Find the position of the first occurrence of any number in string

Question

Can someone help me with algorithm for finding the position of the first occurrence of any number in a string?

The code I found on the web does not work:

function my_offset($text){
    preg_match('/^[^\-]*-\D*/', $text, $m);
    return strlen($m[0]);
}
echo my_offset('[HorribleSubs] Bleach - 311 [720p].mkv');

Solution 1

function my_ofset($text){
    preg_match('/^\D*(?=\d)/', $text, $m);
    return isset($m[0]) ? strlen($m[0]) : false;
}

should work for this. The original code required a - to come before the first number, perhaps that was the problem?

Solution 2

function my_offset($text) {
    preg_match('/\d/', $text, $m, PREG_OFFSET_CAPTURE);
    if (sizeof($m))
        return $m[0][1]; // 24 in your example

    // return anything you need for the case when there's no numbers in the string
    return strlen($text);
}

Solution 3

The built-in PHP function strcspn() will do the same as the function in Stanislav Shabalin's answer when used like so:

strcspn( $str , '0123456789' )

Examples:

echo strcspn( 'That will be $2.95 with a coupon.' , '0123456789' ); // 14
echo strcspn( '12 people said yes'                , '0123456789' ); // 0
echo strcspn( 'You are number one!'               , '0123456789' ); // 19

HTH

Comments

[Tim Pietzcker]

Makes more sense. I (and RegexBuddy) didn't know about PREG_OFFSET_CAPTURE :)

[Gasper]

if there is no number in string it returns length of original string?

[Tim Pietzcker]

@gašper: Right. Can be fixed, but Stanislav's solution is much better.

[Gasper]

Why is his solution better? if there is no number in string it returns me 1 (with his algorithm). if i use yours it returns me something logical (string length).

[Stanislav Shabalin]

@gašper I've edited mine. If there's no number it now returns string length as well. I guess it's better because a) it searches exactly for numbers, not for non-numbers and b) it utilizes preg_match built-in feature which is always better than reinventing a bike. I hope Tim agrees with me :-)

[Tim Pietzcker]

@gašper: Well, my solution is kind of bent-over backwards (count the number of non-digits until the first digit). Also, thinking about it, I expect my solution to require an additional check whether there has been a match at all. I'm not a PHP programmer, so I'm building that code upon snippets created by RegexBuddy using my regex.

[Roland Seuhs]

I bet that this is about 20 times faster than any regex solution

[djdance]

note: it works fine with non-latin text, but then don't use mb_substr