Do you mean, you already have the duplicate_reason variable in your dataset? Hmm, i think i am confused with how your data looks like, See if my understanding is right:   /*Another version*/ data want; set test; by id; length temp $100; retain temp ; if first.id then do; call missing(temp); temp= catx('|',run_id,run_date); end; else temp=catx('~',temp,trim(run_id)||run_date); if last.id and index(duplicate_reason,temp)=0 then duplicate_reason= catx('~',duplicate_reason,temp); keep id duplicate_reason; run;   Regards, Naveen Srinivasan ... View more

data test; length id $3 run_id $3 run_date $10; input id run_id run_date ; cards; 123 ABC 2017/01/01 123 CDE 2017/01/01 ; run;   data want; set test; by id; length duplicate_reason $100; retain duplicate_reason ; if first.id then do; call missing(duplicate_reason); duplicate_reason= catx('|',run_id,run_date); end; else duplicate_reason=catx('~',duplicate_reason,trim(run_id)||run_date); if last.id; keep id duplicate_reason; run;   Regards, Naveen Srinivasan ... View more

    hi.   data have; input CompanyName $ Year                      month                 TYPE $; datalines; Abc                        2016                       1                              A Abc                        2016                       10                           C Abc                        2015                       2                           B ;   proc sort data=have; by companyname year month; run;   data want(rename=(_month=month)); length     _type _type1 $1; do _n_=1 by 1 until(last.year);   set have;   by year ;   if first.year and last.year then    do;       do _month=month to 12;          output;       end;   end;   if _n_=1 and not  last.year then   do;     call missing(_month,_type,_month1,_type1);     _month=month;     _type=type;   end; if _n_>1    then     do;      _type1=type     ;         do _month=_month to month-1;            _month1=month;          type=_type;           output;        end;          type=_type1;          output;     end;  if  (_n_>1 and last.year and month<=12) or (first.year and last.year)  then    do;       do _month=_month+1 to 12;          output;       end;   end; end;  drop month _type _type1 _month1; run;   Regards, Naveen Srinivasan ... View more

If I understand you correctly, this is perhaps what you need:   %LET DrugUse1 = 'ABC' ; %LET DrugUse2 = 'XYZ' ; %LET DrugUse3 = '' ; %LET DrugUse4 = '' ; %LET DrugUse5 = '' ; data want; set have;/*your source input dataset*/ array use[5] DrgUse1-DrgUse5 ; %MACRO DRUG ; %do i = 1 %to 5 ; if drugHave = "&&drugUse&i" then drugUse&i=1 ; %end; %MEND ; run; %DRUG Try and let me know when you can- Regards, Naveen Srinivasan ... View more

Assigning a month format doesn't change the stored values of your date variable that is they still remain as date values displaying with your month format. I would suggest to extract month values using month function and the group by month so that you would get your desired result: /*sample*/ data have; input defined_period ddmmyy10.; format defined_period ddmmyy10.; month=month(defined_period); datalines; 01/03/2017 02/03/2017 01/04/2017 02/04/2017 ; proc sort data=have; by month; run; data want; set have; by month; if first.month then count=0; count+1; if last.month; run;   Regards, Naveen Srinivasan ... View more

hi, data want(drop=_profit); set have(firstobs=min); set have(rename=(profit=_profit)) nobs=_nobs_ point=_nobs_; d=_profit-profit; output; stop; run;   Regards, Naveen Srinivasan ... View more