proc sort data= proc sort data = comp_3 out = comp_4; by pid pl_date descending bill_amount; run; data want; set comp_4; by pid pl_date; retain fst_pl2; if first.pid then fst_pl2=fst_pl; run; ... View more

Seems like you are not saving the final1 in sas library, you need to update the proc transpose step to use out=sas.final1 to again use the sas.final1 in later steps.   proc transpose data = fp out = sas.final1 by Scheme name; var performance; run; ... View more

@UcheOkoro you question is not clear, you said you want to create 3 new variables but its not clear. However if you want to concatenate all the drugs in drug1-drug48 variables then you can do as below with arrays   data want; set merged_data2; array drug_list(*) $ drugflag1-drugflag48; array drug_name(*) $ drugnam1-drugnam48; do i = 1 to dim(drug_list); if drug_list(i)='Y' then drug_name(i)=vname(drug_name(i)); if drug_name(i)='opioid' the opioid='Yes'; if drug_name(i)='benzo' the benzo='Yes'; if drug_name(i)='anticholinergic' the anticholinergic='Yes'; end; all_drugs=catx(' ', of Drug_nam[*]); run; ... View more

please try the proc stdsize procedure, here the missing values within the group will be replaced with the method specified. Since you have the method as median, the missing values within the group will be replaced with the median of that group as we used by statement as group. The reponly option does help in replacing only the missing values with median. A very good procedure perfect for your requirement and it helps you avoid the step2.   PS it is an untested code as i am not sure of your, so please test and let me know if it is working. Per your response I am assuming that we need to keep only kilograms between 40 and 200 inclusive and then get the median value and then replace the missing values with median.   data nonmissing missing; set cleandata36; if 40<=kilograms<=200 then output nonmissing; if missing(kilograms) then output missing; run; data cleandata36_2; set nonmissing missing; run; proc sort data=cleandata36_2; by group; run; proc stdize data=cleandata36_2 out=results.output36 method=median reponly; var kilograms; by group; run; ... View more

Please try the below code , however more information is required on indicator to program   data have; input Acc_no IndustryCode$; cards; 283 MB 286 PL 286 IN 286 CE 286 NL 286 FL 287 VF 287 CL 287 NL ; proc sql; create table want as select count(Acc_no) as count, Acc_no from have group by Acc_no; quit; ... View more

You can try the below code   proc sql; create table work.W10Z35LY as select A,B,C from work.W10Y7DVC union all select A,B,C from work.W10VUM5G; quit; ... View more