如何使用DOMXPath获取<h2>标签后div内的<p>内容？

with DOMXPath #

Hey there! I’ve hit this exact XPath roadblock before—let’s work through how to fix it. First, let’s assume your HTML structure looks like this (super common for this scenario):

<h2>My Awesome Heading</h2>
<div class="section-content">
  <p>This is the paragraph we need to grab!</p>
  <p>Maybe there's a second one too!</p>
</div>

The most likely issue is that you’re using child-node syntax (/div) instead of targeting sibling nodes. Here’s how to fix it:

1. Grab the first

immediately after a specific

#

If the

is the direct next sibling of your

, use following-sibling::div[1] to target it, then drill down to the

tags:

$dom = new DOMDocument();
libxml_use_internal_errors(true); // Suppress messy HTML parsing warnings
$dom->loadHTML($yourHtmlString);
$xpath = new DOMXPath($dom);

// Target the <h2> with specific text, then its next div, then all <p>s
$targetHeading = "My Awesome Heading";
$paragraphs = $xpath->query("//h2[normalize-space(text())='$targetHeading']/following-sibling::div[1]/p");

// Loop through results if there are multiple paragraphs
foreach ($paragraphs as $p) {
    echo trim($p->nodeValue) . "\n";
}

2. Grab a

that comes after the

(not necessarily immediately) #

If there are other elements between the

and

, use following::div[1] instead—it finds the first

anywhere after the

:

$paragraphs = $xpath->query("//h2[normalize-space(text())='$targetHeading']/following::div[1]/p");

Common Mistakes to Avoid #

• Using /div instead of sibling syntax: /div looks for child nodes of the

  , but your

  is a sibling, not a child. That’s almost certainly why your initial queries failed!

• Forgetting normalize-space(): If your

  has extra whitespace (like newlines or tabs), text() won’t match exactly. normalize-space() cleans that up.

• Not specifying [1]: If there are multiple
  s after the

  , omitting [1] will grab all of them—adjust based on your needs.

If your HTML structure is a bit different (like nested elements), share the exact markup and I can tweak the XPath further—but these patterns work for most common cases.

内容的提问来源于stack exchange，提问作者Gislef