Error while using UTF-8 Unicode Format in Join

[Vijay Kumar]

Hi,

Error Message :
Fatal Error : Virtual dataset; out of <join stage name>; Record to big to fit in a block; length requested : 301924548

I am running a job which uses UTF-8 format. The job consists a join stage.
The Join keys are in Unicode Format ( under Columns tab I have selected Unicode under Extended section for key columns). When ever I run the job, it is getting aborted with the above error message. The join internally uses hash partision and sort. I tried to remove sort from the join stage, but it is of no use.
But If I uncheck the unicode for the key columns under Columns tab, the job is executing successfully.

I am not able to trace the issue, whether it is due to unicode format or anything else. Could any one help me in this issue would be verymuch helpful.

Thanks in advance.

Regards
Vijay