Regulgar Expression syntax help...

[Randell D.]

Folks,

I could do some basic regular expressions (which I picked up from sed,
in unix) but I'm afraid its been a while. I'm wondering if someone
could help out. I basically want to check that a string variable
contains only alphanumeric characters, in addition check for the
underscore. No decimal point check is required, and it should be a case
insensitive check.

I'm using PHP 4.x and would basically like to replace whatever is not
named above with an empty character (thus if a bang or question mark was
in a string, it would be removed from the string).

I know I could use something like preg_replace, but I don't know how to
say "when not found"... In sed, I could do something like

sed -e -i 's/[a-z]|[0-9]|_//g'

This would do the opposite of what I want to do (I think) but I'm lost
when it comes to PHP...

Can someone help me out, via the newsgroup (so all can learn?)

thanks in advance,
randell d.

[Tim Van Wassenhove]

On 2005-07-21, Randell D. <su*****@fiprojects.moc> wrote:

I'm using PHP 4.x and would basically like to replace whatever is not
named above with an empty character (thus if a bang or question mark was
in a string, it would be removed from the string).

I know I could use something like preg_replace, but I don't know how to
say "when not found"... In sed, I could do something like

sed -e -i 's/[a-z]|[0-9]|_//g'

preg_replace("#[^a-z0-9_]#", "", $subject);

--
Met vriendelijke groeten,
Tim Van Wassenhove <http://timvw.madoka.be>

[John Dunlop]

Tim Van Wassenhove wrote:

#[^a-z0-9_]#

and #2 of the 100 ways to write that expression - probably
*the* most common in c.l.php - involves + optimi[sz]ation
(likely not appreciable though):

#[^a-z0-9_]+#i <- case-insensitive, mind

in at no. 3, the less explicit:

#[^[:alnum:]_]+# //no need for 'i'; can match non-US-ASCIIs

down at number 4, the shorter and even more implicit:

#\W+# //as with 2, can match non-US-ASCIIs, no need for 'i'

they're all different in their own inimitable ways.

--
Jock

[Randell D.]

John Dunlop wrote:

Tim Van Wassenhove wrote:

#[^a-z0-9_]#

and #2 of the 100 ways to write that expression - probably
*the* most common in c.l.php - involves + optimi[sz]ation
(likely not appreciable though):

#[^a-z0-9_]+#i <- case-insensitive, mind

in at no. 3, the less explicit:

#[^[:alnum:]_]+# //no need for 'i'; can match non-US-ASCIIs

down at number 4, the shorter and even more implicit:

#\W+# //as with 2, can match non-US-ASCIIs, no need for 'i'

they're all different in their own inimitable ways.

Thanks to you both for the suggestions...

My internet access is limited at the moment but I will test this in a
few hours (when I have my LAMP machine)...

However... I'm curious, the solutions provided, when used with
preg_replace, would replace any of those characters, as opposed to
replace anything BUT those characters... (ie I want to replace anything
that is NOT alphanumeric or an underscore with a nospace...)

Thanks again for the help,
Randell D.

[John Dunlop]

Randell D. wrote:

However... I'm curious, the solutions provided, when used with
preg_replace, would replace any of those characters, as opposed to
replace anything BUT those characters... (ie I want to replace anything
that is NOT alphanumeric or an underscore with a nospace...)

That's what they do, for differing values of 'alpha'. The ^
negates the character class, so [^a-z] matches everything
except a, b, c, ... z. [[:alnum:]] in addition to the decimal
digits can cover other letters. \W matches any non-word
character (word = letters, numbers and underscore, where
'letters' depends on your locale settings).

--
Jock