TalkPHP - Using regex

TalkPHP (http://www.talkphp.com/forums.php)

- General (http://www.talkphp.com/general/)

- - Using regex (http://www.talkphp.com/general/4758-using-regex.html)

benton

07-20-2009 01:56 AM

Using regex

The following code, the way I understand it, is supposed to strip all punctuation characters from the string. From what I've found on the web, a space is not one of those characters. But the code below outputs "abstractmodern." Would someone please explain how I can strip all special characters except the spaces?

PHP Code:


		
			
$pattern = "([[:punct:]])"; 

$str = "abstract modern"; 

$str = ereg_replace($pattern, '', $str)); 

echo $str;

Village Idiot

07-20-2009 03:25 AM

Use this range [!-@]

This is because ranges go by ASCII value, where ! is the first punctuation (after space) and @ is the last.

EGYG33K

07-20-2009 03:32 AM

PHP Code:


		
			
echo ereg_replace("[^[:alpha:],[:digit:],_,[:blank:]]","",'this is 123 $ *&^_ text test');

Enfernikus

07-20-2009 04:38 AM

As a side note, it's best just to stop using the ereg_* function and it's POCL ( I think it is ) engine because as of 5.3.0 it's going to through an E_DEPRECIATED error. Use the preg_* family

benton

07-20-2009 03:14 PM

Thanks for the suggestions but they don't seem to be working. I changed the input string to more properly illustrate the problem. So given this string

PHP Code:


		
			
$str = '#!$abstract modern-fields';

it should output

Quote:

abstract modern-fields

If I use this

PHP Code:


		
			
echo ereg_replace("[^[:alpha:],[:digit:],_,[:blank:]]","",$str).'<br>';

It displays

Quote:

abstract modernfields

And if I use this

PHP Code:


		
			
ereg_replace("[[!-@],[:punct:]]","",$str).'<br>';

It displays

Quote:

#!$abstract modern-fields

The #!S are all listed as characters in :punct: so those should be stripped. The first method does that but it also removes the hyphen. The second doesn't remove the punctuation but handles the hyphen correctly. Would someone point out my mistake, please?

Salathe

07-20-2009 04:27 PM

Do you want to use a whitelist or a blacklist. The former would remove all characters except a specific set, e.g. allow only alphanumeric and hyphen characters. The latter, a blacklist, would remove only specific characters, e.g. !@#$%^&*().

It is also advised to use the PCRE functions (preg_replace) since the POSIX family of functions (ereg_*, split, etc.) is deprecated.

With regards to why ereg_replace("[[!-@],[:punct:]]","",$str) won't work, it's just how you've constructed the character set. There shouldn't be a set of square brackets around !-@. Note that that range includes the dash/hyphen character - so that will be removed.

Back to white- and blacklists, here are a few examples:
Only allow alphanumeric (case insensitive) and hyphen characters

PHP Code:


		
			
echo preg_replace('/[^a-z0-9-]/i', '', $str);

Only remove !@#$%^&*()+= characters

PHP Code:


		
			
echo preg_replace('/[!@#$%^&*()+=]/', '', $str);

EGYG33K

07-20-2009 07:52 PM

PHP Code:


		
			
echo ereg_replace("[^[:alpha:],[:digit:],_,-,[:blank:]]","",$str).'<br>';

benton

07-23-2009 05:48 PM

My thanks to everyone. The code is working as expected now and I have a better understanding of how to manipulate the string. I do appreciate it.

All times are GMT. The time now is 01:18 PM.