Reg Ex Help!

whaites

I'm failing miserably to come up with what should be an easy reg ex. I need to find any line that starts with a "<p" with a "!supportEmptyParas" in the middle and ends with a "/p>" with any characters in between like:

<![if !supportEmptyParas]> <![endif]><o:p></o:p>

Why doesn't something as simple as this work?:
'^{<p!supportEmptyParas\/p>$'i}

I'm new to reg ex, any help is very appreciated.

Thanks,
Tom

gid

$t = "<p class=MsoNormal><span style='font-size:12.0pt'><b><![if !supportEmptyParas]> <![endif]><o:p></o:p></b></span></p>";

if(preg_match('/^<p.*?!supportEmptyParas.*?\/p>$/i', $t)) {
    echo 'matched';
}

Note the ? after . this makes it non-greedy. . alone was matching the whole rest of the line, so the !support... was not being matched, because .* already matched it.

whaites

I tried this, but with no luck:

$text = preg_replace("/^{<p.?!supportEmptyParas.?\/p>$/i",} "MATCH", $text);

Didn't work on a block of text like this:

<![if !supportEmptyParas]> <![endif]><o:p></o:p>

which was embedded multiple times in a block of html.

I'm trying to delete these lines entirely from a block of html.

Could I have issues with the new lines?

Thanks,

Tom

examancer

Originally posted by whaites
I tried this, but with no luck:

$text = preg_replace("/^{<p.?!supportEmptyParas.?\/p>$/i",} "MATCH", $text);

Didn't work on a block of text like this:

<![if !supportEmptyParas]> <![endif]><o:p></o:p>

which was embedded multiple times in a block of html.

I'm trying to delete these lines entirely from a block of html.

Could I have issues with the new lines?

Thanks,

Tom

you've got the syntax all wrong, here is a better example:

// $text contains the string you want to search. The string must be stored in a variable.
// If you want to scan an html file, you must load it into a string using a file function, like fread()
$text = "<p class=MsoNormal><span style='font-size:12.0pt'><b><![if !supportEmptyParas]> <![endif]><o:p></o:p></b></span></p>";

// $num_matches contains the number of times the pattern was matched
// $matches is an array that contains each individual match.
$num_matches = preg_match('/^<p.*?!supportEmptyParas.*?\/p>$/i', $text, $matches));

whaites

My example was not clear.
Let's say I have html in a variable like so:

$text="<![if !supportEmptyParas]> <![endif]><o:p></o:p>
<![if !supportEmptyParas]> <![endif]><o:p></o:p>
PURPOSE: 
Acting as a liaison between the sections and the MBA leadership and staff, the...";

I can't get this pattern to match the   lines. I want to use preg_replace to remove those lines entirely from the block of text like this:
$text = preg_replace("/^{<p.?!supportEmptyParas.?\/p>$/i",} "", $text);

Will this expression: /^{<p.?!supportEmptyParas.?/p>$/i}
only work with preg_match? What is the syntax for the search expression for preg_replace (I assumed it would be the same).

Thanks,
Tom

Weedpacket

The problem is probably the and $ characters you have at the start and end of the regexp. meaning "start of the string" and $ meaning "end of the string", so the regexp will only match if the entire string - from start to end - is of the form you're looking for. Since the bits you're lookinf for don't make up the entire string, they won't match.

Try without those and see if you get any improvement.

Additional: the regexp will only work if the entire paragraph from is on a single line with no newline characters, because . doesn't match a newline character unless you ask for it with the 's' modifier. But then you'll get a new problem with the regexp starting its match way too soon (it always tries to match as soon as it can - in this case, the very first <p it sees).

/<p((?!\/p>).)*!supportEmptyParas.*?\/p>/is

Takes care of this by making sure that the bit it matches between - which would presumably always appear somewhere between one <p and the next.

whaites

THANKS! This works great.

On the topic, is there well recommended site/book for regular expressions?

examancer

i learned from microsoft's msdn tutorial on regular expressions, but I can't bring that site up any, so this is where i go now:

http://www.pcre.org/man.txt

its written for C++ i think, but the PCRE (perl compatible regular expressions) examples and explanations are still useful, so I think thats what you want.

Weedpacket

Originally posted by whaites
THANKS! This works great.

On the topic, is there well recommended site/book for regular expressions?

Book is easy: O'Reilly & Associate's Mastering Regular Expressions. As for sites - there's a tutorial on this site but it's for the ereg-style regexps (which you wouldn't want to use in the multiline situation I noted). Perl programming sites may have tutorials on the subject of regular expressions also, particularly since they're more likely to discuss the Perl-style pattern modifers and assertions (like (?!...))