How can I combine my macro variable conditions to recode a variable in a data step? I try not to use macro variables when part of the data step but even when I change the code below to %do %if/%then outside of the dataset I run into errors, plus I'd rather run one data step rather than a data step for each value of a.    I have a  list of variable names in list a and a list of variable names in list b. I need to do a loop within a loop to check that the names in list a = names in list b. If they are equal then I want to recode the variable name in list a to equal the element in a different list, b2 that is associated with list b.      %let a = gender, age  ;  /*(in this example, yes, commas)*/ %let a_n = 2 ; %let b= age height weight gender ; %let b_n = 4; %let b2= "unk1" "unk2" "unk3"  "unk4";  /*(in my example they are quoted)*/   Data want ; Data have ; do x = 1 to &a_n. ;   do y=1 to &b_n. ;       if scan(&a, &x, ',') = scan(&b., &y.) then do ;         if scan(&a,&x,' ') = '' then scan(&a,&x)= scan(&b2,&y.) ;         end;      end ; end ; run;   I want the above to resolve to :   data want ; data have ; if age = '' then age="unk1" ; if gender = "" then gender="unk4" ; run;   ... View more

I am trying to loop through variable names in a list to determine if it meets a condition. If it does then I need to revise the list of variables in  a variable named grp_lst. I would like to replace the variable name in the list with blank which by the end of the loop I will have a final variable list. Here is my sample code. I am using this in a macro so I created a little test macro around the code.  Let's assume that the &actual = 2 for Age and that the value in the drug1 dataset is 'age gender type' . After this code I would expect to see 'gender type' in the grp_lst.    Variables that are needed and that are created prior to this code:      %let var_lst = age gender type ; %let varCount = 3 ; %let expected = 3 3 6 ; /*expected value for each variable in the var_lst */ %macro test(); %DO k=1 %to &varCount; /*get the actual count for each variable in variable list*/ proc sql NOPRINT; Select count(distinct %Scan(&var_lst, &k)) as %Scan(&var_lst, &k)_lvls into :actual separated by ' ' from drug1; quit; /* create revised grp_lst by removing variables in the list if actual < expected */ /* currently data in grp_lst is equal to var_lst for all observations - just the way it was created..*/ DATA drug1; set drug1; if &actual. < %Scan(&expected., &k.) then grp_lst=%sysfunc(compress(%sysfunc(tranwrd(grp_lst, "%Scan(&var_lst,&k)", " ")))); run; %END; %MEND; %test();      This is what gets resolved in the log for the grp_list statement in the datastep: MPRINT(test 😞 if 2 < 3 then grp_lst = grp_lst ;       ... View more