[
https://issues.apache.org/jira/browse/HBASE-5208?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13188406#comment-13188406
]
Nicholas Telford commented on HBASE-5208:
-----------------------------------------
Regarding test length, without my additions (i.e. clean trunk) the tests take a
very long time. Most of the time is spent doing the original tests as they spin
up 11 MapReduce jobs. I imagine running the tests in parallel might improve
things, but I haven't tested that.
My additional test adds another MapReduce job, so it will increase the test
length, but not substantially.
> Allow setting Scan start/stop row individually in TableInputFormat
> ------------------------------------------------------------------
>
> Key: HBASE-5208
> URL: https://issues.apache.org/jira/browse/HBASE-5208
> Project: HBase
> Issue Type: Improvement
> Components: mapreduce
> Reporter: Nicholas Telford
> Priority: Minor
> Attachments: HBASE-5208-001.txt, HBASE-5208-002.txt,
> HBASE-5208-003.txt, HBASE-5208-004.txt
>
>
> Currently, TableInputFormat initializes a serialized Scan from
> "hbase.mapreduce.scan". Alternatively, it will instantiate a new Scan using
> properties defined in "hbase.mapreduce.scan.*". However, of these properties
> the "start row" and "stop row" (arguably the most pertinent) are missing.
> TableInputFormat should permit the specification of a start/stop row as with
> the other fields using a new pair of properties:
> "hbase.mapreduce.scan.row.start" and "hbase.mapreduce.scan.row.end"
> The primary use-case for this is to permit Oozie and other job management
> tools that can't call TableMapReduceUtil.initTableMapperJob() to operate on a
> contiguous subset of rows.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
