Regex Groups, Backreferences & Lookarounds ...

Once you're past character classes and quantifiers, groups and assertions are what make regex genuinely powerful — extracting parts of a match, matching repeats, and matching based on context. Here's how they work.

This is a supporting guide in The Complete Guide to Regular Expressions.

Capturing groups

Parentheses do two things: they group part of a pattern so a quantifier or alternation applies to the whole thing, and they capture what matched so you can pull it out afterward.

(\d{4})-(\d{2})-(\d{2})   on   2026-08-28
group 1 = 2026   group 2 = 08   group 3 = 28

Groups are numbered left to right by their opening parenthesis. In a replacement, you refer back to them with $1, $2, … — so replacing with $3/$2/$1 turns the date into 28/08/2026.

Named groups

Numbered groups get hard to track. Named groups give them labels: (?<year>\d{4}) captures into a name you reference as year (or $<year> in a replacement). Clearer, and stable if you reorder the pattern.

Non-capturing groups and alternation

Sometimes you want to group without capturing — to apply a quantifier or an alternation. Use (?:...):

(?:ab)+ — one or more "ab", but not captured
(cat|dog|fish) — alternation: any one of the three (captured)
(?:https?|ftp):// — group the scheme options without a capture slot

Backreferences

A backreference matches the same text a previous group captured. (["']) followed by \1 matches a matching pair of quotes:

(["']).*?\1   matches   "hello"   and   'hi'
              but not   "mismatched'

\1 refers to whatever group 1 actually captured, so the closing quote must be the same character as the opening one.

Lookahead and lookbehind

Lookarounds assert that something does (or doesn't) come next, without consuming it — the match position doesn't move past the asserted text:

foo(?=bar) — "foo" only if followed by "bar" (positive lookahead)
foo(?!bar) — "foo" only if not followed by "bar" (negative lookahead)
(?<=\$)\d+ — digits only if preceded by "$" (positive lookbehind)
(?<!\$)\d+ — digits not preceded by "$" (negative lookbehind)

A common use: (?=.*\d)(?=.*[a-z]) as password rules — assert a digit exists and a lowercase letter exists, without those assertions consuming any characters.

Part of The Complete Guide to Regular Expressions. See also Character Classes, Quantifiers & Anchors.

Try it

Build a pattern with groups and lookarounds in the Regex Tester and watch each capture group appear as you type — all in your browser.

Regex Groups, Backreferences & Lookarounds