I am still having trouble getting this to work and some parts have changed.
I have to match times within 5 seconds the first time around, and take those that didn't match and match them within 15 seconds. Anything that does not match the second time, will be dropped from the final dataset for those variables (missing data) .
I also have 13 additional variables that are associated with the times, so when I match up the times, those associated variables also have to stay with those times.
So, i would have ID, time1-time37, timef1-timef37, a1-a37, b1-b37, c1-c37, etc.
If I take real data:
id time1 time2 time3 time4 time5 timef1 timef2 timef3 timef4 timef5
1 5:56:33 6:01:44 6:10:47 6:16:00 6:17:55 5:56:17 6:01:45 6:10:35 6:15:58 6:17:54
2 2:05:03 2:14:05 2:18:48 2:31:15 . 2:05:04 2:08:55 2:14:00 2:18:44 2:23:35
3 8:16:19 8:18:18 8:21:57 8:25:04 8:27:41 8:16:18 8:18:07 8:20:17 8:25:04 8:27:20
Some match within 5 seconds, some within 15 seconds and some not at all.
In the end, I need to have the times ordered 1-37 in time order, but only if they match and with associated time descriptive variables (a1-a37, b1-b37, etc).
So there will be the same amount of time1-time37 values as timef1-timef37 all matching one another where Time1=timef1 time2=timef2, etc. within 5 or 15 seconds depending on how it was selected.
What I envision doing with SAS is matching up the times within 5 seconds, output to a dataset, match up the leftover times within 15 seconds (not including those that are already matched within 5) and ouput them to another dataset. Remove all other times, merge back the other two datasets and sort and reorder the variables by time. The times should all fall within an ~30 minute time range.
I've been mulling over this for a week and writing code with no success other than to determine which times match and which do not.
Thanks again!