned partition (p) (a,p) values(1,1)_ is rewritten to
> _insert into acid_partitioned partition (p) (a,p) select * from
> values__tmp__table__1_
> CC [~ashutoshc]
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
