Thanks guys. That's what I figured out eventually. It works well in Apache 
Hadoop and CDH but not so well in Datastax package (DSE 2.1.1) though. I filed 
a bug report to Datastax on that. 

Cheers,
-Keeyong 

> Subject: Re: How to read Mahout generated sequence files in Pig
> From: [email protected]
> Date: Wed, 6 Feb 2013 15:37:05 -0500
> To: [email protected]
> 
> The elephant bird sequence file loader should work, you'll just need to 
> register the mahout jar with the vector writable they use.
> 
> David
> 
> On Feb 4, 2013, at 7:06 PM, Harsha <[email protected]> wrote:
> 
> > keeyong, 
> >    we used elephantbird( https://github.com/kevinweil/elephant-bird ) from 
> > twitter to read/write sequence files.
> > Take a look at these classes 
> > com.twitter.elephantbird.pig.store.SequenceFileStorage, 
> > com.twitter.elephantbird.pig.load.SequenceFileLoader.
> > 
> > -- 
> > Harsha
> > 
> > 
> > On Monday, February 4, 2013 at 3:58 PM, keeyong han wrote:
> > 
> >> I am wondering how I can read Mahout generated sequence files in Pig? I 
> >> guess there might be a UDF but I can't find one yet.
> >> 
> >> Cheers,
> >> -Keeyong
> >> 
> >> 
> > 
> 
                                          

Reply via email to