[jira] [Commented] (CASSANDRA-9835) Restore collectTimeOrderedData behaviour post-8099

["Sylvain Lebresne (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-9835?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14631123#comment-14631123 ] 

Sylvain Lebresne commented on CASSANDRA-9835:
---------------------------------------------

What makes you think that? We do still do this, it's now in {{SinglePartitionNamesCommand}} (in the {{queryMemtableAndDiskInternal}} method to be precise).

In fact, if anything, CASSANDRA-8099 makes this a lot more useful since pre-CASSANDRA-8099, {{collectTimeOrderedData}} is _never_ used for non-compact tables (we just never create names queries for non-compact tables for reasons explained at length in CASSANDRA-7085 and related issues).

> Restore collectTimeOrderedData behaviour post-8099
> --------------------------------------------------
>
>                 Key: CASSANDRA-9835
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-9835
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Benedict
>             Fix For: 3.0.x
>
>
> AFAICT, we no longer prune the sstables we iterate once we know we've satisfied a query. This is not only still possible, but possible in more scenarios (since we can do it for any single CQL-row lookup).
> Affected workloads may have noticeably degraded behaviour, and this will impact CASSANDRA-6477.



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