how to jump over specific row under cumulative sum?

[km0927]

hello. i made data set, counting patient's visiting number(except admission) before they got surgery.

raw data is here.

id date admission_YN surgery_YN

1 2002/01/28 0 0
1 2002/02/15 0 0
1 2002/04/06 1 0

1 2002/05/05 0 0
1 2002/06/07 1 1
1 2002/08/10 0 0
1 2003/01/12 0 0
2 2003/01/15 0 0
2 2003/02/10 1 0
2 2003/03/10  1 1
2 2003/04/08 0 0

and with this code, i made a data set. but i can't exclude(=jump over) visiting for admission.

proc sort data=have;
by id date;
run;
data want;
retain had_surgery;
set have;
by Id date;
if first.id then
do;
had_surgery = 0;
count = 0;
end;
if surgery_yn = 0 and had_surgery = 0 then count+1;
else if surgery_yn = 1 then
do;
had_surgery = 1;
count=0 ;
end;
run;

id date admission_YN surgery_YN count

1 2002/01/28 0 0 1
1 2002/02/15 0 0 2
1 2002/04/06 1 0 3

1 2002/05/05 0 0 4
1 2002/06/07 1 1 0
1 2002/08/10 0 0 0
1 2003/01/12 0 0 0
2 2003/01/15 0 0 1
2 2003/02/10 1 0 2
2 2003/03/10  1 1 0
2 2003/04/08 0 0 0

i want make data set like this. how can i jump over admission event?

id date admission_YN surgery_YN count

1 2002/01/28 0 0 1
1 2002/02/15 0 0 2
1 2002/04/06 1 0 2

1 2002/05/05 0 0 3
1 2002/06/07 1 1 0
1 2002/08/10 0 0 0
1 2003/01/12 0 0 0
2 2003/01/15 0 0 1
2 2003/02/10 1 0 1
2 2003/03/10 1 1 0
2 2003/04/08 0 0 0

[PaigeMiller · Accepted solution]

if surgery_yn = 0 and had_surgery = 0 and admission_yn=0 then count+1;

[novinosrin]

data have;
input id date : yymmdd10. surgery_YN;
format date yymmdd10.;
cards;
1 2002/01/28 0 0
1 2002/02/15 0 0
1 2002/04/06 1 0
1 2002/05/05 0 0
1 2002/06/07 0 1
1 2002/08/10 0 0
1 2003/01/12 0 0
2 2003/01/15 0 0
2 2003/02/10 1 0
2 2003/03/10 0 1
2 2003/04/08 0 0
;





data want;
set have;
by id date;
if first.id then do;f=0;count=0;end;
f+surgery_YN;
if f<1 then count+1;
if lag(surgery_YN)=1 and surgery_YN=0 then count=0;
drop f;
run;