You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hbase.apache.org by Rajeshkumar J <ra...@gmail.com> on 2017/05/12 09:08:35 UTC

mapreduce over hbase table

Hi,

   I am running a mapreduce job over hbase table. I went through some blogs
and in that it is mentioned that data locality will be achieved for this
job other than some cases like region reassignment and others. When I run
mapreduce job over hbase table it takes each region as input for each
mapper. My question is how this job knows which region exists where? can
anyone explain this flow

Thanks

Re: mapreduce over hbase table

Posted by Ted Yu <yu...@gmail.com>.

Please take a look at the getSplits() methods in:

hbase-server/src/main/java/org/apache/hadoop/hbase/mapreduce/TableInputFormat.java
hbase-server/src/main/java/org/apache/hadoop/hbase/mapred/TableInputFormatBase.java

Cheers

On Fri, May 12, 2017 at 2:08 AM, Rajeshkumar J <ra...@gmail.com>
wrote:

> Hi,
>
>    I am running a mapreduce job over hbase table. I went through some blogs
> and in that it is mentioned that data locality will be achieved for this
> job other than some cases like region reassignment and others. When I run
> mapreduce job over hbase table it takes each region as input for each
> mapper. My question is how this job knows which region exists where? can
> anyone explain this flow
>
> Thanks
>