Logic help...

[SJN]

Hi All,

Input					Output
EXPOSURE	LOSS	FLAG			EXPOSURE	LOSS	FLAG
0	10	0			0	10	0
1	10	1			1	10	1
2	20	0			2	20	0
3	20	1			3	20	1
4	30	0			4	30	0
5	40	0			5	40	0
6	50	0			6	50	0
7	50	1
8	50	1
9	50	1
10	50	1

Above is my test data. We are looking into data from 2014 Actually we are having exposure till 60. 

Ex:

For Current month (Aug16): Exposure ranges from 0 to 30

       prior month (jul16):Exposure ranges from 0 to 27 ( Current month -3)

       pre-prior month (Jun16) : Exposure ranges from 0 to 24 ( prior month -3)

Like this it goes on....

I created a variable lag and flag. 

if loss=lag(Loss) then Flag=1; else Flag=0;

I couldn't able to get logic. If I take only Aug16 data exposure is till 60 from row 31 value is duplaicated. I dont want to cosider that duplicated which are from 31 row . How you help how to achive this logic. 

[Kurt_Bremser]

So, from a series of exposure values, you only want to keep the first one with the highest loss (and discard all later obs with equal loss), and all observations up to that?

[SJN]

Thank you for the reply. I'm trying to keep the value from low to high.

My data looks like this.

Exposure	Loss
0	0
1	0
2	0.000001978
3	0.000001978
4	0.000001978
...........	0.000053544
...........	...............
29	0.0091486
30	0.0092539
31	0.0092539
32	0.0092539
...........	...............
...........	...............
60	0.0092539

[Kurt_Bremser · Accepted solution]

data have;
input exposure loss;
cards;
1 0
2 0
3 .1978
4 .1978
5 .1978
6 .2
7 .2
8 .2
9 .2
;
run;

proc sql noprint;
select max(loss) into :maxloss from have;
quit;

data want;
set have;
retain flag 1;
if flag;
if loss = &maxloss then flag = 0;
drop flag;
run;

proc print noobs;
run;

Result:

exposure     loss

    1       0.0000
    2       0.0000
    3       0.1978
    4       0.1978
    5       0.1978
    6       0.2000