%let last_sat=%sysfunc(intnx(week.7,%sysfunc(today()),0,b),date9.); %put last_sat=&last_sat ; ... View more

You didn't show your code using Proc Transpose, but I believe with some dataset options,  Proc transpose ALONE can produce the outcome you listed:     data have; infile cards truncover ; input application_ID status 'total income'n :dollar10. Type 'code Desc'n & $18.; cards; 74628 1 $45,787 3 main income 74628 2 $4,689 2 other income 74628 3 $57,812 1 one off income 213514 1 $12,789 3 main income 213514 2 $54,713 2 other income 213514 3 $13,574 1 one off income ; run; PROC TRANSPOSE DATA=WORK.have OUT=WORK.want(drop=_name_ where=('main income'n ne .)) LET ; BY application_ID; ID "code Desc"n; VAR "total income"n; COPY "total income"n; run; ... View more

You need to speak to the SAS sales rep in charge of your area. SAS software licensing is a compound processing based on many factors, and in many cases, negotiable. Your Mileage May Vary, a lot. ... View more

Rules seem not entirely clear to me. 1. What if you have more than 2 groups within the same id, do you only count the first one? 2.  What if you have more than 2 obs that are >=200 after the first >=170. Are they going to the first group or you kick off a new group/just ignore it? 3. What if first >=170 hitting the bottom, do you consider itself as the average or ignore it? ... View more

Now I have no clue what you need. The 6 is the length of the interval when dose is above 15 (>15), the unit is 'hour', if you need more accuracy, you can easily switch it to 'minutes'. 6 hours is between first time point when dose>15, which is 07JUN2017:21:56:00,  to the last time point where dose > 15, which is 08JUN2017:03:14:00.   " it should be  07JUN2017:21:56:00 - 07JUN2017:19:39:00"? This is when the dose is not above 15, you did not mention you need that as well, but that can be easily done.   The thing is, there is a conflict between your description and your wanted outcome (this part is confusing).     ... View more

data want; do until (last.var); set have; by id var notsorted ; if first.id and var=1 then flag=1; if flag ne 1 then output ; end; run; ... View more

It would be great if you can offer the sample data in data step in the future. Not knowing your wanted outcome, I would do this in two steps, first step is to flag the eligible period, after this, there could be more than one approaches:   data have; input ID TIME MDYAMPM15. DOSE; format time datetime25.; cards; 101 6/7/17 7:39 PM 14 101 6/7/17 8:48 PM 15 101 6/7/17 9:56 PM 16 101 6/8/17 12:15 AM 18 101 6/8/17 1:12 AM 20 101 6/8/17 3:14 AM 14 101 6/8/17 6:42 AM 13 101 6/8/17 7:38 AM 6 101 6/8/17 8:06 AM 7 101 6/8/17 9:03 AM 8 102 6/8/17 10:18 AM 12 102 6/8/17 11:02 AM 10 102 6/8/17 12:55 PM 11 102 6/8/17 3:17 PM 20 102 6/8/17 4:07 PM 21 102 6/8/17 5:48 PM 14 102 6/8/17 7:21 PM 14 102 6/8/17 8:05 PM 13 ; run; data h1; set have; by id; if dose>15 or (lag(dose)>15 and not first.id) then flag=1;else flag=0; run; data want; do _n_=1 by 1 until (last.flag); set h1; by id flag notsorted ; if _n_=1 and flag then _beg=time; end; time_abve_15=intck('hour', _beg,time); do _n_=1 by 1 until (last.flag); set h1; by id flag notsorted ; output; end; drop _beg; run; ... View more