still regular expression

seanlyon

Cheers,

I want to thank those who replied to my previous questions. As I play with regular expression more, one thing comes up:

How do we find something NOT including a pattern? i.e.

<a href = " http://www.yahoo.com" >.....
<a href= investment.html >.....
<a href = "/books/search.asp?isbn=12345678" >.....
......

I want to parse a html page with all links similar to the above and get the values of href. So I don't want anything after the '='. Note that the right side values come in different format, some quoted, some don't, some start with http and some don't, some with space and some don't.

I was using explode but it didn't work for the third case. Any simple way to extract values here? Can I use preg_match in one step?

Thanks!

seanlyon

I sort of figuring it out by using preg_split.

$results = preg_split("/\s+href\s=\s/i", $inputs);

The right side value should be one of $results.

pyro

Yes, it can be done with regex:

<?PHP
$str = "blah <a href=\"http://www.yahoo.com\"> blah <a href= investment.html > blah <a href = \"/books/search.asp?isbn=12345678\" >";
preg_match_all("/<a +href *= *\"? *(.*?) *\"? *>/i", $str, $matches); #be sure to include the spaces
$links = $matches[1]; #$links is an array of all the links.
?>

jpr

This seems to work really good. However when something extra is in the link, such as target="new", it does funny things. Any way to get around this?

$str = "blah <a href=\"http://www.yahoo.com\" target=\"new\"> blah <a href= investment.html > blah <a href = \"/books/search.asp?isbn=12345678\" >";
preg_match_all("/<a +href = \"? (.?) \"? >/i", $str, $matches); #be sure to include the spaces
$links = $matches[1]; #$links is an array of all the links.

for ($i=0; $i<count($links); $i++) {
echo htmlentities($links[$i]) . "<br>\n";
}

That printed this:

http://www.yahoo.com" target="new
investment.html
/books/search.asp?isbn=12345678

pyro

This should work, as long as your link attributes are quoted (as they should be):

$str = "blah <a href=\"http://www.yahoo.com\" target=\"new\"> blah <a href = \"/books/search.asp?isbn=12345678\" >"; 
preg_match_all("/<a +href *= *[\"'] *(.*?) *[\"'] *[^>]*/i", $str, $matches); #be sure to include the spaces
$links = $matches[1]; #$links is an array of all the links.

for ($i=0; $i<count($links); $i++) {
	echo htmlentities($links[$i]) . "<br>\n";
}

pyro

Ok, a bit more playing and now you don't have to have the links quoted (though it's a good idea to do so):

$str = "blah <a href=\"http://www.yahoo.com\" target=\"new\"> blah <a href= investment.html > blah <a href = \"/books/search.asp?isbn=12345678\" >"; 
preg_match_all("/<a +href *= *[\"']? *([^\"' >]*)/i", $str, $matches); #be sure to include the spaces
$links = $matches[1]; #$links is an array of all the links.

for ($i=0; $i<count($links); $i++) {
	echo htmlentities($links[$i]) . "<br>\n";
}

seanlyon

Cheers, Pyro and JPR:

Thanks for your kind replies. My purpose is to extract the link value of href so I can prepend my dispatcher to create pages with my look n feel (without using frames). Each time user clicks on a link or submits a form, a new page is created the same way.

So far, I have been able to replace most img src, href, form action etc. with a few broken links due to the fact that original source uses relative links without giving enough path info. Not sure how to deal with that.

🙂

pyro

Glad it worked for you... 🙂