📋 Contents
Why Regex Looks Scary (And Why It's Not)
This is a regex for email validation:
^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\.[a-zA-Z]{2,}$
At first glance: gibberish. But broken down, it's just a list of instructions:
^— start of the string[a-zA-Z0-9._%+-]+— one or more letters, numbers, dots, underscores, etc.@— literal @ symbol[a-zA-Z0-9.-]+— one or more domain characters\.— literal dot[a-zA-Z]{2,}— two or more letters (TLD)$— end of string
Regex isn't magic. It's a pattern-matching DSL (domain-specific language) with about 15 rules total. Most people never learn it because they try to memorize every metacharacter at once instead of learning the core building blocks and composing them.
The 10 Regex Patterns You'll Actually Use
This is the Pareto principle applied to regex. Learn these and you can solve 90% of real-world pattern matching:
| Pattern | Meaning | Example match |
|---|---|---|
. | Any single character (except newline) | h.t matches "hat", "hot", "h3t" |
[abc] | Any one character in brackets | [aeiou] matches any vowel |
[^abc] | Any one character NOT in brackets | [^0-9] matches any non-digit |
\d | Digit (0-9) | \d{3} matches "123", "456" |
\w | Word character (letter, digit, underscore) | \w+ matches "hello_world123" |
\s | Whitespace (space, tab, newline) | a\s+b matches "a b", "a b" |
+ | One or more | \d+ matches "5", "42", "1337" |
* | Zero or more | go*al matches "gal", "goal", "goooal" |
? | Zero or one (optional) | colou?r matches "color", "colour" |
{n,m} | Between n and m times | \d{2,4} matches "42", "1337" |
^ | Start of string | ^Hello matches "Hello world" |
$ | End of string | world$ matches "Hello world" |
(...) | Capture group | (\d+)-(\d+) captures "123" and "456" from "123-456" |
| | OR (alternation) | cat|dog matches "cat" or "dog" |
a, @), metacharacter (has special meaning, like ., +, \d), or escaped (metacharacter made literal with backslash, like \., \+). That's it. Three categories.Real Examples for Common Tasks
Extract all email addresses from text
[\w.+-]+@[\w-]+\.[\w.-]+
Validate a phone number (loose)
\+?\d[\d\s()-]{7,}\d
Matches: +1 (555) 123-4567, 555-123-4567, 15551234567
Extract URLs from HTML
https?://[^\s"'<>]+
Matches http:// and https:// URLs until a whitespace, quote, or angle bracket.
Find duplicate words
\b(\w+)\s+\1\b
\b = word boundary, (\w+) captures a word, \s+ = whitespace, \1 = backreference to the captured word. Matches "the the", "is is".
Strip HTML tags
<[^>]*>
Matches anything between angle brackets. But: don't parse HTML with regex. For tag stripping from user input, it's fine. For anything structural, use an HTML parser.
Validate ISO date format (YYYY-MM-DD)
^\d{4}-(0[1-9]|1[0-2])-(0[1-9]|[12]\d|3[01])$
Matches "2026-07-28". Rejects "2026-13-01" (month 13), "2026-00-05" (month 00).
How to Build and Test Any Regex
Don't write regex blind. Build it incrementally and test at each step:
- Start with a literal example of what you want to match. "I want to match '2026-07-28'".
- Replace literal parts with patterns one at a time. Year →
\d{4}, hyphen →-, month →(0[1-9]|1[0-2]). - Test against positive cases. Does it match all the variations you want?
- Test against negative cases. Does it correctly REJECT things that shouldn't match?
- Add anchors (
^and$) if you want to match the entire string, not a substring.
Use our Regex Tester for instant feedback. Paste your test text, write your pattern, see matches highlighted in real time. Also check out the Regex Explainer — paste any regex and it breaks down what each character does in plain English.
When NOT to Use Regex
Regex is a hammer, but not everything is a nail. Don't use regex for:
- Parsing HTML/XML/JSON — These are nested structures. Regex matches flat patterns. You'll write a monster that breaks on edge cases. Use a parser.
- Email validation (RFC 5322 compliant) — The full email spec allows nested comments, quoted local parts with spaces, IP address domains with CIDR notation. The RFC-compliant regex is ~6,400 characters. Your 30-character regex will reject valid emails and accept invalid ones. Use a library, or a simple "has @ and dot" check + send a verification email.
- Calendar date arithmetic — Regex can match "2026-02-29" (which doesn't exist most years) but can't do leap year math. Validate format with regex, validate semantics with code.
- Arbitrary-depth nesting — Regex (without recursion extensions) can't match balanced parentheses, nested brackets, or matching open/close tags at arbitrary depth. This is a fundamental limitation of regular languages.