URL fetching from paragraph

[singhsahab]

Hi All,

In my variable there are multiple URL were entered , but not in sequence. I need to filter all the URL in new dataset. Below is the sample data . 

data have;
infile cards ;
input url & $5000.;
cards;
https://www.youtube.com/w ahsjskjk https://www.yahoo.com jhjkj https://www.google.com
;
run;


Output want:

URL
https://www.youtube.com/w  https://www.yahoo.com  https://www.google.com

Thank you in advance 🙂 

[Kurt_Bremser · Accepted solution]

Count through "words" separated by blanks, identify those beginning with the wanted token, and concatenate to a new string:

data have;
infile cards ;
input url & $5000.;
cards;
https://www.youtube.com/w ahsjskjk https://www.yahoo.com jhjkj https://www.google.com
;

data want;
set have (rename=(url=_url));
length url $5000;
do _i = 1 to countw(_url," ");
  _word = scan(_url,_i," ");
  if substr(_word,1,4) = "http" then url = catx(" ",url,_word);
end;
drop _:;
run;

[ed_sas_member]

Hi @singhsahab 

Here is another approach:

proc sql noprint;
	select max(countw(url,' ')), max(count(url,"http")) into:nb1,:nb2 from have;
quit;

data want;
	set have;
	array _URL (&nb1) $ 200;
	array URL_ (&nb2) $ 200;
	do i=1 to countw(url,' ');
		if prxmatch('/http/',scan(url,i,' ')) then _URL(i) = scan(url,i,' ');
	end;
	URL_extract = catx(',',of _URL(*));
	do j=1 to count(url,"http");
		URL_(j) = scan(URL_extract,j,',');
	end;
	drop i j URL_extract _:;
run;