Date:         Thu, 7 Sep 2006 16:51:02 -0400
Reply-To:     Peter Constantinidis <peter@CONSTANTINIDIS.CA>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Peter Constantinidis <peter@CONSTANTINIDIS.CA>
Subject:      Re: Trying to figure out the best way to count unique
              observations by variable in a dataset
In-Reply-To:  <7ffd0c20609071321k7c01bae2w8bad0aeb7049a934@mail.gmail.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed

I found code on the web that helps, it will produce 14 observations in
a data set, but it creates a data set.. :( So I would still have to
merge stuff together? so it could be counted as part of proc summary,
or the like? And still have to make a data set for each leave code?



data testing;
set perm.f20032004 (rename=( actualhrs=actlhrs leavecd=leave_code));
where leave_code='952';
by emplid Leave_code;
 * reset score at each change of recno;
if first.emplid then score = 0;
 * increment score at each change in comp, except when comp=0;
if first.leave_code and leave_code <> 0 then score + 1;
 * output the accumulate score before the next change in recno;
if last.emplid then output;

run;

data testing2;

   do until (last.emplid) ;
set perm.f20032004 (rename=( actualhrs=actlhrs
leavecd=leave_code));where leave_code='952';
by emplid leave_code ;
      score = sum (last.emplid) ;
   end ;
run ;