[
https://issues.apache.org/jira/browse/CASSANDRA-5250?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13607685#comment-13607685
]
Jonathan Ellis commented on CASSANDRA-5250:
-------------------------------------------
I would say, go ahead and rip out EmptyCompactionScanner.
I was also hoping that we could get a better estimate for sstables where the
range we're scanning is only part of the sstable. One way would be to do
lookups against our in-memory IndexSummary for the start and end of the range,
and assume that 1/10 of the rows in the summary correspond to 1/10 of the size
on disk. Obviously not 100% accurate but a lot better than throwing the entire
sstable size in each time.
> Improve LeveledScanner work estimation
> --------------------------------------
>
> Key: CASSANDRA-5250
> URL: https://issues.apache.org/jira/browse/CASSANDRA-5250
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Jonathan Ellis
> Assignee: Marcus Eriksson
> Labels: compaction
> Attachments: 0001-CASSANDRA-5250.patch
>
>
> See
> https://issues.apache.org/jira/browse/CASSANDRA-5222?focusedCommentId=13577420&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-13577420
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
