You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@pig.apache.org by Marek Miglinski <mm...@seven.com> on 2011/12/02 13:37:40 UTC
FOR - EACH?
Hello guys,
I have a group of users aggregated by day period, lets assume there are 10mill users and each user has got 1000 transactions.
Now is there a way to iterate through this group NOT foreach, but foreach 10! records in that group? So I could safe space and count more users per action, export less records.
Thanks,
Marek M.
Re: FOR - EACH?
Posted by pablomar <pa...@gmail.com>.
what about?
b = group a by ...
c = foreach b {
generate ...
}
On 12/2/11, Marek Miglinski <mm...@seven.com> wrote:
> Hello guys,
>
> I have a group of users aggregated by day period, lets assume there are
> 10mill users and each user has got 1000 transactions.
>
> Now is there a way to iterate through this group NOT foreach, but foreach
> 10! records in that group? So I could safe space and count more users per
> action, export less records.
>
>
>
> Thanks,
> Marek M.
>
RE: FOR - EACH?
Posted by Marek Miglinski <mm...@seven.com>.
That was a good advice Dima, thanks ;)
Marek M.
________________________________________
From: Dmitriy Ryaboy [dvryaboy@gmail.com]
Sent: Saturday, December 03, 2011 1:06 AM
To: user@pig.apache.org
Subject: Re: FOR - EACH?
group by (user_id, semi-random-number_between-0-and9) ?
On Fri, Dec 2, 2011 at 4:37 AM, Marek Miglinski <mm...@seven.com> wrote:
> Hello guys,
>
> I have a group of users aggregated by day period, lets assume there are 10mill users and each user has got 1000 transactions.
>
> Now is there a way to iterate through this group NOT foreach, but foreach 10! records in that group? So I could safe space and count more users per action, export less records.
>
>
>
> Thanks,
> Marek M.
Re: FOR - EACH?
Posted by Dmitriy Ryaboy <dv...@gmail.com>.
group by (user_id, semi-random-number_between-0-and9) ?
On Fri, Dec 2, 2011 at 4:37 AM, Marek Miglinski <mm...@seven.com> wrote:
> Hello guys,
>
> I have a group of users aggregated by day period, lets assume there are 10mill users and each user has got 1000 transactions.
>
> Now is there a way to iterate through this group NOT foreach, but foreach 10! records in that group? So I could safe space and count more users per action, export less records.
>
>
>
> Thanks,
> Marek M.
Re: FOR - EACH?
Posted by Thejas Nair <th...@hortonworks.com>.
I haven't understood what you are trying to do. Can you explain with an
example of input and output ?
-Thejas
On 12/2/11 4:37 AM, Marek Miglinski wrote:
> Hello guys,
>
> I have a group of users aggregated by day period, lets assume there are 10mill users and each user has got 1000 transactions.
>
> Now is there a way to iterate through this group NOT foreach, but foreach 10! records in that group? So I could safe space and count more users per action, export less records.
>
>
>
> Thanks,
> Marek M.