[ 
https://issues.apache.org/jira/browse/HBASE-3005?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12910057#action_12910057
 ] 

HBase Review Board commented on HBASE-3005:
-------------------------------------------

Message from: "Pranav Khaitan" <[email protected]>

-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.cloudera.org/r/859/
-----------------------------------------------------------

Review request for hbase and Kannan Muthukkaruppan.


Summary
-------

Skip between timestamps to avoid the unnecessary read for all the blocks which 
come in between.

For example:

If we have timestamps from 1 to 1000 for a particular row, family, qualifier 
and we want to read only timestamps 990 and 10, then we should be able to read 
990 and then jump to 10 instead of doing a sequential read including everything 
in between.


This addresses bug HBASE-3005.
    http://issues.apache.org/jira/browse/HBASE-3005


Diffs
-----

  trunk/src/main/java/org/apache/hadoop/hbase/KeyValue.java 997639 
  trunk/src/main/java/org/apache/hadoop/hbase/filter/TimestampsFilter.java 
997639 

Diff: http://review.cloudera.org/r/859/diff


Testing
-------

All tests using TimestampFilter passing

Reviewed by Kannan


Thanks,

Pranav




> Add reseek optimization for fetching multiple versions via TimestampsFilter
> ---------------------------------------------------------------------------
>
>                 Key: HBASE-3005
>                 URL: https://issues.apache.org/jira/browse/HBASE-3005
>             Project: HBase
>          Issue Type: Improvement
>          Components: filters
>            Reporter: Pranav Khaitan
>            Assignee: Pranav Khaitan
>         Attachments: HBASE-PUBLIC-3005-V1.patch
>
>
> Skip between timestamps to avoid the unnecessary read for all the blocks 
> which come in between.
> For example:
> If we have timestamps from 1 to 1000 for a particular row, family, qualifier 
> and we want to read only timestamps 990 and 10, then we should be able to 
> read 990 and then jump to 10 instead of doing a sequential read including 
> everything in between.

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

Reply via email to