It looks like your code can be converted to SAS data step code like this:   1. Get the FULL_DATE from TBL: Proc sql noprint; SELECT MAX(full_date) into :full_date from tbl; quit; 2. Process the data: data want; set final; by id rownum; if rownum=1 then do; f_dose=fill_dt; l_dose=min(fill_dt+fill_days_supply-1,&full_date); span=1; end; else do; f_dose=min(max(fill_dt,l_dose+1),&full_date+1); if fill_dt-l_dose>=60 then span+1; l_dose=min(max(fill_dt,l_dose+1)+fill_days_supply-1,&full_date); end; retain l_dose span; format l_dose f_dose date9.; run; The code shown assumes that the tables exist in SAS.WORK, and that the dates are SAS date values. The FINAL table should be sorted by ID and ROWNUM. The code also assumes that rows exist for consecutive ROWNUM values (1,2,3...) for all IDs. ... View more

@Sathish_jammy I strongly recommend to amend your code as per the highlighted lines below.  If your condition &count>0 is false then your macro will currently still generate a run; statement.   As for the SQL the following code would suffice and populate macro var &record_cnt the same as your code. proc sql; select count(*) into :record_cnt trimmed from table1 where id not in (select id from table2); quit;     ... View more

There is no automated way to migrate SAS code to other software tools. SQL will be much the same in the new tool but you will have to change SAS table and column names to whatever you have defined in BigQuery. DATA step code will need to be rewritten in SQL or other supported BigQuery languages. ... View more

data have; input Have $; cards; 6.9. 6.9 7.8. 7.8 6.2. 6.2 8.9.2 8.92 4.5.2 4.52 6.5.7 ; data want; set have; call scan(have,-1,p,l,'.'); want=cats(substr(have,1,l-1),substrn(have,p)); drop p l; run; ... View more

data have; infile datalines dlm=";" dsd truncover; input (Pv_Drugs Fv_Drugs) (:$30.); datalines4; Insu;Met,Sulp,SGLT Insu;Sulp Insu,DPP;Met Insu,DPP;Sulp,DPP,Thiaz Met;Insu,Met Met;Met,Sulp,Thiaz Met,Sulp;Met,Sulp,AGI Sulp,AGI;Met,AGI Sulp,DPP;Met,Sulp,SGLT Thiaz;Met,Sulp,AGI AGI;Insu,Met,Sulp ;;;; data want; set have; length effect $300; do i = 1 to countw(fv_drugs,", "); temp=scan(fv_drugs,i,','); if not findw(Pv_Drugs,strip(temp),', ') then effect=catx(',',effect,temp); end; drop i temp; run; ... View more

Also, what exactly do you mean by a "trend line"? Elsewhere you use the word "regression", but a "trend line" and "regression line" don't have to be the same thing. ... View more

If you are into drug development or evaluation, you should use the procedures and methods suggested by your statistician.   If this is an exercise, what have you tried so far? ... View more

I see no start_date in that screenshot. Please post data in usable form (I will NOT repeat that request for a third time), and post the complete log of your data step by copy/pasting it into a window opened with the </> button. Do not, I repeat NOT, post logs or code in pictures. Use the "little running man" button right next to the </> for code. ... View more