PHP preg_replace doesn't act as expected with file names strings

Question

I'm trying to create a function which removes all none English characters (except spaces,dots and hyphens) from a string. For this I tried using preg_replace, but the function produces strange results.

I have a file called "example-נידדל.jpg"

Here is what I'm getting when trying to sanitize the file name:

echo preg_replace('/[^A-Za-z0-9\.]/','','example-נידדל.jpg');

The above produces: example.jpg as expected.

But when I try to pull the file name from a $_FILES array after uploading it to the server I get:

echo preg_replace('/[^A-Za-z0-9\.]/','',$_FILES['file_upload']["name"]);

The above produces example-15041497149114911500.jpg

The numbers I'm getting are in fact the HTML numbers of the characters which were suppose to be removed, see the following for character reference: http://realdev1.realise.com/rossa/phoneme/listCharactors.asp?start=1488&stop=1785&rows=297&page=1

I can't figure out why doesn't the preg_replace work with file names.

Can anyone help?

Thanks,

Roy

just echo $_FILES['file_upload']["name"] and see the results. — lovesh
– lovesh, Commented Jun 18, 2011 at 18:07
I've renamed an image to the name above and tried uploading it. Regardless of whether or not I specify an accept-charset for the form or add a charset meta tag, I always got it to return example.jpg. Are you sure the file you're uploading has isn't in fact named example-15041497149114911500.jpg before you upload it? — Francois Deschenes
– Francois Deschenes, Commented Jun 18, 2011 at 18:18
I checked now and the numbers I'm getting correlate to the respected HTML character that was suppose to be replaced: realdev1.realise.com/rossa/phoneme/… — Roy Peleg
– Roy Peleg, Commented Jun 18, 2011 at 20:35

Francois Deschenes · Accepted Answer · 2011-06-18 20:40:22Z

2

What about using mb_convert_encoding to convert the HTML entities back into UTF-8 before the preg_replace?

echo preg_replace('/[^A-Za-z0-9\.]/', '', mb_convert_encoding($_FILES['file_upload']["name"], 'UTF-8', 'HTML-ENTITIES'));

answered Jun 18, 2011 at 20:40

Francois Deschenes

25.1k4 gold badges67 silver badges61 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

Community · Accepted Answer · 2017-05-23 12:13:19Z

1

I would use a combination of regular expressions and iconv to transliterate it.

Update: Prior transliteration/filtering the filename mabye needs to be urldecoded:

$path = urldecode($path); // convert triplets to bytes.

Here is a code example from here that does something very similar to your question:

function pathauto_cleanstring($string)
{
    $url = $string;
    $url = preg_replace('~[^\\pL0-9_]+~u', '-', $url); // substitutes anything but letters, numbers and '_' with separator
    $url = trim($url, "-");
    $url = iconv("utf-8", "us-ascii//TRANSLIT", $url); // TRANSLIT does the whole job
    $url = strtolower($url);
    $url = preg_replace('~[^-a-z0-9_]+~', '', $url); // keep only letters, numbers, '_' and separator
    return $url;
}

It expects your into to be UTF-8 encoded.

Reference

edited May 23, 2017 at 12:13

CommunityBot

11 silver badge

answered Jun 18, 2011 at 20:20

hakre

200k55 gold badges454 silver badges865 bronze badges

2 Comments

Roy Peleg Over a year ago

Tried it, still the same result :-(

hakre Over a year ago

@Roy Peleg: I'm smelling that the filenname is urlencoded. So it needs to be urldecoded first. I'll add another piece of code, maybe it helps.

Collectives™ on Stack Overflow

PHP preg_replace doesn't act as expected with file names strings

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