REGEX help needed: is this redundant ??

[mOrloff]

Could I just get rid of the first of these lines ?

PHP Code:

$text=preg_replace('/[^A-Za-z0-9-]/', ' ', $body); // strip all non-word, non-number, non-hyphen characters & punctuation
$text=preg_replace('/\b[^A-Za-z0-9-]{1,2}\b/', ' ', $text); // strip all words with fewer than three (min=1, max=2) characters 


Can you think of any instance where eliminating it would potentially miss something?

~ Mo

[MattF]

Not unless you want any string above three characters in length to be able to contain any character whatsoever.They're both parsing different strings too, btw.

[mOrloff]

Originally Posted by MattF

Not unless you want any string above three characters in length to be able to contain any character whatsoever...

I presume that's an answer to my first question about eliminating the 1st line, rather than my second question about potentially missing something ??

Originally Posted by MattF

... They're both parsing different strings too, btw.

Yeah, they're actually in sequence exactly like that.

It grabs a copy of body, then replaces any non-desired char's with a space, then replaces any string of <3 of those same non-desired chars with a space.

~ Mo

[MattF]

Originally Posted by mOrloff

I presume that's an answer to my first question about eliminating the 1st line, rather than my second question about potentially missing something ??

Yup. You need both of those lines if you want your text to be alphanumeric. If you don't mind allowing underscores also, you could just change that first line to:

$text=preg_replace('/[^\w\-]/', ' ', $body);

Yeah, they're actually in sequence exactly like that.

It grabs a copy of body, then replaces any non-desired char's with a space, then replaces any string of <3 of those same non-desired chars with a space.

I'd overlooked the $text assignment. Muppet mode there.

[mOrloff]

Thanks-a-bunch.