Matching Time variables plus or minus 5 seconds

[statadm]

Yes, essentially, I have to rearrange the time variables (no matter the location) until they match based on the criteria I indicated. But, I do have to keep the 13 additional variables associated with the time variables when I reorder the times. In the end, I have to actually compare the 13 variables to one another (13 variables for Time1-Time37 and compare the Timef1-Timef37), not the times. The times are just being used to match up the correct events. This is for a reliability study. I'm trying to get some help from a fellow programmer because this is taking me too long 🙂


Message was edited by: statadm

[Flip]

I did it this way;
Transpose your data and split it such that you have a data set time with:
ID Time Var1 - Var13
data set timef with
ID TimeF
Do a cartesian join;

create table diff as
select time.*, timef.timef , abs(time-timef) as diff, min(abs(time-timef) as mindif
from time, timef where time.id = timef.id
group by time.id, time;

now select diff.* from diff where diff = mindif;

You now have the time difference on the rows with all of your variables with times matched to the closest timef. If diff is more than 5 or 15 you can drop that row.

[statadm]

Thank you all for your recommendations. I actually do have a code that works now. I appreciate everyone's input so much.

The code is pretty long, so I'm not going to post it unless anyone is interested in seeing it.