str_replace won't replace arabic characters

Question

<?php 
$utf8_string = 'مع السلامة مع السلامة مع السلامة مع السلامة مع السلامة مع السلامة مع السلامة السلامة الرائعة على الطويلة ';
echo $utf8_string;
echo'<br/><br/>';

$patterns = array("على", "مع");
$replacements   = array("", "");

$r_string = str_replace($patterns, $replacements, $utf8_string);

//echo $r_string;
print_r ($r_string);
echo'<br/>';
//$words = preg_split( "/ ( |مع|على) /",$r_string);
$words = explode(" ",$r_string);

$num = count($words);
echo 'There are <strong>'.$num.'</strong> words.';
?>

I have this code to count the number of words in an arabic sentence.however i want to remove some words and count the rest.i tried to use str_replace, but this way is counting the number of words of the original sentence. can anyone help me?

Answer 1

You could use:

$num = count(
    explode(
        " ", 
        str_replace(
            $word, //Word you want to remove from your text.
            "",
            $string //String you want the word to be removed from.
        )
    )
);

Or even:

$num = count(
    explode(
        " ", 
        str_replace(
            array("word1", "word2", [...]), //Words you want to remove from your text.
            "",
            $string //String you want the word to be removed from.
        )
    )
);

---

EDIT: As pointed out, the above won't work. I tried pinpointing where the error is, and apparently str_replace can't handle arabic characters, even though explode can. PHP is not reliable with non-ascii characters.

What you can do, alternatively, is:

$num = Count(explode(" ", $utf8_string)) - Count(array_intersect(explode(" ", $utf8_string), $patterns))

It should return the value you want.

You could also try writing your own string replacement function, but I would advice against it, seeing you'd have to manually loop through your array and compare each word. Doing so should take longer to run, and make it much more verbose.

---

Coming here to warn yall that the correct way to handle this is with the mbstring extension (http://php.net/manual/en/book.mbstring.php). Please use this extension, no the ugly hack/workaround above.