Hi experts,   This is with regard to a programming query to calculate the change from minimum of the previous visit values. The calculation would check each visit against all prior visits for smallest value and calculate a percentage of change. An example would be something like visit 3 checks visit 1 and 2 and get the minimum of visit 1 and visit 2 to derive the change. I tried the below code in datastep and i am getting the expected output.   Would like to check if there are any better approaches in datastep or proc sql   Subjid Visit Value 6 Visit1 75 6 Visit2 82 6 Visit3 14 6 Visit4 73 6 Visit5 67 6 Visit6 61 6 Visit7 38 proc sort data=have out=have2(keep=subjid visit value); by subjid visit ; run; data want; length new new2 $100; set have2; by subjid visit; retain new; if first.subjid then new=''; if value ne . then new=catx(',',new,value); new2=lag(new); if first.subjid then new2=''; array cnt(15) x1-x15; do i=1 to 15; cnt(i)=input(scan(new2,i,','),best.); end; base=min(of x1-x15); if nmiss(value,base)=0 then pchg=((value-base)/base)*100; drop new new2 x: i; run; Subjid Visit Value BASE PCHG 6 Visit1 75 . . 6 Visit2 82 75 7 6 Visit3 14 75 -61 6 Visit4 73 14 59 6 Visit5 67 73 -6 6 Visit6 61 67 -6 6 Visit7 38 61 -23   ... View more