hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Schubert Zhang (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-1841) If multiple of same key in an hfile and they span blocks, may miss the earlier keys on a lookup
Date Fri, 13 Nov 2009 10:44:40 GMT

    [ https://issues.apache.org/jira/browse/HBASE-1841?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12777440#action_12777440
] 

Schubert Zhang commented on HBASE-1841:
---------------------------------------

@ stack

Thanks stack. A new issue HBASE-1978 for the future patch.

@ ryan and stack

I think the "seekBefore" should not be reason to change lastKey indexing to firstKey indexing.

I my opinion, our major use cases are "seek to a point, and scan next....", these use cases
need lastKey indexing, i.e. (startKey, endKey] range/block scheme.
But the other scheme, i.e.  [startKey, endKey) range/block is more suited to "seek to a point,
and scan prev....".

I think the Bigtable paper's describe is a good choice.

> If multiple of same key in an hfile and they span blocks, may miss the earlier keys on
a lookup
> -----------------------------------------------------------------------------------------------
>
>                 Key: HBASE-1841
>                 URL: https://issues.apache.org/jira/browse/HBASE-1841
>             Project: Hadoop HBase
>          Issue Type: Bug
>            Reporter: stack
>            Assignee: Schubert Zhang
>             Fix For: 0.21.0
>
>         Attachments: HBASE-1841-step1-v2.patch, HBASE-1841-step2-v2.patch
>
>
> See HBASE-818 for description by Schubert Zhang -- discovered by him doing a code review
of hfile.

-- 
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