How to use str_replace() to remove text a certain number of times only in PHP?

I am trying to remove the word "John" a certain number of times from a string. I read on the php manual that str_replace excepts a 4th parameter called "count". So I figured that can be used to specify how many instances of the search should be removed. But that doesn't seem to be the case since the following:

$string = 'Hello John, how are you John. John are you happy with your life John?';

$numberOfInstances = 2;

echo str_replace('John', 'dude', $string, $numberOfInstances);

replaces all instances of the word "John" with "dude" instead of doing it just twice and leaving the other two Johns alone.

For my purposes it doesn't matter which order the replacement happens in, for example the first 2 instances can be replaced, or the last two or a combination, the order of the replacement doesn't matter开发者_StackOverflow社区.

So is there a way to use str_replace() in this way or is there another built in (non-regex) function that can achieve what I'm looking for?

As Artelius explains, the last parameter to str_replace() is set by the function. There's no parameter that allows you to limit the number of replacements.

Only preg_replace() features such a parameter:

echo preg_replace('/John/', 'dude', $string, $numberOfInstances);

That is as simple as it gets, and I suggest using it because its performance hit is way too tiny compared to the tedium of the following non-regex solution:

$len = strlen('John');

while ($numberOfInstances-- > 0 && ($pos = strpos($string, 'John')) !== false)
    $string = substr_replace($string, 'dude', $pos, $len);

echo $string;

You can choose either solution though, both work as you intend.

You've misunderstood the wording of the manual.

If passed, this will be set to the number of replacements performed.

The parameter is passed by reference and its value is changed by the function to indicate how many times the string was found and replaced. Its initial value is discarded.

There are a few things you could do to achieve this, but I can't think of one specific php function that will easily let you do this.

One option is to create your own replace function and utilize strripos and substr to do the replaces.

Another thing you can do is use preg_replace_callback and count the number of replacements you have done in the callback.

There's probably more ways but that's all I can think of on the fly. If performance is an issue I suggest you give both a try and do some simple benchmarks.

The cleanest, most-direct, single function call is to use preg_replace(). Its replacement limiting parameter makes the task intuitive and readable.

$string = preg_replace('/John/', 'dude', $string, $numberOfInstances);

The function is also attractive because making the search case-insensitive is as simple as adding the i pattern modifier to the end of the pattern. I won't delve into the usefulness of word boundaries (\b).

If a search string might contain characters with special meaning to the regex engine, then preg_quote() will be necessary -- this diminishes the beauty of the technique but not prohibitively so.

$search = '$5.99';
$pattern = '/' . preg_quote($search, '/') . '/';
$string = preg_replace($pattern, 'free', $string, $numberOfInstances);

For anyone who has an unnatural bias against regex functions, this can be done without regex and without looping -- it will be case-sensitive though.

Limited Explode & Implode: (Demo)

$numberOfInstances = 2;

$string = 'Hello John, how are you John. John are you happy with your life John?';
// explode here -^^^^ and ---------^^^^ only to create the following array:
// 0 => 'Hello ',
// 1 => ', how are you ',
// 2 => '. John are you happy with your life John?'

echo implode('dude', explode('John', $string, $numberOfInstances + 1));

Output:

Hello dude, how are you dude. John are you happy with your life John?

Notice the explode's limiting parameter dictates how many elements are generated, not how many explosions are executed on the string.

function str_replace_occurrences($find, $replace, $string, $count = -1) {
    // current occrurence 
    $current = 0;
    // while any occurrence 
    while (($pos = strpos($string, $find)) != false) {
        // update length of str (size of string is changing)
        $len = strlen($find);
        // found next one 
        $current++;
        // check if we've reached our target 
        // -1 is used to replace all occurrence 
        if($current <= $count || $count == -1) {
            // do replacement 
            $string = substr_replace($string, $replace, $pos, $len);
        } else {
            // we've reached our 
            break;
        }
    }
    return $string;
}

Artelius has already described how the function works, ill just show you how to do this via the manual methods:

function str_replace_occurrences($find,$replace,$string,$count = 0)
{
    if($count == 0)
    {
        return str_replace($find,$replace,$string);
    }

    $pos = 0;
    $len = strlen($find);
    while($pos < $count && false !== ($pos = strpos($string,$find,$pos)))
    {
        $string = substr_replace($string,$replace,$pos,$len);
    }

    return $string;
}

This is untested but should work.