Mapper that creates multiple rows for one tuple

[dbahir]

I have a situation where it would be convenient if the HBase mapper class create multiple rows per one tuple. Working a project where each item we get from a Kafka queue could be written to 1 to 4 HBase tables depending on the content.

Do you see a benefit of adding such capability or should this be handled by bolts upstream of the HBase bolt that will chop this up and multiple ?