You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hbase.apache.org by "Lars George (JIRA)" <ji...@apache.org> on 2009/06/28 13:24:47 UTC

[jira] Issue Comment Edited: (HBASE-1385) Revamp TableInputFormat, needs updating to match hadoop 0.20.x AND remove bit where we can make < maps than regions

    [ https://issues.apache.org/jira/browse/HBASE-1385?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12724944#action_12724944 ] 

Lars George edited comment on HBASE-1385 at 6/28/09 4:23 AM:
-------------------------------------------------------------

Stack,

bq. We don't need to add specific types as inputs as you suggest. I like your improvement leaving them generic.

That is the question I had though, they seem to always have been generic and the previous person who worked on the Table classes thought it being important and added it. I am fine with leaving them - or making them - generic, just don't want to step on someone's toes.

OK, will rename classes/methods and add to next patch.

bq. The \n in below is probably in wrong place now that numMapTasks have been removed:

Changed the \n in that line to wrap after each second argument, I think that is what was intended in the first place.

bq. IRW is all ^Ms as are other classes. Not important. Probably not you.

I trust that is my Eclipse saving with stupid Windows line endings. Switched it to use Unix line endings and will save them again.

bq. Only objection I had to 1548 was this: scan.setMaxVersions();

Yeah, I was not sure why that was in there and will remove as you say to leave it up to the caller to set if needed.

bq. Some lines are > 80 characters. Convention is to try and keep them < 80 if makes sense.

Wrapped all lines where I could. I had an issue with JavaDoc @see lines that when wrapped would cause a JavaDoc compile error. So I unwrapped them. Let me know if that is an issue.



      was (Author: larsgeorge):
    Stack,

bq. We don't need to add specific types as inputs as you suggest. I like your improvement leaving them generic.

That is the question I had though, they seem to always have been generic and the previous person who worked on the Table classes thought it being important and added it. I am fine with leaving them - or making them - generic, just don't want to step on someone's toes.

OK, will rename classes/methods and add to next patch.

bq. The \n in below is probably in wrong place now that numMapTasks have been removed:

Changed the \n in that line to wrap after each second argument, I think that is what was intended in the first place.

bq. IRW is all ^Ms as are other classes. Not important. Probably not you.

I trust that is my Eclipse saving with stupid Windows line endings. Switched it to use Unix line endings and will save them again.

bq. Only objection I had to 1548 was this: scan.setMaxVersions();

Yeah, I was not sure why that was in there and will remove as you say to leave it up to the caller to set if needed.

bq. Some lines are > 80 characters. Convention is to try and keep them < 80 if makes sense.

Warpped all lines.


  
> 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-v2.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.