[jira] Commented: (HIVE-591) create new type of join ( 1 row for a given key from multiple tables)

["Adam Kramer (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-591?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12740438#action_12740438 ] 

Adam Kramer commented on HIVE-591:
----------------------------------

The other thing missing from this syntax is joining on multiple keys.

So, ...FROM UNIQUEJOIN A(a.c1,a.c2), B(b.c1,b.c2), C(c.c1,c.c2) WHERE... would be very very useful in this implementation, matching all of the cited columns conjunctively.

Also, is there any reason this needs to be trinary? Would it be possible to make this n-ary? I could see wanting to join 4 or 5 tables.

> create new type of join ( 1 row for a given key from multiple tables)
> ---------------------------------------------------------------------
>
>                 Key: HIVE-591
>                 URL: https://issues.apache.org/jira/browse/HIVE-591
>             Project: Hadoop Hive
>          Issue Type: New Feature
>          Components: Query Processor
>            Reporter: Namit Jain
>
> It will be useful to support a new type of join:
> say:
> select .. from JOIN TABLES (A,B,C) WITH KEYS (A.key, B.key, C.key) where ....
> The semantics are that for a given key only 1 row is created - nulls are present for the the tables which do not contain a row for that key.
> There is no limit on the number of tables, the number of keys should be the same as the number of tables.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.