I need to get the email adress from an anchor with a mailto attribute with regex.
this pattern: (.*)<a\s(.*?)(.*)\s*href\=['"]mailto:([-a-z0-9_]+)@([a-z0-9-]+).([a-z]+)['"]>(.*)</a>(.*)
Works in regex coach though it doesnt work with PHP.
Code:
preg_match("'(.*)<a (.*?)(.*) *href\=['\"]mailto:([-a-z0-9_]+)@([a-z0-9-]+).([a-z]+)['\"]>(.*)</a>(.*)'si", "<a href=\"mailto:someemail@ohio.com\"">Some email</a>", $matches);
print_r($matches);
So why doenst it work in php?
PHP’s PCRE require the regular expression to be wrapped into delimiters that separate the pattern from optional modifiers. In this case the first non-alphanumeric character is used (i.e. '
) so the pattern is actually just (.*)<a (.*?)(.*) *href\=[
and the rest are treated as modifiers. And that is an invalid regular expression as the [
is not properly escaped and the rest are not valid modifiers neither.
As the others have already suggested, you can fix this by escaping any occurrence of the delimiter '
inside the regular expression or choose a different delimiter that does not appear in the regular expression.
But besides that, trying to parse HTML with regular expressions is very error prone. In you case using that many .*
will also result in a horrible performance behavior (it’s just due to how regular expressions are processed).
Better use a proper HTML parser that returns a DOM that can be queried like PHP’s DOM library:
$doc = new DomDocument();
$doc->loadHTML($str);
foreach ($doc->getElementsByTagName("a") as $a) {
if ($a->hasAttribute("href")) {
$href = trim($a->getAttribute("href"));
if (strtolower(substr($href, 0, 7)) === 'mailto:') {
$components = parse_url($href);
}
}
}
Your delimiter is a quote '
, and there are some instances of it in the regex:
preg_match("'(.*)<a (.*?)(.*) *href\=['\"]mailto:([-a-z0-9_]+)@([a-z0-9-]+).([a-z]+)['\"]>(.*)</a>(.*)'si", "<a href=\"mailto:someemail@ohio.com\"">Some email</a>", $matches);
^ ^
Escape them (i.e.: \'
) or change your delimiter.
if (preg_match('#<a\s.*?href=[\'"]mailto:([A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,6})[\'"].*?>.*?</a>#i', $subject, $regs)) {
$result = $regs[0];
} else {
$result = "";
}