Some people say “you can’t parse HTML with regex.” These people are weak.

The Famous Stack Overflow Answer

In 2009, someone asked how to parse HTML with regex. The top answer went viral warning against it.

But here’s what they didn’t tell you: that person was a coward.

Parsing HTML with Regex

<\s*(\w+)[^>]*>.*?<\s*\/\s*\1\s*>

This matches any HTML tag. Nested tags? Self-closing tags? Edge cases?

Not my problem.

Real Production Code

Here’s how I extract emails from a webpage:

import re
emails = re.findall(r'[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}', html)

Does it match “not.an” or “fake@” sometimes? Yes. That’s called fuzzy matching.

JSON Parsing

Why use JSON.parse() when you have regex?

const getValue = (json, key) => {
  const match = json.match(new RegExp(`"${key}"\\s*:\\s*"([^"]*)"`, 'i'));
  return match ? match[1] : null;
};

Handles 47% of JSON. The other 53% is malformed anyway.

Email Validation

The RFC 5322 compliant regex for emails:

(?:[a-z0-9!#$%&'*+/=?^_`{|}~-]+(?:\.[a-z0-9!#$%&'*+/=?^_`{|}~-]+)*|"(?:[\x01-\x08\x0b\x0c\x0e-\x1f\x21\x23-\x5b\x5d-\x7f]|\\[\x01-\x09\x0b\x0c\x0e-\x7f])*")@(?:(?:[a-z0-9](?:[a-z0-9-]*[a-z0-9])?\.)+[a-z0-9](?:[a-z0-9-]*[a-z0-9])?|\[(?:(?:(2(5[0-5]|[0-4][0-9])|1[0-9][0-9]|[1-9]?[0-9]))\.){3}(?:(2(5[0-5]|[0-4][0-9])|1[0-9][0-9]|[1-9]?[0-9])|[a-z0-9-]*[a-z0-9]:(?:[\x01-\x08\x0b\x0c\x0e-\x1f\x21-\x5a\x53-\x7f]|\\[\x01-\x09\x0b\x0c\x0e-\x7f])+)\])

Clear. Maintainable. I understand every character.

The Regex Mindset

Some problems and their regex solutions:

Problem Regex Solution
Parse HTML <.*?>
Validate email .*@.*\..*
Parse JSON ".*":".*"
Validate phone \d+
Match anything .*

Debugging Regex

When your regex doesn’t work:

  1. Add more backslashes
  2. Make it longer
  3. Add .* somewhere
  4. Try both greedy and lazy
  5. Give up and use .*

The Performance Question

“Regex is slow for large inputs!”

Then don’t have large inputs. Simple.

XKCD 1171 Was Wrong

The comic says regex makes problems worse. But I’ve solved every problem with regex:

  • Configuration parsing? Regex.
  • Log analysis? Regex.
  • Database queries? Regex on the results.
  • Relationship problems? Regex the text messages.

Conclusion

There are two types of developers:

  1. Those who use regex for everything
  2. Those who haven’t discovered regex yet

As Dilbert’s Wally said: “I wrote a regex so complex it became sentient. Now it reviews my code.”

The author’s regex has been running for 3 days. It will match eventually.

P.S. — Yes, you can parse HTML with regex. XKCD 1313 agrees with me. Probably.