finding consecutive different values across variables

[fama]

Hi all,

I am pretty new in SAS programming and do need your help. thx

i have a data set like this ( 5 variabes and too many observations)

num  a b c d e

1       .  3 4  .  .

2       .  5 .  6 .

3       1 2 2  .  .

4       2 .  .   . 4

desired result

num  a b c d e

1       .  3 4  .  .

2       1 2 2  .  .

Indeed, i want to have just those rows that contain 2 consecutive different values.

i do appreciate your help

[jdwaterman91]

This logic should give you your desired output.

/* Creating the dataset */

Data Input;
Input Num a b c d e ;
Datalines;
1 . 3 4 . .
2 . 5 . 6 .
3 1 2 2 . .
4 2 . . . 4
;
Run;

/* Eliminate rows that do not have consecutive non-missing values */

Data Final;
Set Input;
If Num ^= "" Then do;
If ( (b*a = "") and (c*b = "") and (d*c = "") and (e*d = ""))
Then Delete;
End;
Run;

[fama]

Thanks for your help. it did not work. still shows undesired observations.

[jdwaterman91]

I'm sorry that did not work.  

Just out of curiousity, what is an example of an undesired row that was showing up in your output?

I will say that the array method in the post below is a better way of approaching the problem compared to what I did. 

[Astounding · Accepted solution]

It's a little complex, since it uses arrays.  But not too lengthy:

data want;

set have;

array vars {5} a b c d e;

do _n_=1 to 4;

   if (vars{_n_} ne .) and (vars{_n_+1} ne .) and (vars{_n_} ne vars{_n_+1}) then keep_flag='Y';

end;

if keep_flag='Y';

drop keep_flag;

run;

[fama]

it worked well. really appreciate it.

thanks 🙂