Thank you again!! ... View more

Thank you so much! I have accepted the first responder's solution since he/she responded to the original question, but your solution works great as well! 🙂 Thanks! ... View more

Thank you again! ... View more

That works great. Thank you so much!! ... View more

I see. Thank you! I would like the final dataset to look as shown below. If can first subset the dataset so that it only includes the two most recent test dates, then the next step of calculating days_between might use less memory, as my dataset is large   ID testdate days_between 1  6/4/17  88 1  8/1/17  88 2  7/5/16  31 2 10/1/16  31   I was wondering if you could advise me on an additional data step that would allow me to subset the dataset as I have described. Thank you very much for your help! ... View more

Thank you so much for your reply! I ran your code as well and get the same issue. If you could see my reply to draycut, and advise me on how to address those issues, I would really appreciate it! Thank you! ... View more

Thank you very much for your response!   I ran the code. It seems to be working for ID 2 but not for person ID 1. The difference between 8/1/17 and 6/04/17 is 58 days but the code seems to be giving 88 days.  I was wondering if you are getting the same thing and if so how would one correct it.    Also, on a related topic, I was wondering if you could advise me on how to first subset the dataset so that the new dataset contains only the two most recent test dates per person.   Thank you very much! ... View more

Thank you for your help! I needed it for a GWAS problem where there are lots of multiple comparisons 😃    Code that worked was: ods output ParameterEstimates=pval1; proc reg data= abc; model aa = x1 x2 ; run; proc contents data=pval1; /*to identify label for pvalue*/ run; data pval2; set pval1; format Probt pvalue15.13; run;   ... View more

Thank you very much for your response. I think your code seems to be exactly what I need. Curiously, the 'overlap' variable calculated using the long sql code above agrees with your code (i.e. the difference between the two dates identified using min and max) agrees about 99% of the time but not a 100%. I may have to troubleshoot why that's happening but your code seems to be more accurate for calculating 'overlap' 🙂 Thank you again! ... View more

Thank you so much Art for your help! I am trying to extract the actual dates that were used to compute 'overlap' so that I would have two more columns containing the overlap start date and the overlap end date. The post by PGStats seems to be working. Thank you again for your help! Appreciate it! 🙂 ... View more

Thank you very much for your response. It is just that in some cases, the 'overlap' column is the difference between have.start_dt and have2.end_t, or between have2.start_dt and have.end_dt, or have.start_dt and have.end_dt or have2.start_dt and have2.end_dt. To illustrate:   overlap is the difference between dates a and b      a. _______ b. c.______________d.   or   overlap is the difference between dates c and d a.__________________b.        c.________d.   or   overlap is the difference between dates c and b a.____________b.          c._______________d.   or   overlap is the difference between dates a and d               a.___________b. c.____________d.   I was wondering if there is a way to output the two dates that were used to compute the overlapping days. Thank you very much for any help you could give! ... View more

Thank you very much for your help once again. I was wondering if there is a way to modify the sql code so that it outputs 2 additional columns containing the 2 dates that were used to compute that difference/overlapping days. I can do it for two combinations in a simple separate data step, but I found that it gets quite hairy when considering 3-5 combinations. Thank you very much! ... View more