Dear community, In my data I want to use a counter variable that counts under special conditions. The counter should start to count everytime primary is equal to 4 until it stops being 4. In the case primary = 5 prior to primary = 4 the counter shouldn't start. The same goes (backwards looking)  for a primary=5 which appears after a sequence of primary=4 (the counter should be set to zero for the whole sequence before). Think of it as the employment status where the counter counts the time in unemployment (primary=4). While this duration is just interesting if the individual was in the labor force before and stays in after, I don't want the duration to be count if primary = 5 (appears when the individual is not in the labor force). Here is an example(normally there are 15000 individuals and time is  up to 1500): data work.sample; input id $ primary $ time ($); datalines; 1 5 1 1  5  2 1  4  3 1  4  4 1  4  5 1  1  6 2  1  1 2  1  2 2  4  3 2  4  4 2  4  5 2  5  6 3  1  1 3  4  2 3  4  3 3  4  4 3  1  5 3  1  6 ; I am able to set up the counter but so far I have no clue how to bring in my conditions on primary = 5 and the following and earlier periods. data test1;        set test1;        if primary = 4 and lag(primary) ^= 4 then duration = 1;        if primary = 4 and lag(primary) = 4 then duration+1;       if primary ^= 4 then duration = 0; run; The output in the end should look like this: input id $ primary $ time $ duration ($) 1 5 1 0 1  5  2  0 1  4  3  0 <- no count after because of the primary=5 1  4  4  0 1  4  5  0 1  1  6  0 2  1  1  0 2  1  2  0 2  4  3  0 2  4  4  0 2  4  5  0 2  5  6  0  <- no count before because of the primary=5 3  1  1  0 3  4  2  1 3  4  3  2 3  4  4  3 3  1  5  0 3  1  6  0 I would be really thankful for any help! ... View more