Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Pig >> mail # dev >> Question about loader and storer


Copy link to this message
-
Re: Question about loader and storer
I agree. PigStorage is the default constructed by LogicalPlanBuilder and
it's not configurable.

Jeff, can you open a JIRA? It would be a nice feature to add.

-Prashant

On Thu, Feb 21, 2013 at 6:26 PM, Aniket Mokashi <[EMAIL PROTECTED]> wrote:

> I think default loader is hardcoded in the pig code. You can open a jira if
> you need such a feature.
>
> Thanks,
> Aniket
>
>
> On Thu, Feb 21, 2013 at 6:08 PM, Johnny Zhang <[EMAIL PROTECTED]>
> wrote:
>
> > Jeff:
> > Basically, edit the pig.properties to
> > .....
> > pig.load.default.statements=/tmp/.temppigbootup
> > .....
> >
> > and in file /tmp/.temppigbootup, you have load statement
> > data = LOAD 'top_queries_input_data.txt' AS (query:CHARARRAY, count:INT);
> >
> > you can edit content to use other loader here.
> >
> > Hope it is helpful. This is different from what you want, and I am also
> > searching if we can define default loader other than PigStorage.
> >
> > Johnny
> >
> >
> > On Thu, Feb 21, 2013 at 6:03 PM, Johnny Zhang <[EMAIL PROTECTED]>
> > wrote:
> >
> > > Hi, Jeff Yuan:
> > >
> > > On Thu, Feb 21, 2013 at 5:53 PM, Jeff Yuan <[EMAIL PROTECTED]>
> > wrote:
> > >
> > >> Hi,
> > >>
> > >> I am new to the pig community, and have a couple of questions
> > >> regarding loader/storers.  Because I am writing code that closely
> > >> couples with Pig, I thought I'd ask on the dev mailing list, but
> > >> please let me know if this is more appropriate for the user list.
> > >>
> > >> 1) Is there a way to set the default pig loader as something other
> > >> than PigStorage via configuration? What I mean is, by default if a
> > >> loader is not specified, PigStorage is assumed. Can I change things so
> > >> that if no loader is specified in the load statement, another custom
> > >> loader is used?
> > >>
> > > I am not sure if there is another way, but you can edit "
> > > pig.load.default.statements=" in pig.properties file. So in your Pig
> > > script you don't have to write load statement, but Pig will always load
> > it
> > > for you (of course with the loader you specified)
> > >
> > >>
> > >> 2) Is there a way to output the entire query result to stdout? Is the
> > >> easiest way to do so by writing a custom storer that doesn't actually
> > >> store to a file but instead just output to stdout?
> > >>
> > > can you try 'DUMP' ?
> > >
> > >>
> > >> If no specific answer is available, advice on where in the pig code I
> > >> should look to discover answers for myself would also be highly
> > >> appreciated.
> > >>
> > >> Thanks,
> > >> Jeff
> > >>
> > >
> > > Johnny
> > >
> >
>
>
>
> --
> "...:::Aniket:::... Quetzalco@tl"
>