Date:         Fri, 4 May 2007 11:21:28 -0400
Reply-To:     Wensui Liu <liuwensui@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Wensui Liu <liuwensui@GMAIL.COM>
Subject:      Re: cumulative product?
Comments: To: celiacai@gmail.com
In-Reply-To:  <1178191211.064832.152770@e65g2000hsc.googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

sorry, celiaca,
there is small mistake in my previous hash code. following should be correct.

data _null_;
  if 0 then set test;

  dcl hash h (hashexp: 16);
  h.definekey ('VarA', 'VarB');
  h.definedata('VarA', 'VarB', 'product');
  h.definedone();

  do until (eof);
  set test end = eof;
    if h.find() ~= 0 then do;
      product = 1;
    end;
    product = product * VarC;
    h.replace();
  end;

  rc = h.output(dataset: 'product');
run;


On 5/3/07, celiacai@gmail.com <celiacai@gmail.com> wrote:
> Hi
>
> I knew there's certain way to calculate subgroup cumulative sum in a
> table...but how can I calculate the cumulative product of subgroups
> within a dataset?
>
> Say, the ovservations are sorted by variable A,B and C (namely
> subgrouped twice), and I want to find out the product of all Cs within
> each B-group. How could I do that?
>
> celia
>


--
WenSui Liu
A lousy statistician who happens to know a little programming
(http://spaces.msn.com/statcompute/blog)