I added SET mapreduce.task.classpath.user.precedence true; to the top of my
script and it worked.
On Tue, Dec 10, 2013 at 11:36 PM, Serega Sheypak
> You have to specify classpath precedence. See here
> 11.12.2013 7:02 пользователь "Shawn Hermans" <[EMAIL PROTECTED]>
> > I am having an issue with a Pig UDF I wrote. The Pig UDF uses a newer
> > version of Guava than is bundled with the version of Hadoop I am using
> > 4.2). I explicitly register the newer version using a REGISTER
> > However, it still loads the version bundled with CDH. How do I get Pig
> > recognize the and load the correct version of Guava?
> > Thanks,
> > Shawn