Row count based on change in two variables

[Q1983]

data have1;

input ln $ setup_dt Date9. ;

format setup_dt mmddyy10.;

datalines;

12000 29Jul2019

12000 30Jul2019

12000 30Jul2019

12000 31Jul2019

12500 31Aug2019

15600 1Aug2019

15700 2Aug2019

;run;

data want;

set have1;

by ln ;

if first.ln then group_id+1;

run;

Output

ln	setup_dt	group_id
12000	07/29/2019	1
12000	07/30/2019	1
12000	07/30/2019	1
12000	07/31/2019	1
12500	08/31/2019	2
15600	08/01/2019	3
15700	08/02/2019	4
Desired as follows
ln	setup_dt	group_id
12000	07/29/2019	1
12000	07/30/2019	2
12000	07/30/2019	2
12000	07/31/2019	3
12500	08/31/2019	1
15600	08/01/2019	1
15700	08/02/2019	1

The desire is that when the same ln shows a new setup_dt, it increments from 1 to 2 or 3 etc depending on the number of dates.  It then should reset to 1 when the ln changes

[novinosrin · Accepted solution]

data want;
set have1;
by ln setup_dt;
if first.ln then group_id=1;
else if first.setup_dt then group_id+1;
run;