hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Zoltan Haindrich (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (HIVE-17332) NullPointer exception when processing query
Date Mon, 21 Aug 2017 11:53:00 GMT

    [ https://issues.apache.org/jira/browse/HIVE-17332?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16135051#comment-16135051
] 

Zoltan Haindrich edited comment on HIVE-17332 at 8/21/17 11:52 AM:
-------------------------------------------------------------------

Hello [~luky],

I was not able to reproduce the issue on the current master...it might be possible that its
not anymore present...
I've just guessed the column types..they might be more relevant....and my guess is that the
problem might be sensitive to some null value in some column - does the snapshot column contains
null values?
Could you check the following case:

{code}
create table EXM_BASE_DATA (key string) partitioned by (snapshot int) stored as orc;
insert into EXM_BASE_DATA partition(snapshot=1) values ('asd') ;
insert into EXM_BASE_DATA partition(snapshot=2) values (null) ;
set hive.auto.convert.join = false;
select count(*) from
        (select * from EXM_BASE_DATA join (select max(snapshot) max_snapshot from EXM_BASE_DATA)
s0 on (snapshot = max_snapshot)) t;
{code}


or this more sophisticated one:

{code}
set hive.exec.dynamic.partition.mode=nonstrict;

create table t0 (key string,snapshot int);
insert into t0 values (1,'asd');
insert into t0 values (null,'asd2');
insert into t0 values (2,null);
insert into t0 values (null,null);


create table EXM_BASE_DATA (key string) partitioned by (snapshot int) stored as orc;

from t0
insert into EXM_BASE_DATA partition(snapshot) select key,snapshot;

insert into EXM_BASE_DATA partition(snapshot=1) values ('asdx') ;
insert into EXM_BASE_DATA partition(snapshot=1) values ('asd') ;
insert into EXM_BASE_DATA partition(snapshot=2) values (null) ;

set hive.auto.convert.join = false;

select count(*) from
        (select * from EXM_BASE_DATA join (select max(snapshot) max_snapshot from EXM_BASE_DATA)
s0 on (snapshot = max_snapshot)) t;

{code}


was (Author: kgyrtkirk):
Hello [~luky],

I was not able to reproduce the issue on the current master...it might be possible that its
not anymore present...
I've just guessed the column types..they might be more relevant....and my guess is that the
problem might be sensitive to some null value in some column - does the snapshot column contains
null values?
Could you check the following case:

{code}
create table EXM_BASE_DATA (key string) partitioned by (snapshot int) stored as orc;
insert into EXM_BASE_DATA partition(snapshot=1) values ('asd') ;
insert into EXM_BASE_DATA partition(snapshot=2) values (null) ;
set hive.auto.convert.join = false;
select count(*) from
        (select * from EXM_BASE_DATA join (select max(snapshot) max_snapshot from EXM_BASE_DATA)
s0 on (snapshot = max_snapshot)) t;
{code}


> NullPointer exception when processing query
> -------------------------------------------
>
>                 Key: HIVE-17332
>                 URL: https://issues.apache.org/jira/browse/HIVE-17332
>             Project: Hive
>          Issue Type: Bug
>    Affects Versions: 1.2.1
>            Reporter: Lukas Waldmann
>
> Hive query:
> {code}
> select count(*) from (select * from EXM_BASE_DATA, (select max(snapshot) max_snapshot
from EXM_BASE_DATA) s0 where snapshot == max_snapshot) t;
> {code}
> finish with NullPointer exception
> while 
> {code}
> select * from EXM_BASE_DATA, (select max(snapshot) max_snapshot from EXM_BASE_DATA) s0
where snapshot == max_snapshot
> {code}
> is executed without error



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message