Date:         Fri, 11 Apr 2008 21:27:41 -0400
Reply-To:     Muthia Kachirayan <muthia.kachirayan@GMAIL.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Muthia Kachirayan <muthia.kachirayan@GMAIL.COM>
Subject:      Re: how to deal with multiple ID
In-Reply-To:  <3a237784-cd86-45d6-a20f-491e08221473@t54g2000hsg.googlegroups.com>
Content-Type: text/plain; charset=ISO-8859-1

On Fri, Apr 11, 2008 at 7:44 PM, AnoniDeer <irain168@gmail.com> wrote:

> I have a dataset like this.How could I get the difference between the
> third and fourth observations for each of the group? for example for
> the group 1 I need to get the difference for the observation of 45 and
> 34. For the second group I need the difference of 76-34 and so on...
> How to write a program to realize this? Thanks so much!
>
> group x
> 1 23
> 1 .
> 1 34
> 1 45
> 2 45
> 2 89
> 2 34
> 2 76
> 4 23
> 4 12
> 4 23
> 4 12
>

Here is a lengthy program. It does not assume that there should be 4
observations per group and that they need  be together within group. Hash
object is used to sense the group change. When the 3rd and 4th observations
within a group are met they are used to find the difference and saved to
output data set.


data have;
 input group x;
cards;
1 23
1 .
1 34
1 45
2 45
2 89
2 34
2 76
4 23
4 12
4 23
4 12
;
run;

data need;
if _n_ = 1 then do;
   if 0 then set have;
   declare hash h();
   h.definekey('group');
   h.definedone();
end;
do until(eof);
   set have end = eof;
      if h.find() ne 0 then do;
      count = 1;
      x3 = 0;
      h.add();
   end;
   else do;
      count ++ 1;
      if count = 3 then x3 = x;
      if count = 4 then do;
         x_34 = x - x3;
         output;
      end;
   end;
end;
stop;
drop count x3;
run;
proc print data = need;
run;

Muthia Kachirayan