2
\$\begingroup\$

I need to check the end of a URL for the possible existence of /news_archive or /news_archive/5 in PHP. The below snippet does exactly what I want, but I know that I could achieve this with one preg_match rather than two. How can I improve this code to treat the /5 as an optional segment and capture it if it exists?

if (preg_match('~/[0-9A-Za-z_-]+_archive/[0-9]+$~', $_SERVER['HTTP_REFERER'], $matches) || preg_match('~/[0-9A-Za-z_-]+_archive$~', $_SERVER['HTTP_REFERER'], $matches)) {
    $page_info['parent_page']['page_label'] = ltrim($matches[0], '/');
}
\$\endgroup\$
2
\$\begingroup\$

Consider your first pattern:

~/[0-9A-Za-z_-]+_archive/[0-9]+$~

Let's break it down:

  1. / a literal string /
  2. [0-9A-Za-z_-]+ one or more of 0-9, A-Z, a-z, _ or -
  3. _archive a literal string _archive
  4. / literal slash again
  5. [0-9]+ one or more digits
  6. $ the end of the string must follow the one or more digits

So basically you want to make #4 and #5 optional. To be more specific, you want either both 4 and 5, or neither 4 nor 5.

Consider this:

(a[b]+)?

This means that you have one a followed by one or more b, and that this grouped a/b entity is optional.

Letting a be #4 and b be digits like in #5, we're left with:

(/[0-9]+)?

Or:

~/[0-9A-Za-z_-]+_archive(/[0-9]+)?$~

This will capture the entire group though, like /5:

php -r "preg_match('~/[0-9A-Za-z_-]+_archive(/([0-9]+))?$~', '/news_archive/5', $m); var_dump($m);"
array(2) {
  [0] =>
  string(15) "/news_archive/5"
  [1] =>
  string(2) "/5"
}

You can just add another group to remedy that though:

~/[0-9A-Za-z_-]+_archive(/([0-9]+))?$~

Example:

php -r "preg_match('~/[0-9A-Za-z_-]+_archive(/([0-9]+))?$~', '/news_archive/44', $m); var_dump($m);"
array(3) {
  [0] =>
  string(16) "/news_archive/44"
  [1] =>
  string(3) "/44"
  [2] =>
  string(2) "44"
}

You could technically make the outside group a non-capturing group (like (?:/([0-9]+))?), but I don't think the added complication is worth not grabbing the / part too.

(By the way, sorry if you're familiar with regex and you found this excessive. I tend to take a very verbose approach to any regex related question :).)

\$\endgroup\$
  • \$\begingroup\$ This is a fantastic response, and certainly more than I expected. In a good way. I am fairly unfamiliar with regex itself, so the thorough analysis was a pleasant and refreshing lesson! Thank you. \$\endgroup\$ – davo0105 Oct 5 '12 at 22:43
  • \$\begingroup\$ @davo0105 Glad I could help! :) \$\endgroup\$ – Corbin Oct 6 '12 at 3:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.