Finding a value in text from a variable that contains a list of values

[SIgnificatif]

Hi Dear All.

I have this working code code :

data have;
re="all";
ro="hello";
ra=catx('',ro,re);
texto='blablabla hello all, people';
if index(texto,(trim(ra)))>0 then do;
bi="wee";
end;
run;

I would like to transorm the re variable to a list , so index could find the string if one of the values of the list 're' is found, here is an example 

data have;
re="all|you";  /* will find if word 'all' or 'you' is found the text 'texto' below , text is delimited with | */
ro="hello";
ra=catx('',ro,re);
texto='blablabla hello all, people';   /* <= instead of word all , the text can contain the word you*/
if index(texto,(trim(ra)))>0 then do;
bi="wee";
end;
run;

Thank you for your response.

[gamotte]

Hello,

data have;
re="\b(all|you)\b";  /* will find if word 'all' or 'you' is found the text 'texto' below , text is delimited with | */
ro="hello";
ra=catx('',ro,re);
texto='blablabla hello all, people';   /* <= instead of word all , the text can contain the word you*/
if prxmatch(cats("/",ra,"/"),texto) then do;
    bi="wee";
end;
run;

[SIgnificatif]

Thank you for the response, but I would like to find the answer using sas code and not re.
(* sorry, since i forgot to mention that 😞 )

[gamotte]

prxmatch is sas code but i understand you don't want to use regular expressions (why ?).

Here is a solution without regex. Note that it will match the string "hello young fellow" for instance.

data have;
re="all|you"; 
ro="hello";

texto='blablabla hello all, people';

do i=1 to countw(re,'|');
	re2=scan(re,i,'|');
	ra=catx('',ro,re2);

	if index(texto, trim(ra)) then do;
	    bi="wee";
	end;
end;
run;

[PGStats]

Why not Regular Expressions? They are your best bet for this task.

[SIgnificatif]

I think that they are more expensive in terms of cpu usage.... unfortunately I did not find any article that compares re to sas functions ...

[Ksharp · Accepted solution]

Using a array.

data have;
re="all|you";  
ro="hello";
ra=catx('',ro,re);
texto='blablabla hello all, people';   /* <= instead of word all , the text can contain the word you*/


do i=1 to countw(re,'|');
temp=scan(re,i,'|');
if index(texto,(trim(temp)))>0 then do;
bi="wee";leave;
end;
end;

run;


proc print;run;

[SIgnificatif]

Thank you Dear Ksharp and All for your responses 🙂 it's ok now