Regex capture group

The basics

Parentheses create a numbered group. Groups are numbered by the position of their opening parenthesis, left to right, and group 0 is always the whole match.

Pattern: (\d{4})-(\d{2})-(\d{2}) applied to 2026-09-18:

Group	Text
0	2026-09-18
1	2026
2	09
3	18

Variations

• Non-capturing group: (?:...) groups without remembering. Use it when you only need grouping for a quantifier or alternation, such as (?:cat|dog)s?.
• Named group: (?<year>\d{4}) gives the group a name, so you read match.groups.year instead of counting positions. Named groups are part of modern JavaScript, and syntax varies between regex flavors.
• Backreference: \1 matches the same text group 1 matched. The pattern (\w+) \1 finds a repeated word such as “the the”.
• In a replacement: JavaScript uses $1, or $<year> for a named group. Swapping a date’s order looks like '2026-09-18'.replace(/(\d{4})-(\d{2})-(\d{2})/, '$3/$2/$1'), which gives 18/09/2026.

Common pitfalls

• A repeated group keeps only its last match. Applying (a|b)+ to abab leaves group 1 as b, not all four letters. To capture the whole run, wrap the repetition: ((?:a|b)+).
• An optional group that did not take part is undefined, not an empty string, in JavaScript. Check before using it.
• Miscounting nested groups. In ((a)(b)), group 1 is the outer one, group 2 is a and group 3 is b.
• Greedy versus lazy. (.*) grabs as much as it can. Use (.*?) to stop at the first opportunity.
• Catastrophic backtracking. Nested quantifiers like (a+)+$ can take exponential time on a non-matching input, which can freeze a page or a server. Keep groups with quantifiers simple, and test patterns on hostile input.
• Flavor differences. Lookbehind support, named-group syntax and Unicode handling all differ between JavaScript, Python, PCRE and others. Test in the flavor you deploy.