Regex in PHP: how do I replace hyphens between alphanumeric characters?

Question

I'm trying to find and then replace instances of hyphens with en and em widths.

So, in the example: "10-100" the hyphen would be replaced by an en width. Also, in the example: "It is - without doubt - the worst" or: "It is -- without doubt -- the worst" either instances would be replaced by an em width.

However, I just can't figure out the proper pattern for preg_replace() in PHP.

"/[0-9]+(\-)[0-9]+/"

... appears to do the replace, but removes the numbers.

How do I get preg_replace() to ignore the patterns either side of the subject?

I suppose it goes without saying that Regular Expressions are weird, but I got it working, so thanks! — Wayne Smallman
– Wayne Smallman, Commented May 3, 2014 at 16:30

Sverri M. Olsen · Accepted Answer · 2014-05-05 01:32:58Z

1

You can use lookbehinds and lookaheads:

function prettyDashes($string) {
    static $regex = array(
        '/(?<=\d)-(?=\d)/' => '&ndash;',  // EN-dash
        '/(?<=\s)-(?=\s)/' => '&mdash;',  // EM-dash
        '/(?<=\w)--(?=\w)/' => '&mdash;', // EM-dash
    );
    return preg_replace(array_keys($regex), array_values($regex), $string);
}
$tests = array(
    'There are 10-20 dogs in the kennel.',
    'My day was - without a doubt - the worst!',
    'My day was--without a doubt--the worst!',
);
foreach ($tests as $test) {
    echo prettyDashes($test), '<br>';
}

The problem is that it is difficult to detect and avoid false-positives when replacing stuff like this. Normal hyphenated words, like "to-do", are not tangential (em-dash), and dates, like 18-12-2014, are not ranges (en-dash). You have to be quite conservative in what you replace, and you should not be surprised if something is changed erroneously.

edited May 5, 2014 at 1:32

answered May 3, 2014 at 19:29

Sverri M. Olsen

13.2k3 gold badges38 silver badges54 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Wayne Smallman Over a year ago

"The problem is that it is difficult to detect and avoid false-positives when replacing stuff like this." Which is exactly the problem I'm having at present, such as detection (and ignore) within a URL, as an example.

Sverri M. Olsen Over a year ago

@WayneSmallman Yeah, that is just a problem that comes with doing this. You have to know every false-positive (or at least know how to identify them somehow) in order to avoid replacing correctly used hyphens. Implementing it would also be quite nontrivial.

Wayne Smallman · Accepted Answer · 2014-05-03 16:35:47Z

0

So, with thanks to @mario, the answers are:

"/(?=.*?[0-9])(\-)(?=.*?[0-9])/"

"/(?=.*?\w)( \- )(?=.*?\w)/"

"/(?=.*?\w)( \-- )(?=.*?\w)/"

answered May 3, 2014 at 16:35

Wayne Smallman

1,72012 gold badges35 silver badges60 bronze badges

Collectives™ on Stack Overflow

Regex in PHP: how do I replace hyphens between alphanumeric characters?

2 Answers 2

2 Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related