The SQL approach will maintain all table attributes like indexes so it's imho the cleanest way of coding a change of variable lengths. ... View more

I like the use of data step functions, since they are as efficient as it can be. And wrapped into a macro, it can be as flexible that you want. But if "is empty" means exists with no rows, perhaps you need to use the ATTRN() function instead. ... View more

As a followup to anotherdream's comment the following would also work, with your macro variable in the where clause. data want;   informat Date_of_Birth mmddyy10.;   format Date_of_Birth mmddyy10.;   input Date_of_Birth;   cards; 1/1/2012 1/1/2012 1/1/2013 1/1/2013 1/1/2014 1/1/2014 1/1/2015 1/1/2015 1/1/2016 1/1/2016 ; Run; proc sql noprint;   select distinct Date_of_Birth format=best8.           into :date  separated by ','   from want ; quit; data test; set want; where Date_of_Birth in (&date); run; ... View more

You'll need to do a better job explaining your goal here.  The sample code you provided has many flaws, including referring to a data set "a" that doesn't exist, and referring to a variable "original_amt" that doesn't exist.  There are other problems, such as the effects of the IF statement.  You would be better off showing the "after" picture to match the "before" data in your post.  There would be many posters willing to help at that point. Good luck. ... View more

Try below code : Will get the same result as required. data tourrevenue; input Country :$ 1-11 LandCost Vendor $ customer; cards;   France       575 Express  10   Spain        510 World    12   Brazil       540 World     6   Japan        720 Express  10   Greece       698 Express  20   Venezuela    425 World     8   Italy        468 Express   9   Russia       924 World     6   Switzerland  734 World   20   Ireland      558 Express   9   ;  Run; Proc sort data = tourrevenue ;by vendor; run; data Percen; input Vendor $ Customer Percentages; cards; Express 58 0.52727 World 52 0.47273 ; run; proc sort data = percen; by vendor; run; data merge_both;   merge  tourrevenue(in=aa) percen(In=bb); by Vendor; if aa = 1; New_Landcost =  LandCost*Percentages;  run; ... View more

If what you mean "trailing" is to find how many blanks are padded, you could try: lengthc(product)-length(product); Haikuo ... View more

Borrowed Art's idea: Data Table_3b (drop=in:);   merge Table_1 Table_2 (in=in2 rename=(SalesQ=in_S Price=in_P));   By ID;   if in2 then do;     SalesQ=in_S;     Price=in_P;   end; Run; ... View more