[
https://issues.apache.org/jira/browse/CASSANDRA-6318?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13847413#comment-13847413
]
Sylvain Lebresne commented on CASSANDRA-6318:
---------------------------------------------
bq. How important is it deemed by the team ?
I can't answer for "the team", but can tell you that personally I don't
consider CASSANDRA-4386 as a big priority. It's almost only a performance
issue, you can always replace a IN by multiple queries, and while
CASSANDRA-4386 will probably yield "some" benefits over doing multiple queries,
it's unlikely to be a huge difference. Besides, it is no secret that 2ndary
indexes in C* are not its most peformance feature and that when performance is
your main concern, denormalization is your friend. So that relying on IN
queries on 2ndary index a lot is likely not the best of ideas.
bq. What is the mental image of C* usage in the team ?
Hopefully you understand this question is pretty generic, has no simple answer
and has no place in a comment of this JIRA ticket. If you have data modeling
questions, the user mailing list is probably the right place.
bq. How hard is it to implement ?
My best estimation would be "likely not terribly hard, though not 2 lines of
code either".
bq. If i comment out the exception in
cassandra.cql3.statements.SelectStatement.RawStatement#prepare, what and where
will break ?
Because I can't really answer that precisely without some code dive, because
it's relatively simple for you to try it out and see for yourself and because
you'll likely learn something in the process, I'm gonna let you do it if you
care enough about an answer.
> IN predicates on non-primary-key columns (%s) is not yet supported
> ------------------------------------------------------------------
>
> Key: CASSANDRA-6318
> URL: https://issues.apache.org/jira/browse/CASSANDRA-6318
> Project: Cassandra
> Issue Type: Bug
> Reporter: Sergey Nagaytsev
> Labels: cql3
> Attachments: CASSANDRA_6318_test.cql
>
>
> Query:
> SELECT * FROM post WHERE blog IN (1,2) AND author=3 ALLOW FILTERING --
> contrived
> Error: IN predicates on non-primary-key columns (blog) is not yet supported
> Please either implement, set milestone or say will never be implemented !
> P.S. Did search, seemingly found no issue/plan related to it. Maybe
> CASSANDRA-6048 ?
> P.S.2 What is recommended workaround for this ? Manual index tables, what are
> design guidelines for them ?
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
