cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gerrit Jansen van Vuuren (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (CASSANDRA-5640) Allow JOIN on partition key for tables in the same keyspace
Date Thu, 18 Dec 2014 21:21:14 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-5640?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14252271#comment-14252271
] 

Gerrit Jansen van Vuuren edited comment on CASSANDRA-5640 at 12/18/14 9:20 PM:
-------------------------------------------------------------------------------

A basic example of how the patch work is:

cqlsh:test> select * from src ;
id | a
---+---
5 | a5
1 | a1
2 | a2
4 | a4
3 | a3
(5 rows)
cqlsh:test> select * from dst;
a | b
---+---
a4 | b4
a3 | b3
a2 | b2
a1 | b1
cqlsh:test> select a, dst.b as q from src join dst on a;
a | q
---+-----
a5 | null
a1 | b1
a2 | b2
a4 | b4
a3 | b3



was (Author: gerritjvv):
A basic example of how the patch works is:

cqlsh:test> select * from src ;
id | a
---+---
5 | a5
1 | a1
2 | a2
4 | a4
3 | a3
(5 rows)
cqlsh:test> select * from dst;
a | b
---+---
a4 | b4
a3 | b3
a2 | b2
a1 | b1
cqlsh:test> select a, dst.b as q from src join dst on a;
a | q
---+-----
a5 | null
a1 | b1
a2 | b2
a4 | b4
a3 | b3


> Allow JOIN on partition key for tables in the same keyspace
> -----------------------------------------------------------
>
>                 Key: CASSANDRA-5640
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-5640
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: API
>            Reporter: Jeremiah Jordan
>            Priority: Minor
>              Labels: cql
>         Attachments: SIMPLE-CRUDE-OUTER-JOIN.patch
>
>
> People whine about there not being any JOIN in CQL a lot.  I was thinking we could add
JOIN with the restriction that you can only join on partition key, and only if the tables
are in the same keyspace.  That way it could be done local to each node, and would not need
to be a true distributed join.
> Feel free to shoot me down ;)



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

Mime
View raw message