Benjamin Reed
Wed, 12 Mar 2008 10:26:55 -0700
It's exactly like Hadoop reduce: each task produces one file. So a foreach of a group by with parallel of 100 will produce 100 output files. ben On Wednesday 12 March 2008 09:58:29 Johannes Zillmann wrote: > Hi folks, > > i like to know more about the characteristics of a StoreFunc. > Does it produce always 1 file or can it produce more files (like the > hadoop reduce phase does/can) ? > Does it run as map or reduce task or at the pig-server side ? > > Thank You > Johannes