Cool - thanks Sonal.  The docs were just unclear about that, but I guess that 
makes sense since it should be pretty trivial to specify the outputformat as 
part of the job configuration.

On Feb 3, 2011, at 8:19 PM, Sonal Goyal wrote:

> Yeah, you can use any OutputFormat in your EMR job. If it is file based, it 
> will write to the given file output path, else to Cassandra, DB, whatever you 
> specify..
> 
> Thanks and Regards,
> Sonal
> Connect Hadoop with databases, Salesforce, FTP servers and others
> Nube Technologies 
> 
> 
> 
> 
> 
> 
> 
> On Wed, Feb 2, 2011 at 12:54 AM, Jeremy Hanna <[email protected]> 
> wrote:
> Thanks!
> 
> Okay - is there a way to specify a completely different outputformat - I 
> would like to output to my Cassandra cluster in ec2.
> 
> 
> On Feb 1, 2011, at 11:21 AM, Koji Noguchi wrote:
> 
> > Input and output filesystems don’t have to match.
> >
> > You should be able to read from S3 and write to hdfs as long as you give a 
> > path like  s3:// and  hdfs:// .
> >
> > Koji
> >
> >
> > On 2/1/11 11:13 AM, "Jeremy Hanna" <[email protected]> wrote:
> >
> > I wanted to input from s3 but output to someplace else in aws with
> > elastic mapreduce.  Their docs seem to only suggest that they only
> > read from/write to s3.  Is that correct?
> >
> 
> 

Reply via email to