[
https://issues.apache.org/jira/browse/HBASE-1385?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12725496#action_12725496
]
Lars George commented on HBASE-1385:
------------------------------------
bq. Should GroupingTableMap become GroupingTableMapper to match?
Yes, an oversight, have renamed it now.
bq. The configuration keys should be mapreduce rather than mapred as in
"hbase.mapred.inputtable" should be "hbase.mapreduce.inputtable"
Yes, although this is only used internally, I will make it match. Makes sense.
bq. Should this be a convenience method on Scan?
I strongly agree. I added three methods to Scan:
* addColumn(byte[])
This one assumes the old style format notation. I have also replaced the
separate parsing in addColumns(byte[][]) to use this new method instead.
*Important!* I have removed the silent try/catch in addColumns(byte[][]) as a)
silent catches are evil and in this case where malformed input can be expected
at a reasonable rate not wanted and b) it bit me changing the TestTableIndex to
the new API.
* addColumns(String)
This one parses a space separate list.
* getInputColumns()
This the reverse to the above. It is needed to create a "human readable"
version of the families and qualifiers.
Does that make sense?
> Revamp TableInputFormat, needs updating to match hadoop 0.20.x AND remove bit
> where we can make < maps than regions
> -------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-1385
> URL: https://issues.apache.org/jira/browse/HBASE-1385
> Project: Hadoop HBase
> Issue Type: Bug
> Reporter: stack
> Fix For: 0.21.0
>
> Attachments: 1385-v4.patch, 1385-v5.patch, 1385-v6.patch, 1385.patch,
> mr.patch
>
>
> Update TIF to match new MR.
> Remove the bit of logic where we will use number of configured maps as splits
> count rather than regions.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
