Dear SAS-L:

I have a dataset with 4 variables, the last variable bothtime is SAS datetime. I need get a new variable, which values are depending on the differences between bothtime variable. I need to give a new variable (interval)=1 for the first obs in each patno, in this case obs 1, 9, 17, as they are these first obs in patno 301, 302, 303, then get differences between the first bothtime and each individual bothtime for each observation, if the difference is <=24 hr, interval=1, if 25-48 hrs, interval=3, if 49-72 interval=3.

OBS PATNO SAMPLE BOTHTIME

1 301 1 1238583300 2 301 2 1238627400 3 301 3 1238648580 4 301 4 1238674500 5 301 5 1238718600 6 301 6 1238755200 7 301 7 1238799300 8 301 8 1238842800 9 302 1 1239793200 10 302 2 1239837300 11 302 3 1239859200 12 302 4 1239879900 13 302 5 1239923220 14 302 6 1239966180 15 302 7 1240009200 16 302 8 1240052400 17 303 1 1240491900 18 303 2 1240534800 19 303 3 1240557300 20 303 4 1240578300 21 303 5 1240624200 22 303 6 1240664700 23 303 7 1240707600 24 303 8 1240751700

So the resulted dataset will be something like this:

OBS PATNO SAMPLE BOTHTIME Interval 1 301 1 1238583300 1 2 301 2 1238627400 1 3 301 3 1238648580 1 4 301 4 1238674500 2 5 301 5 1238718600 2 6 301 6 1238755200 3 7 301 7 1238799300 3 8 301 8 1238842800 3 9 302 1 1239793200 1 10 302 2 1239837300 1 11 302 3 1239859200 2 12 302 4 1239879900 2 13 302 5 1239923220 2 14 302 6 1239966180 3 15 302 7 1240009200 3 16 302 8 1240052400 8 17 303 1 1240491900 1 18 303 2 1240534800 1 19 303 3 1240557300 1 20 303 4 1240578300 1 21 303 5 1240624200 2 22 303 6 1240664700 3 23 303 7 1240707600 3 24 303 8 1240751700 3

Thanks in advance for your help.

John Zhang