I have a bag of tuples like this:

{ (product, unwanted, count), (product, unwanted, count) }

Is it possible in Pig to generate a new bag with a revised tuple structure
with one of its columns removed?

The desired structure I want is:

{ (product, count), (product, count) }

Thanks,
James

FLATTEN() the bag, re-project (foreach/generate) leaving out the unwanted items, then group back together if you like.

_____________
Steve Bernstein
VP, Analytics
Rearden Commerce, Inc.

+1.408.499.0961 Mobile

deem.com | reardencommerce.com
-----Original Message-----
From: James Newhaven [mailto:[EMAIL PROTECTED]]
Sent: Wednesday, May 09, 2012 10:43 AM
To: [EMAIL PROTECTED]
Subject: Removing unwanted items in tuple

I have a bag of tuples like this:

{ (product, unwanted, count), (product, unwanted, count) }

Is it possible in Pig to generate a new bag with a revised tuple structure with one of its columns removed?

The desired structure I want is:

{ (product, count), (product, count) }

Thanks,
James

Thanks. But what if my bag is grouped like this:

group: (id), bagname: { (product, unwanted, count), (product, unwanted,
count) }

I want to retain the group tuple  but remove the "unwanted" column from the
bag.

I don't think I can perform a flatten here without losing the group tuple?

Thanks,
James

On Wed, May 9, 2012 at 7:47 PM, Steve Bernstein <[EMAIL PROTECTED]>wrote:

> FLATTEN() the bag, re-project (foreach/generate) leaving out the unwanted
> items, then group back together if you like.
>
> _____________
> Steve Bernstein
> VP, Analytics
> Rearden Commerce, Inc.
>
> +1.408.499.0961 Mobile
>
> deem.com | reardencommerce.com
>
>
> -----Original Message-----
> From: James Newhaven [mailto:[EMAIL PROTECTED]]
> Sent: Wednesday, May 09, 2012 10:43 AM
> To: [EMAIL PROTECTED]
> Subject: Removing unwanted items in tuple
>
> I have a bag of tuples like this:
>
> { (product, unwanted, count), (product, unwanted, count) }
>
> Is it possible in Pig to generate a new bag with a revised tuple structure
> with one of its columns removed?
>
> The desired structure I want is:
>
> { (product, count), (product, count) }
>
> Thanks,
> James
>

You can use a nested foreach to project only the wanted columns.

Cheers,
--
Gianmarco
On Wed, May 9, 2012 at 10:22 PM, James Newhaven <[EMAIL PROTECTED]>wrote:

> Thanks. But what if my bag is grouped like this:
>
> group: (id), bagname: { (product, unwanted, count), (product, unwanted,
> count) }
>
> I want to retain the group tuple  but remove the "unwanted" column from the
> bag.
>
> I don't think I can perform a flatten here without losing the group tuple?
>
> Thanks,
> James
>
> On Wed, May 9, 2012 at 7:47 PM, Steve Bernstein <[EMAIL PROTECTED]
> >wrote:
>
> > FLATTEN() the bag, re-project (foreach/generate) leaving out the unwanted
> > items, then group back together if you like.
> >
> > _____________
> > Steve Bernstein
> > VP, Analytics
> > Rearden Commerce, Inc.
> >
> > +1.408.499.0961 Mobile
> >
> > deem.com | reardencommerce.com
> >
> >
> > -----Original Message-----
> > From: James Newhaven [mailto:[EMAIL PROTECTED]]
> > Sent: Wednesday, May 09, 2012 10:43 AM
> > To: [EMAIL PROTECTED]
> > Subject: Removing unwanted items in tuple
> >
> > I have a bag of tuples like this:
> >
> > { (product, unwanted, count), (product, unwanted, count) }
> >
> > Is it possible in Pig to generate a new bag with a revised tuple
> structure
> > with one of its columns removed?
> >
> > The desired structure I want is:
> >
> > { (product, count), (product, count) }
> >
> > Thanks,
> > James
> >
>