Replace Accented Characters

Question

I've been reading a few other questions but I am still stuck on the problem of converting strings containing accented characters into plain characters (by which I mean a-z)

I have a product name "Áhkká" which is already encoded as "Áhkká"

I want to decode this to the string with accents, and then convert it to read "Ahkka"

So far, I have tried:

function convert($name) {
   $name = html_entity_decode($name,ENT_COMPAT,"UTF-8");
   $name = iconv('UTF-8', 'ASCII//TRANSLIT', $name);
   return $name;
}

I get an error from iconv: "Detected an illegal character in input string"

I have also tried using htmlspecialchars_decode($name); but that gives me �hkk�

I also found a string replace function to clear accents, but I can't seem to pass a non-html string to it

$name = strtr($name,'àáâãäçèéêëìíîïñòóôõöùúûüýÿÀÁÂÃÄÇÈÉÊËÌÍÎÏÑÒÓÔÕÖÙÚÛÜÝ','aaaaaceeeeiiiinooooouuuuyyAAAAACEEEEIIIINOOOOOUUUUY');

Can someone please offer a solution? The server is running PHP 5.2.13. iconv is enabled glibc 2.5 (input/internal/output encoding is ISO-8859-1 in phpinfo)

`I have a product name "Áhkká" which is already encoded as "Áhkká"` <- why are you storing the html representation of the string in the first place? — Mahn, Oct 02 '12 at 19:54
Looks like your syntax error is from forgetting a semicolon: `return $name;` — Mr. Llama, Oct 02 '12 at 19:55
@Musa - I still get ?hkk?, Mahn, the data comes from a 3rd party (CSV data) so I have no control over this. GigaWatt - thanks, I updated my question, but this isn't the problem. — Alexander Holsgrove, Oct 02 '12 at 21:24
@Musa setlocale(LC_CTYPE, 'en_GB.utf8'); fixed it. Thanks :) — Alexander Holsgrove, Oct 03 '12 at 08:51

score 1 · Accepted Answer · edited May 23 '17 at 10:28

1

Trying to find a solution to your problem I have found this question:

multibyte strtr() -> mb_strtr()

In the chosen answer Alix Axel writes a function which is exactly what you need:

function Unaccent($string)
{
    return preg_replace('~&([a-z]{1,2})(?:acute|cedil|circ|grave|lig|orn|ring|slash|th|tilde|uml|caron);~i', '$1', htmlentities($string, ENT_QUOTES, 'UTF-8'));
}

echo Unaccent(html_entity_decode('&Aacute;hkk&aacute;'));

prints Ahkka

edited May 23 '17 at 10:28

Community

1
1

answered Oct 02 '12 at 21:59

Michal Trojanowski

10,641
2
22
41

This looks promising, but I end up with "ampaacutehkkampaacute" as the output. I isolated my code and that was the output. I've updated my answer with the test code based on your answer. Perhaps I need to set my default charset? – Alexander Holsgrove Oct 02 '12 at 22:13
Have you tried decoding html entities with the parameters you used before? `html_entity_decode($name,ENT_COMPAT,"UTF-8");` Do you get a proper UTF-8 string after decoding? If you do the `Unaccent` function should work well. – Michal Trojanowski Oct 03 '12 at 07:20
Sorry for my mistake. Your code did indeed work, but I also need to remove ' type entries. Secondly, This code will be run around 400,000 times along with 2 more preg_replace to make a "clean" url - to clean some product URLs - is this the fastest way? – Alexander Holsgrove Oct 03 '12 at 08:24
I've accepted your answer as it is a solution, although in the end I did use iconv (needed to setlocate); – Alexander Holsgrove Oct 03 '12 at 08:53

Replace Accented Characters

1 Answers1