%let end=1808; %let n=24; data have; do i=0 to &n by 1; do yymm=intnx('month',input(cats(&end,'01'),yymmdd6.),-i); format yymm yymmn4.; output; end; end; run; ... View more

Hi I tried the below and it worked with no issues data have; input date:anydtdte.; weeknum = ceil(day(date) / 7); format date date9.; cards; 20021030 ; ... View more

data have; input date:anydtdte.; week=intck('week',intnx('month',date,0),date)+1; format date date9.; cards; 20180825 ; ... View more

please try the below code    if there is only year like 2011 then I imputed it to 01-01, let me know if that works for the other imputation suggested is followed.   data have; input id date1 $10. date2 $11. ; if length(date1)<10 then cd1=intnx('month',input(cats(date1,'-01'),yymmdd10.),0,'e'); else if length(date1)>=10 then cd1=input(date1,anydtdte10.); if length(date2)<=4 then cd2=intnx('year',input(cats(date2,'-01-01'),yymmdd10.),0,'s'); else if 4< length(date2)<10 then cd2=intnx('month',input(cats(date2,'-01'),yymmdd10.),0,'e'); else if length(date2)>=10 then cd2=input(date2,anydtdte10.); if cd2<cd1 then want='early'; else want='late'; format cd: date9.; datalines; 1 05/21/2010 2013-03-21 2 2011-03-01 2011-01-21 3 2014-03 2011 4 2015-01 2015-05 ; ... View more

  data have; input Sample:$ v1; cards; A 1 B 1 B 2 C 1 ; proc sql; create table want as select sample,sum(v1) as sum, v1 from have group by sample having sum=1; quit; ... View more

Alternatively you could try the below code   data want; length new_on_off new_CATEGORY new_group $100; set have; by id ; retain new_on_off new_CATEGORY new_group; if first.id then new_on_off=''; if first.id then new_CATEGORY=''; if first.id then new_group=''; if on_off ne '' and new_on_off ne on_off then new_on_off=catx('_',new_on_off,on_off); if CATEGORY ne '' and new_CATEGORY ne CATEGORY then new_CATEGORY=catx('_',new_CATEGORY,CATEGORY); if Group ne '' and new_group ne Group then new_group=catx('_',new_group,Group); if last.id; keep new_on_off new_CATEGORY new_group; run;     ... View more

It is possible replace the insert with union as below   proc sql; create table want as select mem length=100,sum(mem in ('F' 'C')) as freq, sum(percent) as percents from have where mem in ('F' 'C') group by mem union select 'Grand Total' as mem length=100, count(*) as freq, sum(percent) as percents from have; quit; ... View more

As per my understanding functions would be a best approach, alternatively you could create a customized function by proc fcmp ... View more

then please something as below   data foo; format y percent7.1; y=input(tranwrd('-0,01259870',',','.'),best.); run; proc print data=foo; run; ... View more