topic Nested Do Loop Terminating After 1 Row in SAS Programming

Nested Do Loop Terminating After 1 Row

cobba — Tue, 14 Dec 2021 00:39:16 GMT

I'm trying to create a new column for each code I'm looking for. Then flag each row that contain those codes.

The code works for 1 row then terminates. How do I keep looping through each row of my data set?

data have;
  input id$ code1$ code2$ code3$ code4$;
  cards;
aa 03f 06x 01a 05a
bb 05n 02b 01a .
cc 02b . . .
dd 01a 02b . .
;
run;

data top_codes;
  input top_code1$ top_code2$ top_code3$;
  cards;
01d 01a 02b
;
run;

data want (drop= i k);
	
	if _n_ eq 1 then;
		do;
			set top_codes;
			retain _all;
			array tc(*) top_code1--top_code3;
		end;

	set have;

	array _vars(*) code1--code4;
	array code_{3}; 
 	
	do i=1 to dim(_vars);
		do k = 1 to dim(code_);
			if _vars(i) = tc(k) then code_{k} = 1 ;
		end;
	end;
run;

This is a image of what I'm getting. I'm expecting to get a line for each row of my data

Re: Nested Do Loop Terminating After 1 Row

ballardw — Tue, 14 Dec 2021 00:48:43 GMT

How about providing a more complete example of what you expect. I am not all sure of what you want.

Note: You are getting one row of data for each line of date in the Top_codes data set. Multiple Set statements are a complex subject.

If your goal is to have every record from Have matched to the Top_code data set then you want a Cartesian product such as

proc sql;
   create table want as
   select a.*,b.*
   from top_codes as a, have as b
   ;
quit;

If your top_codes had two rows of data you have every row of top_codes joined with every row of Have yielding 2 times as many records as Have.

Re: Nested Do Loop Terminating After 1 Row

Tom — Tue, 14 Dec 2021 01:13:29 GMT

You have an extra semi-colon.

Yet another reason to dislike that indentation style. Keep the DO on the end of the IF line and align the END with the IF

Try this instead:

data want (drop= i k);

  if _n_ eq 1 then do;
    set top_codes;
    retain _all;
    array tc(*) top_code1--top_code3;
  end;

  set have;

  array _vars(*) code1--code4;
  array code_{3}; 

  do i=1 to dim(_vars);
    do k = 1 to dim(code_);
      if _vars(i) = tc(k) then code_{k} = 1 ;
    end;
  end;
run;

You don't even need a DO/END block at that point because the SET statement is the only executable statement.

  if _n_ eq 1 then set top_codes;

  retain _all;
  array tc(*) top_code1--top_code3;

I am not sure what the RETAIN statement is doing. You are not creating any variable named _ALL. And if there is a variable named _ALL coming in from one of those two datasets it would already be retained so no need to list it in a RETAIN statement. If you meant to use the _ALL_ variable list then that also would do nothing as the only variables defined at that point are the ones read from TOP_CODES and those are already set to be retained.