[
https://issues.apache.org/jira/browse/HIVE-15666?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Niklaus Xiao updated HIVE-15666:
--------------------------------
Description:
repro steps:
{code}
set hive.execution.engine=spark;
create table base(id int) partitioned by (dt string);
alter table base add partition(dt='2017');
create view view1 as select * from base where id < 10;
select * from view1;
{code}
it requires the access not only for view1 but also for base@dt=2017 partition,
which should not be required.
was:
repo steps:
{code}
set hive.execution.engine=spark;
create table base(id int) partitioned by (dt string);
alter table base add partition(dt='2017');
create view view1 as select * from base where id < 10;
select * from view1;
{code}
it requires the access not only for view1 but also for base@dt=2017 partition,
which should not be required.
> Select query with view adds base table partition as direct input in spark
> engine
> --------------------------------------------------------------------------------
>
> Key: HIVE-15666
> URL: https://issues.apache.org/jira/browse/HIVE-15666
> Project: Hive
> Issue Type: Bug
> Components: Authorization
> Affects Versions: 1.3.0
> Reporter: Niklaus Xiao
> Attachments: TestViewEntityInSparkEngine.patch
>
>
> repro steps:
> {code}
> set hive.execution.engine=spark;
> create table base(id int) partitioned by (dt string);
> alter table base add partition(dt='2017');
> create view view1 as select * from base where id < 10;
> select * from view1;
> {code}
> it requires the access not only for view1 but also for base@dt=2017
> partition, which should not be required.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
