Is this what you want ?
A = LOAD 'data/input' USING PigStorage(',') AS
(name:chararray, aa:float, bb:float, cc:float, dd:float);
B = FILTER A BY cc > 0;
C = GROUP B BY name;
D = FOREACH C GENERATE group,SUM(B.cc)/SUM(B.aa)/SUM(B.bb) as allsum;
E = JOIN B By name,D BY group;
F = FOREACH E GENERATE name,cc*dd/allsum;
DUMP F;
On Thu, Feb 25, 2010 at 3:04 PM, Jumping <[email protected]> wrote:
> Hi Jeff,
>
> Thank you for your reply so quickly.
> I would like to do calculations (cc * dd) * SUM (cc) / SUM (B.aa) / SUM
> (B.bb) on 1.csv.
>
>
> Best Regards,
> Jumping Qu
>
> On Thu, Feb 25, 2010 at 2:50 PM, Jeff Zhang <[email protected]> wrote:
>
> > I am not quit understand your script. For each D there is a allSum, so
> what
> > is the relationship between B and allSum in your last statement.
> >
> >
> >
> > On Thu, Feb 25, 2010 at 2:33 PM, Jumping <[email protected]> wrote:
> >
> > > Hi,
> > > I have this script, how to achive results ?
> > >
> > >
> > >
> > >
> >
> ====================================================================================
> > > A = LOAD 'file:///home/hadoop/1.csv' USING PigStorage(',') AS
> > > (name:chararray, aa:float, bb:float, cc:float, dd:float);
> > >
> > > B = FILTER A BY cc > 0;
> > >
> > > C = GROUP B BY name;
> > >
> > > D = FOREACH C GENERATE group,
> > > SUM(B.cc)/SUM(B.aa)/SUM(B.bb) as allsum;
> > >
> > > E= FOREACH B GENERATE cc*dd/allsum;
> > >
> > >
> >
> =====================================================================================
> > >
> > >
> > > The last one could not work.
> > >
> > > Looking forward to experts input. Thanks.
> > >
> > >
> > > Best Regards,
> > > Jumping Qu
> > >
> >
> >
> >
> > --
> > Best Regards
> >
> > Jeff Zhang
> >
>
--
Best Regards
Jeff Zhang
