You can run 0.4, but it will have the same issue.
0.5+ are only compatible with Hadoop 0.20+ (in fact 0.5 is essentially just
0.4 with the 0.20 compatibility patches).

You can run in hadoop mode, and have hadoop running in pseudo-distributed
mode locally. That should work.

-D

On Sat, Feb 27, 2010 at 10:43 AM, Cory Radcliff <[email protected]>wrote:

> Oh. Ok, I was thinking that gzip didn't work on local files, but its local
> mode.
>
> Thanks. Can I run newer versions of Pig on older versions of Hadoop?
>
> On Fri, Feb 26, 2010 at 10:39 PM, Jeff Zhang <[email protected]> wrote:
>
> >  pig 0.3 do not support load gz file in local mode
> >
> >
> >
> > On Fri, Feb 26, 2010 at 8:17 PM, Cory Radcliff <[email protected]>
> > wrote:
> >
> > > I've been up and down the docs, and I see people using GZipped files.
> But
> > > when I try to load them, i get garbage. Basically it loads it as raw
> data
> > > from the local file system.
> > >
> > > test = LOAD 'file:///home/hadoop/testme1.
> > > gz'  using PigStorage('\u0002');
> > >
> > > dump test;
> > >
> > > Even when I load it from s3, still no dice.
> > >
> > > I'm using pig on amazon. Is it too old for this functionality?
> > >
> > > Apache Pig version 0.3.1-amzn (r2485701)
> > > compiled Aug 10 2009, 11:52:03
> > >
> > > Hadoop 0.18
> > > Subversion  -r
> > > Compiled by root on Sat Jan 16 02:29:24 UTC 2010
> > >
> >
> >
> >
> > --
> > Best Regards
> >
> > Jeff Zhang
> >
>

Reply via email to