As incremental update mechanism that utilizes sliding window concept, I think it is fine.
But for real-time stream processing, I can foresee a lot of workarounds to make it work using batch-oriented map/reduce. Pi On 3/18/08, Chris Olston <[EMAIL PROTECTED]> wrote: > > As Utkarsh said, this is a road we started down a while back, but did > not continue. > > If people have use-cases for stream processing in the pig/hadoop > context, I'd like to hear about it. > > -Chris > > > On Mar 17, 2008, at 1:54 PM, pi song wrote: > > > Dear everyone, > > > > I've just recognized this file "TimeWindowSpec.java" in Pig. Is > > anyone > > currently working on stream processing? > > > > Cheers, > > Pi > > -- > Christopher Olston, Ph.D. > Sr. Research Scientist > Yahoo! Research > > >
