[
https://issues.apache.org/jira/browse/CASSANDRA-261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12725331#action_12725331
]
Jun Rao commented on CASSANDRA-261:
-----------------------------------
Yes, the code has to scan and skip the portion of data between 0 and offset.
So, the larger the offset, the worse the performance. It's hard to avoid this
scanning overhead since offset is based on data merged from all sstables and
it's difficult to push offset into individual sstables.
However, this is still better than the old get_slice that has to materialize
the whole CF. A typical use case of this function is to support page scrolling.
Typically, a user goes to earlier pages much more frequently. So this api could
still be useful. We should document this impact though.
> get_slice needs offset + limit
> ------------------------------
>
> Key: CASSANDRA-261
> URL: https://issues.apache.org/jira/browse/CASSANDRA-261
> Project: Cassandra
> Issue Type: Bug
> Reporter: Chris Goffinet
> Assignee: Jun Rao
> Attachments: issue261.patchv1
>
>
> Right now get_slice does not allow you to provide an offset. This would help
> for pagination.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
