cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gary Dusbabek (JIRA)" <j...@apache.org>
Subject [jira] Commented: (CASSANDRA-350) optimize away unnecessary range scans
Date Tue, 27 Oct 2009 21:02:59 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12770638#action_12770638
] 

Gary Dusbabek commented on CASSANDRA-350:
-----------------------------------------

After working up some test cases and examining the code more closely, it appears that we can't
optimize the case where reversed==true, since it starts reading at the last block anyway,
and when it gets to the first block, it must start at the beginning to find finishColumn.

Am I seeing things correctly?

> optimize away unnecessary range scans
> -------------------------------------
>
>                 Key: CASSANDRA-350
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-350
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Jonathan Ellis
>            Assignee: Gary Dusbabek
>             Fix For: 0.5
>
>         Attachments: cassandra-350-v2-tests.patch, cassandra-350-v2.patch, cassandra-350.patch
>
>
> Even when a the columns exist for a given key, they usually won't exist in all the sstable
segments.  So avoiding the scan when the column index shows we can, will be a win.
> This is what the todo in SSTableSliceIterator is about:
>         // TODO push finishColumn down here too, so we can tell when we're done and optimize
away the slice when the index + start/stop shows there's nothing to scan for

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


Mime
View raw message