如何提取某些HTML标记，例如 <ul>在PHP中使用带有preg_match_all的Regex？

Question

I am new to regular expressions. I want to fetch some data from a web page source. I used file_get_contents("url") to get the page's HTML source. Now I want to capture a portion within some special tags.

I found preg_match_all() works for this. Now I want some help to solve my problem and if possible help me to find out how to solve similar problems like this.

In the example below, how would I get the data within the <ul>? (I wish this sample HTML code could be easier for me to understand.)

<div class="a_a">qqqqq<span>www</span> </div>
<ul>
<li>
    <div class="a_a"><h3>aaaa</h3> aaaa aaaaa</div>
</li>
<li>
    <div class="b_b">bbbbb <span class="s-s">bbbb</span> bbbb</div>
</li>
<li>
    <div class="c_c d-d">cccc cccc ccccc</div>
</li>
</ul>
<table>
<tr>
    <td>sdsdf</td>
    <td>hjhjhj</td>
</tr>
<tr>
    <td>yuyuy</td>
    <td>ertre</td>
</tr>   
</table>

Answer 1

As the comments stated already, it's generally not recommended to parse html with regex. In my opinion, it depends on what exactly you're going to do.

---

If you want to use regex and know, that there are no nested tags of the same kind, the most simple pattern for getting everything that's between <ul> and closest </ul> would be:

$pattern = '~<ul>(.*?)</ul>~s';

It matches <ul> followed by as few characters of any kind as possible to meet </ul>. The dot is a metacharacter, that matches any single character except newlines ( ). To make it match newlines too, after the ending delimiter ~ I put the s-modifier. The quantifier * means zero or more times.

By default quantifiers are greedy, which means, they eat up as much as possible to be satisfied. A question-mark ? after the * makes them non-greedy (or lazy) and match as few characters as possible to meet </ul>. As pattern-delimiter I chose the ~ tilde.

preg_match_all($pattern, $html, $out);

Matches are captured and can be found in the output-variable, that you set for preg_match or preg_match_all, where [0] contains everything, that matches the whole pattern, [1] the first captured parenthesized subpattern, ...

---

If your searched tag can contain attributes (e.g. <ul class="my_list"...) this extended pattern, would after <ul also include [^>]* any amount of characters, that are not > before meeting >

$pattern = '~<ul[^>]*>\K.*(?=</ul>)~Uis';

Instead of the question-mark, here I use the U-modifier, to make all quantifiers lazy. For only getting captured the desired parts, that are <ul> inside </ul>. \K is used to reset beginning of the reported match. Instead of capturing the ending </ul> a lookahead is used (?=, as we neither want that part in the output.

This is basically the same as '~<ul[^>]*>(.*)</ul>~Uis' which would capture whole-pattern matches to [0] and first parenthesized group to [1].

---

But, if your html contains nested tags of same kind, the idea of the following pattern is to catch the innermost ones. At each character inside <ul>...</ul> it checks if there is no opening <ul

$pattern = '~<ul[^>]*>\K(?:(?!<ul).)*(?=</ul>)~Uis';

Get matches using preg_match_all

$html = '<div><ul><li><ul><li>.1.</li></ul>...</li></ul></div>
         <ul><li>.2.</li></ul>';

if(preg_match_all($pattern, $html, $out))
{
  echo "<pre>"; print_r(array_map('htmlspecialchars',$out[0])); echo "</pre>";
} else {

  echo "FAIL";
}

Matches between \K and (?= will be captured to $out[0]

• \K resets beginning of the reported match (supported in PHP since 5.2.4)
• the second pattern, when <ul> matched, looks ahead (?!... at each character, if there's no opening <ul before meeting </ul>, if so starts over until </ul> is ahead (?=</ul>).
• [^>]* any amount of characters, that are not > (negated character class)
• (?: starts a non-capturing group.

Used Modifiers: Uis (part after the ending delimiter ~)

U (PCRE_UNGREEDY), i (PCRE_CASELESS), s (PCRE_DOTALL)