remove non-alphanumeric characters except underscore

[Nodral]

Hi

 

I need to strip all non alphanumeric characters from a string except an underscore.

 

Here's what I have so far which works from the non alphanumeric bit, how do I then ignore underscores?

 

$value = ereg_replace("[^A-Za-z0-9]", "", $value );

 

Any help would be greatly appreciated.

[AyKay47]

Use preg_replace, erg_replace is deprecated. \w includes alpha-numerical characters and an underscore.

 

$value = preg_replace('\W*', '', $value);

[Nodral]

Ok, done that and it removes all characters.  I think i'm getting a certain amount of 'garbage' in my strings, due to them being imported from a tab seperated csv file.  Any thoughts how I can 'clean' the input, or check it is set to the correct encoding?

[AyKay47]

Can you give me an example?

Also note that I forgot the delimiters in my above reply.

[Nodral]

Delimiters?

 

Sorry, not really used regex before, finding it's a whole different ball game!!!

[Nodral]

Ok, found what you mean.

 

Now I have

$value = preg_replace('/\W*/', '', $value);

 

But I still get

 

ÿþcourse_id

 

How can I strip the foreign characters off?

[AyKay47]

There are several ways, I would first use iconv to remove undesired encoded characters:

 

$value = iconv("UTF-8", "ISO-8859-1//IGNORE", $value);
$value = preg_replace('~\W*~', '', $value);
echo $value;