Re: Capture only first match in regular expression

Zapanaz <http://joecosby.com/code/mail.pl> wrote:

> I am parsing/page scraping some HTML. I know the first anchor tag <a>
> contains information I want.
>
> So I do this:
>
> if($content =~ /.*(<a.*<\/a>).*/i){
> $anchorContent = $1;

Another poster suggested that regular expressions aren't sufficient
for this. But you may be able to do it anyway if you can confidently
predict features of the incoming HTML.

That is, if you know "know the first anchor tag <a> contains
information" you want, you may also know other things about the HTML
you're trying to parse.

Given an anchor of the general form:

<a href=foo possible-other-arrtibutes=bar> Anchor-text </a>

If you know in advance that the "Anchor-text" is *not* an <IMG
src=...> tag and that the "Anchor-text" does not itself contain any
other tags (such as, say, "<i>Anchor-text</i>) then you could use:

if($content =~ /(<a\s[^>]+>[^<]*<\/a>)/i)
{
$anchorContent = $1;
}

Match the <a literally
Require some matching whitespace after the 'a'
Match anything that can occur within an opening <A...> tag
Match the closing '>' of the opening <a tag
Match any text except the '<' that will signal the closing </a> tag
Match the closing </a> tag

Won't work if the incoming HTML is arbitrary because you might have:

<a href=foo><img src=bar></a> or
<a href=foo> <i> Yow<b>!</b></i> </a>

I'm no expert but I suspect that to reliably match what you want from
any arbitrary HTML, you'll have to write a more general parser.

--
Mike Spencer Nova Scotia, Canada

Mike Spencer