How does WordPress generate URL slugs?

5 comments

1. As per @SinisterBeard‘s very valid comment to the question already a couple of years back now, this answer has long been outdated and the mentioned function(s) hence been replaced by a newer API:

   See wp_unique_post_slug.

   Original Answer

   Off the bat, I can’t give you a page/tutorial/documentation on how WP slugs are generated, but take a look at the sanitize_title() function.

   Don’t get a wrong impression by the function name, it is not meant to sanitize a title for further usage as a page/post title. It takes a title string and returns it to be used in a URL:

   • strips HTML & PHP
   • strips special chars
   • converts all characters to lowercaps
   • replaces whitespaces, underscores and periods by hyphens/dashes
   • reduces multiple consecutive dashes to one

   There might be edge cases where the core does something additional (you’d have to look at the source to verify that sanitize_title() will always suffice in generating exactly the same you expect), but this should cover at least 99%, if not all, cases.

   Log in to Reply

2. You can use this function:

   static public function slugify($text)
   {
     // replace non letter or digits by -
     $text = preg_replace('~[^pLd]+~u', '-', $text);
   
     // transliterate
     $text = iconv('utf-8', 'us-ascii//TRANSLIT', $text);
   
     // remove unwanted characters
     $text = preg_replace('~[^-w]+~', '', $text);
   
     // trim
     $text = trim($text, '-');
   
     // remove duplicate -
     $text = preg_replace('~-+~', '-', $text);
   
     // lowercase
     $text = strtolower($text);
   
     if (empty($text)) {
       return 'n-a';
     }
   
     return $text;
   }
   

   Its kind of exactly how the wp url sanitize function works.

   Log in to Reply

3. Core at your service

   There’s no developer mode built into WordPress aside from WP_DEBUG, which doesn’t help you too much in this case. Basically WP uses the “Rewrite API”, which is a function based, low level wrapper for the WP_Rewrite class, which you can read about in Codex. The global $wp_rewrite object stands at your service to inspect it or interact with the class.

   Plugins that help looking into it.

   Toschos “T5 Rewrite”-Plugin and Jan Fabrys “Monkeyman Rewrite Analyzer”-Plugin will guide you your way. I’ve written a small extension for “T5 Rewrite” to smoothly integrate it with the “Monkeyman Rewrite Analyzer”, which you can find in the “T5 Rewrite” repos wikie here on GitHub.

   The “Monkeyman”-plugin adds a new page, filed in the admin UI menu under Tools. The “T5 Rewrite”-plugin adds a new help tab to the Settings > Permalinks page. My extension adds the help tabs to the mentioned Tools-page too.

   Here’s a screenshot of what the “T5 Rewrite”-plugins help tab content looks like.

   

   ^{Vorlage = Pattern | Beschreibung = Explanation | Beispiele = Examples}

   Notes

   The “T5 Rewrite”-plugin does a wonderful job with helping you inspect the rewrite object. And it does even more: It adds new possibilities. Therefore it’s (at least in my installations) part of my basic plugins package.

   Log in to Reply

4. Actually, if you look core function wp_insert_post (post.php), you will see that it does the following:

   $data['post_name'] = wp_unique_post_slug( sanitize_title( $data['post_title'], $post_ID ), $post_ID, $data['post_status'], $post_type, $post_parent );
   
   $wpdb->update( $wpdb->posts, array( 'post_name' => $data['post_name'] ), $where );
   

   The key thing to note is that uses both wp_unique_post_slug and sanitize_title:

   wp_unique_post_slug( sanitize_title( 
   

   Log in to Reply

5. Forgive for resuming an old question, but I had the same necessity as found out this method works perfectly for me:

   $some_string = "DON'T STOP ME NOW!";
   $slug = sanitize_title(sanitize_title($some_string, '', 'save'), '', 'query');
   echo $slug; // dont-stop-me-now
   

   This method uses a double sanitization.

   The first one uses the save mode, where HTML and PHP tags are stripped, and accents are removed (accented characters are replaced with non-accented equivalents).

   The second query mode ensures all spaces are replaced with dashes - and other punctuation removed.

   Hope this helps someone! 🙂

   Log in to Reply