topic Re: How to give more than 1 loop control variable to a do loop? in SAS Programming

How to give more than 1 loop control variable to a do loop?

PrudhviB — Wed, 07 Sep 2022 16:03:06 GMT

Hi SAS experts,

I am trying to add 3 control conditions to a do loop to run but not able to figure it out.

what i have is a pseudo logic from where I have to build a sas code to run through my data to find first instance of a specific number in the sequence.

for this example lets use 2. attached the pseudo logic and sample data.

set sequence_event to 0
set in_sequence_position to -1
set k to 0

repeat until (( character in sequence >= 2 or blank) or 
		(k = length of squence)):

{
get next character in sequence 
increment in_sequence_position by 1
increment K by 1 
}

if character in sequence > 2 
then 
increment s by 1 and 

in_sequence_position + 24
}
else if s = 'n/a'
}
return
(s,
in_sequence_position)

sample Data

id	sub id	sequence
1	1	1111111111111111XX111111111111121111121111111111
1	2	111111111111111111111111111111111111111111111111
1	3	1
2	1	554321125555555555555555555555555555555555555555
3	2	111111111111111111111111111111111111111111X11111
3	1	554321111111111111111111111111111111555555555555
4	1	111111111111111111111
4	2	11
4	3	111211111111111111111111111111111111111111111111

desired output

id	sub id	sequence	S	in_sequence_position
1	1	1111111111111111XX111111111111121111121111111111	1	32
1	2	111111111111111111111111111111111111111111111111	0	0
1	3	1	0	0
2	1	554321125555555555555555555555555555555555555555	1	5
3	2	111111111111111111111111111111111111111111X11111	0	0
3	1	554321111111111111111111111111111111555555555555	1	5
4	1	111111111111111111111	0	0
4	2	11	0	0
4	3	111211111111111111111111111111111111111111111111	1	4

any help is much appreciated.

Re: How to give more than 1 loop control variable to a do loop?

Tom — Wed, 07 Sep 2022 15:38:47 GMT

Is that input supposed to be a representation of a dataset structured like this one?

data have;
  input id subid sequence :$50. ;
cards;
1 1 1111111111111111XX111111111111121111121111111111
1 2 111111111111111111111111111111111111111111111111
1 3 1
2 1 554321125555555555555555555555555555555555555555
3 2 111111111111111111111111111111111111111111X11111
3 1 554321111111111111111111111111111111555555555555
4 1 111111111111111111111
4 2 11
4 3 111211111111111111111111111111111111111111111111
;

What is the desired output for that input?

Re: How to give more than 1 loop control variable to a do loop?

Tom — Wed, 07 Sep 2022 15:45:07 GMT

I cannot understand what you want.

But to your question you can include a WHILE condition (or an UNTIL condition) on top of a normal start/stop interval.

do k=1 to length(sequence) while ( <some stopping condition);
...
end;

Re: How to give more than 1 loop control variable to a do loop?

PrudhviB — Wed, 07 Sep 2022 15:48:54 GMT

i should have included full pseudo code

this is after the do loop

if character in seq > 2 
then 
increment s by 1 and 

in_sequence_position + 24
}
else if s = 'n/a'
}
return
(s,in_sequence_position)

my desired out put should hve the "s" and "in_sequence_position" values returned.

Re: How to give more than 1 loop control variable to a do loop?

Reeza — Wed, 07 Sep 2022 15:51:36 GMT

You could do that but have you seen FINDC() or FINDW()?

Do you want to learn how to implement your logic specifically or would you like to learn how the SAS methods of accomplishing your objective?

If the later, it may be easier to specify the logic you'd like to implement. Using built in optimized functions is definitely more efficient and faster.

@PrudhviB wrote:

Hi SAS experts,

I am trying to add 3 control conditions to a do loop to run but not able to figure it out.

what i have is a pseudo logic from where I have to build a sas code to run through my data to find first instance of a specific number in the sequence.

for this example lets use 2. attached the pseudo logic and sample data.
set sequence_event to 0
set in_sequence_position to -1
set k to 0

repeat until (( character in sequence >= 2 or blank) or 
		(k = length of squence)):

{
get next character in sequence 
increment in_sequence_position by 1
increment K by 1 
}
sample Data

id sub id sequence

1 1 1111111111111111XX111111111111121111121111111111

1 2 111111111111111111111111111111111111111111111111

1 3 1

2 1 554321125555555555555555555555555555555555555555

3 2 111111111111111111111111111111111111111111X11111

3 1 554321111111111111111111111111111111555555555555

4 1 111111111111111111111

4 2 11

4 3 111211111111111111111111111111111111111111111111

any help is much appreciated.

Re: How to give more than 1 loop control variable to a do loop?

ballardw — Wed, 07 Sep 2022 15:55:34 GMT

You might describe why you are doing this and what the actual expected output would be.

I would guess that you may be doing something related to sequential values. Which, if that is the case might be easier, depending on actual want, than processing each character.

For instance if you want to find out how many 1's are in the first sequence such as "1111111111111111XX111111111111121111121111111111"

For your consideration:

data example;
   x="1111111111111111XX111111111111121111121111111111";
   sequences= countw(x,'1','k');
   do i= 1 to sequences;
      thisseq =scan(x,i,'1','k');
      lengthseq = countc(scan(x,i,'1','k'),'1');

      output;
   end;
run;

In this example the Countw function uses everything except the 1 as a delimiter. So it says there are 4 sequences containing only the character 1. Then the Scan function can extract the actual sequence. The Countc function counts how many characters. Actually with all a single character the Length function would work as well.

So, what is it you are actually trying to do?

Re: How to give more than 1 loop control variable to a do loop?

Tom — Wed, 07 Sep 2022 16:02:00 GMT

my desired out put should hve the "s" and "in_sequence_position" values returned.

I don't know what you mean by those two quoted words.

Also what is the actual values you expect to get for those two variables for the actual input you provided?

Here: Just replace the missing values in this data step and re-post to show what answers you expect.

data want;
  input id subid sequence :$50. s in_sequence_position;
cards;
1 1 1111111111111111XX111111111111121111121111111111 . .
1 2 111111111111111111111111111111111111111111111111 . .
1 3 1 . .
2 1 554321125555555555555555555555555555555555555555 . .
3 2 111111111111111111111111111111111111111111X11111 . .
3 1 554321111111111111111111111111111111555555555555 . .
4 1 111111111111111111111 . .
4 2 11 . .
4 3 111211111111111111111111111111111111111111111111 . .
;

Re: How to give more than 1 loop control variable to a do loop?

Tom — Wed, 07 Sep 2022 16:29:55 GMT

So now that you have posted what you expect as output it looks like you just want to find the first occurrence of the digit 2.

You also appear to want to create an extra variable is is true when the location was found (not sure what value that adds).

That is want the INDEX() or INDEXC() function will do.

data have;
  input id subid sequence :$50. s in_sequence_position;
cards;
1 1 1111111111111111XX111111111111121111121111111111 1 32
1 2 111111111111111111111111111111111111111111111111 0 0
1 3 1 0 0
2 1 554321125555555555555555555555555555555555555555 1 5
3 2 111111111111111111111111111111111111111111X11111 0 0
3 1 554321111111111111111111111111111111555555555555 1 5
4 1 111111111111111111111 0 0
4 2 11 0 0
4 3 111211111111111111111111111111111111111111111111 1 4
;

data want;
  set have;
  location = indexc(sequence,'2');
  found = location > 0 ;
run;

proc print;
  var id subid s found in: location sequence;
run;

Result

                                    in_sequence_
Obs    id    subid    s    found      position      location    sequence

 1      1      1      1      1           32            32       1111111111111111XX111111111111121111121111111111
 2      1      2      0      0            0             0       111111111111111111111111111111111111111111111111
 3      1      3      0      0            0             0       1
 4      2      1      1      1            5             5       554321125555555555555555555555555555555555555555
 5      3      2      0      0            0             0       111111111111111111111111111111111111111111X11111
 6      3      1      1      1            5             5       554321111111111111111111111111111111555555555555
 7      4      1      0      0            0             0       111111111111111111111
 8      4      2      0      0            0             0       11
 9      4      3      1      1            4             4       111211111111111111111111111111111111111111111111

Re: How to give more than 1 loop control variable to a do loop?

Tom — Wed, 07 Sep 2022 16:35:55 GMT

If you did want to re invent the wheel using your own DO loops here is one way.

data want;
  set have;
  found=0;
  do location=1 to length(sequence) while (char(sequence,location) ne '2');
  end;
  if location > length(sequence) then location=0;
  else found=1;
run;