[jira] [Commented] (HIVE-11845) Bucket map join requires exact number of files as number of buckets

["Prasanth Jayachandran (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-11845?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14790781#comment-14790781 ] 

Prasanth Jayachandran commented on HIVE-11845:
----------------------------------------------

[~vikram.dixit] Please close the jira if you think otherwise.

> Bucket map join requires exact number of files as number of buckets
> -------------------------------------------------------------------
>
>                 Key: HIVE-11845
>                 URL: https://issues.apache.org/jira/browse/HIVE-11845
>             Project: Hive
>          Issue Type: Bug
>    Affects Versions: 1.3.0, 2.0.0
>            Reporter: Prasanth Jayachandran
>            Assignee: Vikram Dixit K
>
> I recently tried to write a test case for bucket map join in tez on ACID table but was unable to do so because bucket map join expects same number of files in table/partition as that of number of buckets. 
> Steps to reproduce:
> create table src_acid(key string, value string) clustered by (key) into 4 buckets stored as orc tblproperties("transactional"="true");
> insert into table src_acid values("key1","value1");
> insert into table src_acid values("key2","value2");
> After inserting 2 rows into the src_acid table, there will be total of 8 files in the table directory which does not match the number of buckets.



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