hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jonathan Gray (JIRA)" <j...@apache.org>
Subject [jira] Commented: (HBASE-783) For single row, single family retrieval, getRow() works half as fast as getScanner().next()
Date Tue, 29 Jul 2008 17:21:32 GMT

    [ https://issues.apache.org/jira/browse/HBASE-783?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12617865#action_12617865
] 

Jonathan Gray commented on HBASE-783:
-------------------------------------

Patch appears to improve performance to the same level as using getScanner.next

This is only in the case of getting all columns for a family.

Not directly related to this issue, but when getting a subset of columns in a single family,
getScanner is still twice as fast as getRow.

> For single row, single family retrieval, getRow() works half as fast as getScanner().next()
> -------------------------------------------------------------------------------------------
>
>                 Key: HBASE-783
>                 URL: https://issues.apache.org/jira/browse/HBASE-783
>             Project: Hadoop HBase
>          Issue Type: Improvement
>    Affects Versions: 0.2.0
>            Reporter: Jonathan Gray
>            Assignee: Jean-Daniel Cryans
>            Priority: Minor
>             Fix For: 0.2.1, 0.3.0
>
>         Attachments: hbase-783-v1.patch
>
>
> We have a very typical use-case of wanting to retrieve all columns under a single row,
single family.
> Benchmarking regularly reproduces the same result.  Using HTable.getScanner().next()
to get the RowResult works twice as fast as HTable.getRow().
> This is related to HBASE-631 by jdcryans which added this functionality to getRow().

-- 
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