Question

I have a dataset that looks something this:

    Unique_ID Date    
    1         03/23/1995
    1         03/27/1995
    1         04/14/1995
    1         08/29/1995
    1         02/14/1996
    .         .
    .         .
    .         .
    1         03/19/1997
    2         10/20/1993
    .         .
    .         .
    .         .
    2         04/20/2000

Within each Unique_ID I need to keep only those observations that have dates at least 3 months apart starting with the first observation (my data set is sorted by Unique_ID and Date). For example, within Unique_ID 1, I would need to keep the next observation that is at least 90 days from 3/23/1995, then the next observation that is 90 days from that day and so on. Could anybody point me in the right direction either with a macro or some kind of loop?

Was it helpful?

Solution

You can try something like this:

data want;
   set have;
      by ID;
   retain date2find;
   if first.ID then do;
      output;   /* This statement will include the first obs from each    */
                /* set of IDs if desired.  If not, delete this statement. */
      date2find = DATE + 90;
      end;

   if DATE >= date2find then do;
      output;                /* Output the found record    */
      date2find = DATE + 90; /* Re-set to date to be found */
      end;
   drop date2find;
run;

This relies on your dataset being sorted as described (by ID and DATE)

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top