[RESOLVED] Retrieve URL array from $string

sssphp

Is there a way to retrieve an array of URL's from a text string? For example:
$string = 'This is a <a href="www.site.com">string</a> of text with 2 URL's, one of which is http://site2.com';

I want code that gives this array:

$url_array = array('www.site.com' , 'http://site2.com')

Shrike

Take a look at the man page for [MAN]parse_url[/MAN]

sssphp

Thanks, but I don't see how I can pull URL's out of a text string with that.

Any other ideas?

NogDog

<?php
$text = <<<EOD
This is <a href="www.site.com">a test</a>.
It is <a href='http://www.php.net/'>only a test</a>.
EOD;

preg_match_all('/href=([\'"])([^\1]+)\1/', $text, $matches);
$urls = $matches[2];
print_r($urls);
?>

sssphp

THANKS -- that's perfect. Some questions though:

1) (this one's probably dumb) -- what's 'EOD'; I got a Parse error: syntax error, unexpected T_SL on your exact code, but it worked when I made $text into a regular string (and escaped inside quotes)

2) Is there a way around the '2' in $matches[2]? The # of URL's will vary each time. I suppose I could use substr_count to get the no. and then plug it in if there's no easier way.

3) Can you come up with a reg ex based on 'http' and 'www' rather than 'href'? -- not all www's are in <a> tags, but all <a> tags have http or www (at least all full URL's, which is what I want)

NogDog

sssphp wrote:
THANKS -- that's perfect. Some questions though:

1) (this one's probably dumb) -- what's 'EOD'; I got a Parse error: syntax error, unexpected T_SL on your exact code, but it worked when I made $text into a regular string (and escaped inside quotes)

The "EOD" is just an arbitrary string. It must match the string after the "<<<" at the start of the string assignment. The "EOD;" must appear on a line by itself with no leading or trailing space of any kind. This is refered to as "heredoc" string delimiting. See http://www.php.net/manual/en/language.types.string.php#language.types.string.syntax.heredoc for more info.

sssphp wrote:
2) Is there a way around the '2' in $matches[2]? The # of URL's will vary each time. I suppose I could use substr_count to get the no. and then plug it in if there's no easier way.

In this case: no. The "2" index indicates that we're looking for all matches found on the second sub-pattern of the regex - it has nothing to do with how many matches are found (there will be one element in this sub-array for each match).

sssphp wrote:
3) Can you come up with a reg ex based on 'http' and 'www' rather than 'href'? -- not all www's are in <a> tags, but all <a> tags have http or www (at least all full URL's, which is what I want)

This is actually a bit easier:

preg_match_all('/(http:\/\/|www\.)[^\s\'">]+/', $text, $matches);
$urls = $matches[0];

sssphp

Thanks VERY much -- that's exactly what I need. 🙂