[
https://issues.apache.org/jira/browse/CASSANDRA-10124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14804575#comment-14804575
]
Sam Tunnicliffe commented on CASSANDRA-10124:
---------------------------------------------
Fair enough, point taken. I've removed the overzealous usage of streams you
mentioned. The worst example (in {{validateTargetsForMultiColumnIndex}}) was
motivated by trying to improve the error message but you're right, it was
certainly a bit overboard. I'm not sure I necessarily agree about the other
case, but I'm more than happy to make the change.
As for the default index name, how about we preserve existing behaviour for
single target indexes whilst for anything else have the generated names be
simply of the form {{table_idx}}, with the existing numeric suffix in the case
of duplicates? It's not a perfect solution, but it fits for indexes without
explicit targets and we can revisit it if we need to when we come to add
functional indexes.
Updated, rebased and pushed to the same location.
> Support for multi-column indexes
> --------------------------------
>
> Key: CASSANDRA-10124
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10124
> Project: Cassandra
> Issue Type: Bug
> Reporter: Sam Tunnicliffe
> Assignee: Sam Tunnicliffe
> Labels: client-impacting
> Fix For: 3.0.0 rc1
>
>
> Since CASSANDRA-6717 decoupled a secondary index from a single column, we can
> expand support for indexes with multiple target columns and for row-based
> indexes with truly dynamic targets.
> Much of the plumbing for this has been done in CASSANDRA-7771, CASSANDRA-6717
> & by the API rework in CASSANDRA-9459. What remains is:
> * Decide on syntax for DDL statements
> * Decide on WHERE clause syntax for SELECT (there is some discussion on this
> in the comments of CASSANDRA-9459)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
