Hbase Custom Splitters in MapReduce

[Nytrix Analytics]

Hi,

Does anybody have an example on how to implement a custom splitter on MapReduce? By default, only one mapper is being spawned

per region.

I want to increase the number of mappers per region. Im using TableMapReduceUtil.initTableMapperJob in my MR job.

thanks.

[Harsh J Cloudera Employee]

Take a look at how the deprecated HBase's TableInputFormat (mapred API) created on-demand splits: https://github.com/cloudera/hbase/blob/cdh5.1.0-release/hbase-server/src/main/java/org/apache/hadoop...

You can basically apply the same logic within the mapreduce API TableInputFormat#getSplits(…) by inheriting and overriding it.