如何删除字符串中的元素内容，只留下最外层的元素标签？

Question

I have a string like this:

<p>
This is some text
</p>

<p>
This is some text
</p>
<p>
This is some text
</p>

<blockquote data-id="1">
    This is some text

    <blockquote data-id="2">
        This is some text
    </blockquote>
</blockquote>

<blockquote data-id="3">
    <blockquote data-id="4">
        This is some text

        <blockquote data-id="5">
            This is some text
        </blockquote>
    </blockquote>
    This is some text
</blockquote>

<blockquote data-id="6">
    This is some text
</blockquote>

I want to keep the outermost blockquote tags, but delete the contents. So I want to convert the above to this:

<p>
This is some text
</p>

<p>
This is some text
</p>
<p>
This is some text
</p>

<blockquote data-id="1"></blockquote>

<blockquote data-id="3"></blockquote>

<blockquote data-id="6"></blockquote>

What is an efficient way to do this in PHP?

Answer 1

Many ways to skin this cat. I'd give the string a dummy root node, ditch all nodes matching the xpath expression /root/blockquote/text() | /root/blockquote/*, then rebuild the string from the root's children.

---

Example:

$string = <<<'STRING'
<p>
This is some text
</p>

<p>
This is some text
</p>
<p>
This is some text
</p>

<blockquote data-id="1">
    This is some text

    <blockquote data-id="2">
        This is some text
    </blockquote>
</blockquote>

<blockquote data-id="3">
    <blockquote data-id="4">
        This is some text

        <blockquote data-id="5">
            This is some text
        </blockquote>
    </blockquote>
    This is some text
</blockquote>

<blockquote data-id="6">
    This is some text
</blockquote>
STRING;

$dom = new DOMDocument();
$dom->loadXML("<root>$string</root>");
$xpath = new DOMXPath($dom);

foreach ($xpath->query('/root/blockquote/text() | /root/blockquote/*') as $node) {
    $node->parentNode->removeChild($node);
}

$string = '';
foreach ($dom->documentElement->childNodes as $node) {
    $string .= $dom->saveHTML($node);
}

echo $string;

---

Output:

<p>
This is some text
</p>

<p>
This is some text
</p>
<p>
This is some text
</p>

<blockquote data-id="1"></blockquote>

<blockquote data-id="3"></blockquote>

<blockquote data-id="6"></blockquote>