I recently needed a routine that would remove the characters in one string from another, like the regex
<?php
$result = preg_replace("/[$chars]/", "", $string);
?>
and I needed it to be fast, and accept pretty much all input. The regex above won't work when strlen($chars) == 0. I came up with this, admittedly pretty horrible-looking code, that is quite fast:
<?php
function RemoveChars($string, $chars)
{
return isset($chars{0}) ? str_replace($chars{0}, "", strtr($string, $chars, str_pad($chars{0}, strlen($chars), $chars{0}))) : $string;
}
?>
According to my own measurements, the regex in ONLY faster for when strlen($chars) == 1; for longer strings, my routine is faster. What does it do? Let's say you want to remove the period, the comma and the exclamation mark from a string, like so:
$result = RemoveChars("Isn't this, like, totally neat..!?", ".?!");
The str_pad function creates a string equal in length to the string that contains the character to be removed, but consisting only of the first character of that string:
The input is ".,!"
The output is "..."
The strtr function translates all characters in the string-to-be-processed ("Isn't this...") that also occur in the input (".,!") to the characters in the same position in the output ("..."). In other words:
Isn't this, like, totally neat..!?
becomes
Isn't this. like. totally neat....
Finally, the first character from the input (".,!") which happens to be, again, the period, is removed from that string by the str_replace call:
Isn't this like totally neat?
The function needs to check is $chars has at least one character, or else the str_pad function will fail. If it's empty, then the unprocessed string is returned.
