正则表达式匹配不在特定 Div 中的标题标签

Question

So I have PHP code that puts out HTML that looks like this:

<div class="wrapper">
<h2>This is a header</h2>
<h2>This is one too/h2>
<h4>Here's one</h4>
<div class="ignore">
<h5>I'm one in here too</h5>
</div>
</div>

What I'm trying to do is preg_match_all of the header tags. My regular expression (<h([1-6]{1})[^>]*)>.*<\\/h\\2> returns all of them appropriately, but I don't want to grab the headers that are in the div with the class "ignore". I was reading about negative lookaheads, but it gets tricky. Anyone with help will be appreciated.

Desired output:

<h2>This is a header</h2>
<h2>This is one too/h2>
<h4>Here's one</h4>

Note I'm one in here too is omitted because it's wrapped in div with class "ignore".

Answer 1

Don't mess around with regular expressions here - unleash the power of DOMDocument in combination with xpath queries:

<?php
$html = <<<EOT
<div class="wrapper">
<h2>This is a header</h2>
<h2>This is one too</h2>
<h4>Here's one</h4>
<div class="ignore">
<h5>I'm one in here too</h5>
</div>
</div>
EOT;

$doc = DOMDocument::loadHTML($html);
$xpath = new DOMXpath($doc);
$headers = $xpath->query("
    //div[not(contains(@class, 'ignore'))]
    /*[self::h2 or self::h4 or self::h5]");

foreach ($headers as $header) {
    echo $header->nodeValue . "\n";
}

?>

This will yield

This is a header
This is one too
Here's one

Answer 2

With DOMDocument and DOMXPath :

$html = <<<'HTML'
<div class="wrapper">
<h2>This is a header</h2>
<h2>This is one too</h2>
<h4>Here's one</h4>
<div class="ignore">
<h5>I'm one in here too</h5>
</div>
</div>
HTML;

$dom = new DOMDocument;
$dom->loadHTML($html);
$xp = new DOMXPath($dom);

$nodeList = $xp->query('
//*
[contains(";h1;h2;h3;h4;h5;h6;", concat(";", local-name(), ";"))]
[not(ancestor::div[
    contains(concat(" ", normalize-space(@class), " "), " ignore ")
    ])
]');

foreach ($nodeList as $node) {
    echo 'tag name: ', $node->nodeName, PHP_EOL,
         'html content: ', $dom->saveHTML($node), PHP_EOL,
         'text content: ', $node->textContent, PHP_EOL,
         PHP_EOL;
}

demo

If you aren't comfortable with XPath take a look at the zvon tutorial .

Answer 3

Since you specify you want to do it with preg_match(), here is an example of a negative look-behind (ie filters out those occurrences NOT preceded by XYZ) : https://regex101.com/r/FeAsuj/1

The lookbehind itself is (?<!<div class=\\"ignore\\">) .

But in the test-snippet, notice how :

• the regex depends on the exact use of whitespace ...
• ... so a platform-dependant \\r\\n can break the regex
• the lookbehind cannot have a variable length, ie \\n? - see Regular Expression Lookbehind doesn't work with quantifiers ('+' or '*')

If you MUST continue to work with regex's, consider a 2-step approach :

• step 1, you use preg_replace() to eliminate all unwanted sections.
• step 2, use your existing regex.

In general, I would concur with the other posters to avoid regex, and go with a HTML parser.