topic Re: Regular expression to a url link in SAS Programming

Regular expression to a url link

melassiri — Wed, 18 Sep 2024 14:24:57 GMT

how can you use regular expression in sas to select the name of the website from a data and display it.

Examples:

A1 = google.com the result should be equal to: google

A2 = http://twitter.com/Marko_met_een_K/status/1725797169897021653 the result should be equal to: twitter

A3 = https://regioonline.nl/regio-den-bosch/schade-aan-stuw-lith/ then the result should be equal to: regioonline

A4 = https://www.aa.com/en/how-to-regex?id=123 the result should be equal to: aa

Re: Regular expression to a url link

Patrick — Wed, 18 Sep 2024 23:43:50 GMT

The following code generated by chatGPT using prompts: Using SAS code <copy/paste your question>

The chatGPT returned code required only one small fix to make it work.

data websites;
    input url :$100.;
    datalines;
google.com
http://twitter.com/Marko_met_een_K/status/1725797169897021653
https://regioonline.nl/regio-den-bosch/schade-aan-stuw-lith/
https://www.aa.com/en/how-to-regex?id=123
;
run;

data extracted_names;
    set websites;
    /* Use PRX to define a regex pattern to extract the website name */
    retain pattern;
    if _N_ = 1 then pattern = prxparse('/(?:https?:\/\/)?(?:www\.)?([^\/\.]+)\./');

    /* Apply the regex to the url and store the result in website_name */
    if prxmatch(pattern, url) then do;
        call prxsubstr(pattern, url, start_pos);
        website_name = prxposn(pattern, 1, url);
    end;

    /* Keep only the relevant columns */
    keep url website_name;
run;

proc print data=extracted_names noobs;
    title "Extracted Website Names";
run;

Re: Regular expression to a url link

Ksharp — Thu, 19 Sep 2024 01:18:40 GMT

Why you have to use PRX ? using classic sas function would be a lot easy.

data websites;
    input url :$100.;
    datalines;
google.com
http://twitter.com/Marko_met_een_K/status/1725797169897021653
https://regioonline.nl/regio-den-bosch/schade-aan-stuw-lith/
https://www.aa.com/en/how-to-regex?id=123
;
run;
data want;
 set websites;
temp=scan(substrn(url,find(url,'//')),1,'/');
if scan(temp,1,'.')='www' then want=scan(temp,2,'.');
 else want=scan(temp,1,'.');
run;

Re: Regular expression to a url link

melassiri — Sat, 21 Sep 2024 10:44:01 GMT

Thank you for your quick response i really appreciate it

Re: Regular expression to a url link

melassiri — Sat, 21 Sep 2024 10:53:52 GMT

thank you for your response.

That is very nice of you.