PHP filename encoding conversion issue

Question

I'm trying to batch rename files in a folder with PHP. It's mostly working, though I'm having problems with accented characters.

An example of a filename with accented characters is ÅRE_GRÖN.JPG.

I would like to rename that file to ARE_GRON.JPG.

If I read the files in like this:

<?php
$path = __DIR__;
$dir_handle = opendir($path);

while ($file = readdir($dir_handle)) {
    echo $file . "\n";
}

closedir($dir_handle);

...And the page displays AÌŠRE_GROÌˆN.JPG.

If I add header('Content-Type: text/html; charset=UTF-8'); to the beginning of my script, it displays the correct file name, but the rename() function seems to have no effect either way.

Here's what I've tried:

while ($file = readdir($dir_handle)) {
    rename($file, str_replace('OÌˆ', 'O', $file)); # No effect
    rename($file, str_replace('Ö', 'O', $file)); # No effect
}

Where am I going wrong?

Do say if you believe I'm using the wrong tool for the job. If anyone knows how to achieve this with a Bash script, show me. I have no Bash chops.

Since he said bash, I would guess he is referring to bash(1) which would suggest Lunix. — Christian Kiewiet
– Christian Kiewiet, Commented Feb 12, 2013 at 21:56
This is what I could find: bugs.php.net/bug.php?id=39660 However I believe there should already be a work-around for this, like using the encoding system PHP is okay with. I'll post an answer if I ever find anything. Also possible duplicate to: stackoverflow.com/questions/873853/… — ASertacAkkaya
– ASertacAkkaya, Commented Feb 12, 2013 at 21:57
It could easily be Bash on cygwin, natively on Windows or on FreeBSD. — that other guy
– that other guy, Commented Feb 12, 2013 at 21:57

Jezen Thomas · Accepted Answer · 2013-02-13 01:49:22Z

I figured out how to do it.

I first ran urlencode() on the filename. This converts the string:

MÖRKGRÅ.JPG

To the URL friendly:

MO%CC%88RKGRA%CC%8A.JPG

I then ran str_replace() on the URL-encoded string, providing needles and haystacks in arrays. I only needed it for a few Swedish characters, so my solution looked like this:

<?php

header('Content-Type: text/html; charset=UTF-8');

$path = __DIR__;

$dir_handle = opendir($path);

while ($file = readdir($dir_handle)) {
    $search = array('A%CC%8A', 'A%CC%88', 'O%CC%88');
    $replace = array('A', 'A', 'O');
    rename($file, str_replace($search, $replace, urlencode($file)));
}

closedir($dir_handle);

Job done :)

I've come to realise this is more versatile than I anticipated. Running another script, url_encode() gave me some slightly different output, but it's easy to change accordingly.

$search = array('%26Aring%3B', '%26Auml%3B', '%26Ouml%3B', '+');
$replace = array('A', 'A', 'O', '_');

that other guy · Accepted Answer · 2013-02-12 22:18:15Z

0

If you have a limited number of characters you want to replace, you can do it with

for f in *; do mv "$f" "${f//Ö/O/}" 2> /dev/null; done

On GNU you could more generally use

expr=""
for char in {A..Z}
do 
    expr+="s/[[=$char=]]/$char/g; "; 
done; 

for f in *; do 
    mv "$f" "$(sed -e "$expr" <<< "$f")" 2> /dev/null; 
done

to replace all A-like accented characters with an ascii A, for every character in the alphabet, but with no guarantees for OS X sed. Beware that this has the side effect of capitalizing all filenames.

answered Feb 12, 2013 at 22:18

that other guy

125k12 gold badges187 silver badges214 bronze badges

2 Comments

Jezen Thomas Over a year ago

Hmm.. I tried running that first script from the directory that holds the files, but it didn't seem to have any effect.

that other guy Over a year ago

Try copy-pasting the Ö character from the filename rather than typing it. Unicode has a lot of pretty identical Ö characters.

Collectives™ on Stack Overflow

PHP filename encoding conversion issue

2 Answers 2

Comments

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related