Conditionally concatenating multiple variables (do loop/iterative macro)

[A_Kh]

Hi Community, 

I would appreciate if someone could share her/his experience in resolving following problem;

In a given dataset there is a character variable with different values in each row (multiple rows). After transposing all records vertically, i need to concatenate all values into a single variable, separated by a comma. As variable numbers differ based on the row numbers, I'm unable to build a programming logic using loop or macro. 

eg. 

have:

var1    var2     var3    var4;

a          b          c        d

need:

var1

a,b,c,d

I used macro var in summing these similar numeric variables, like ;

data need;

set have;

var1= sum(of var1-var&num);

run; 

Now for character variable, trying to use concatenation,  like;

data need;

set have;

var1= catx(",", var1, var&num);... but it does not cover multiple variables...

[Patrick · Accepted solution]

If your variables follow some naming pattern - something you can chose if using Proc Transpose - then below code should work for both character and numeric variables.

data have;
  infile datalines dlm='|' dsd truncover;
  input (var_1-var_4) ($);
datalines;
a|b|c|d
1||3
;

data want;
  set have;
  length new_var $20;
  new_var=catx(',',of var_:);
run;

[A_Kh]

Hi Patrick,

Thank you so much, it was so simple and worked perfectly!