Hi,

I'd like to format the output of some pig transforms to be stored as 
hive-readable partitions.

I'm able to get it to work using Multistorage (in which I write files based 
upon the partition keys).  However, it's not quite there because while it's 
writing into a directory for each partition key, it's storing the key in the 
file as well!

So before I write a UDF based upon Multistorage just for writing a Hive 
partition (which'll squelch writing partition key columns), I'm curious what 
other resources/techniques are available.

Thanks,

Ranjan

Reply via email to