indexing repeated variables

[HBa]

i have a data set where 20 study participants went through 24 tests (8 conditions each repeated three times). my data base consists of the subject number (idsub) and test condition (either "cond" or "fact"). i like to add another column to indicate repeated conditions (REPT). I appreciated any suggestion. the data below is from one subject (idsub=1) and REPT is the variable that i would like to add.

idsub cond    fact           REPT
1          1       NOEO         1
1          3       YOBL          1
1          2       NOEC         1
1          4      YOSC          1
1          3      YOBL           2
1          2      NOEC         2
1          1      NOEO         2
1         4      YOSC          2
1         3      YOBL          3
1         2      NOEC        3
1        4      YOSC         3
1        1      NOEO        3

[Reeza]

Is the rule 4 1's, 4 2's, 4 #s etc? or something else?

If it's every 4 something like this:

data want;
set have;
retain count 0;

if mod(_n_, 4) = 1 then count+1;

run;

[PeterClemmensen · Accepted solution]

If I understand you correctly, then here is how I would do it. I added data for idsub=2 and handled multiple values of the idsub variable as well. Let me know if it works for you.

data have;
input idsub cond fact $;
datalines;
1 1 NOEO
1 3 YOBL
1 2 NOEC
1 4 YOSC
1 3 YOBL
1 2 NOEC
1 1 NOEO
1 4 YOSC
1 3 YOBL
1 2 NOEC
1 4 YOSC
1 1 NOEO
2 1 NOEO
2 3 YOBL
2 2 NOEC
2 4 YOSC
2 3 YOBL
2 2 NOEC
2 1 NOEO
2 4 YOSC
2 3 YOBL
2 2 NOEC
2 4 YOSC
2 1 NOEO
;

data want(drop=rc);                     
   if _n_ = 1 then do;          
      dcl hash h ();             
      h.defineKey ('cond', 'fact');       
      h.defineData ('REPT');      
      h.defineDone ();           
   end;
   set have;           
   by idsub;
   if first.idsub then do;
      rc=h.clear();
      REPT=0;
   end;
   if h.find() ne 0 then REPT=1;
   else                  REPT+1;
   h.replace();                 
run;