As @Reeza suggested provide sample data sets that clearly describe information contained for a proper solution. Based on information provided seems like you are looking for something like this:   data A; input Namecountry $20.; datalines; abcCanada defUS ghqCanada ;   data B; input Rank name $ score avg_score; datalines; 1 abc 23 20 2 def 21 23 ;   proc sql; create table want as select b.*,substr(Namecountry,length(name)+1) as Country from B left join A on substr(Namecountry,1,length(name))=name; quit; ... View more

Seems like you are trying to calculate lift based on a model. Try the following code that will split data into 10 groups based on descending values of Response_score variable.      proc rank data=have groups=10 descending out=ranked; var Response_score; ranks decile; run; ... View more