How to extract the numbers from the text?

[ybz12003]

Good morning,

I would like to extract the numbers from the sample dataset 'Have'.   I list the result I would like to get in the 'Want' dataset.

data Have;
      infile datalines delimiter='#';
  input Comments : $200.  ;
datalines;
	UGESTAGE=34 5/7 WKS #
	UGESTAGE=35WKS #
	UGESTAGE=33.4 WKS #
	UGESTAGE= 24 WKS
;
run;

data Want;
      infile datalines delimiter='/';
  input UGESTAGE  ;
datalines;
	34 /
	35 /
	33 /
	24
;
run;

[gamotte · Accepted solution]

Hello,

One way :

want=input(prxchange("s/^[^\d]*(\d+).*$/$1/",1,Comments), best.);

[ybz12003]

Super!

But what is '^[^\d]*(\d+).*$/$1'?

And what is '1' in front of the 'Comments'?

[gamotte]

It is a perl regular expression (regexp).

https://perldoc.perl.org/perlre

^ : beginning of the string
$ : end of the string
\d a digit
[^\d] any non digit
[^\d]* any sequence of non-digits
\d+ any non empty sequence of digits
(\d+) => the matching string can be referred to in the replace string with $1 ($2, $3, ... if there are other (...) in the regexp)
.* any sequence of characters

So the command searches for a string begining with any sequence of non digits followed by a sequence
of digits (which is given the alias "$1") followed by any sequence of characters.
The input string is replaced by the matching sequence of digits.

prxchange SAS function allows to use perl regular expressions to search/replace elements in a string.
https://documentation.sas.com/?docsetId=lefunctionsref&docsetTarget=n0r8h2fa8djqf1n1cnenrvm573br.htm...

[ybz12003]

Thank you!

[Ksharp]

data Have;
      infile datalines delimiter='#';
  input Comments : $200.  ;
datalines;
	UGESTAGE=34 5/7 WKS #
	UGESTAGE=35WKS #
	UGESTAGE=33.4 WKS #
	UGESTAGE= 24 WKS
;
run;

data Want;
 set have;
 want=scan(comments,1,,'kd');
run;

[ybz12003]

Somehow, I used PERI code is not working properly.   I changed to use Ksharp's code, it works fine.   Thanks much!