JavaScript RegExp Groups

Parentheses in a regex do two jobs. They group pieces together so you can apply a quantifier or alternation to the whole group, and they capture the text that matched so you can pull it out or reuse it. Capturing is what turns regex from "does it match?" into "give me the parts."

// Capturing Groups

Without the g flag, match returns the full match at index 0, then each group in order:

const m = "Due: 2026-10-02".match(/(\d{4})-(\d{2})-(\d{2})/);

console.log(m[0]);   // "2026-10-02"  the whole match
console.log(m[1]);   // "2026"        group 1
console.log(m[2]);   // "10"          group 2
console.log(m[3]);   // "02"          group 3

Groups are numbered by the position of their opening parenthesis, counting from the left.

// Destructuring makes it tidy (skip index 0 with a leading comma)
const [, year, month, day] = "2026-10-02".match(/(\d{4})-(\d{2})-(\d{2})/);

exec() works the same way

The regex method exec() (more in RegExp Objects) returns the same kind of array, plus the position where the match started:

const result = /(\w+) loves (\w+)/.exec("Alice loves Bob-");

console.log(result[0]);
console.log(result[1], result[2]);
console.log(result.index);

// Named Groups

Numbers get confusing with many groups. Give each one a name with (?<name>...), then read it from .groups:

const re = /(?<year>\d{4})-(?<month>\d{2})-(?<day>\d{2})/;
const { year, month, day } = "2026-10-02".match(re).groups;

console.log(`${month}/${day}/${year}`);

// Using Groups in replace()

In the replacement string, $1, $2... insert numbered groups and $<name> inserts named ones:

"2026-10-02".replace(/(\d{4})-(\d{2})-(\d{2})/, "$2/$3/$1");   // "10/02/2026"

"Lovelace, Ada".replace(/(?<last>\w+), (?<first>\w+)/, "$<first> $<last>");   // "Ada Lovelace"

// Grouping Without Capturing: (?:...)

Sometimes you only need parentheses for grouping. (?:...) groups but does not capture, keeping the numbering clean:

// Repeat a whole group
/(?:ha)+/.test("hahaha")      // true: "ha" one or more times

// Limit an alternation
/^(?:cat|dog)$/.test("dog")   // true: exactly "cat" or "dog"

// Only the second group is captured
"img.png".match(/(?:\w+)\.(\w+)/)[1]   // "png"

// Backreferences: Matching the Same Text Again

Inside the pattern, \1 means "whatever group 1 matched." This finds repeated words:

const text = "This is is a test of the the checker";

text.match(/\b(\w+) \1\b/g);                      // ["is is", "the the"]
text.replace(/\b(\w+) \1\b/g, "$1");              // "This is a test of the checker"

// Named version: \k<name>
/(?<q>["']).*?\k<q>/   // a quote, text, then the SAME kind of quote

// Two Group Traps

A repeated group keeps only its last repetition, and a group that did not take part in the match gives undefined:

console.log("abc".match(/(\w)+/));        // the + repeats the group 3 times

const m = "color".match(/colo(u)?r/);
console.log(m[1]);                           // the optional group matched nothing
console.log("color".replace(/colo(u)?r/, "[$1]"));

// Every Match, With Its Groups: matchAll

With the g flag, match returns only the full matches and drops the groups. Use matchAll to get groups for every match:

const log = "ava=92, ben=85, cal=78";

for (const m of log.matchAll(/(?<name>\w+)=(?<score>\d+)/g)) {
  console.log(m.groups.name, Number(m.groups.score));
}

// Where Each Group Was Found: the d Flag

As promised in RegExp Flags: with d, the result's indices holds a [start, end] pair for the whole match and for every group. Named groups also appear under indices.groups:

const m = "Due: 2026-10-02".match(/(?<y>\d{4})-(?<mo>\d{2})/d);

console.log(m[0], m.indices[0]);         // whole match
console.log(m.indices[1], m.indices[2]);   // group 1, group 2
console.log(m.indices.groups.y);           // the same pair, by name

// Group Modifiers: (?flags:...)

RegExp Flags showed how to turn a flag on for one group with (?i:...). A minus sign turns a flag off for one group instead:

const re = /(?-i:W3)schools/i;   // i everywhere EXCEPT inside the group

console.log(re.test("W3SCHOOLS"));   // true: "schools" ignores case
console.log(re.test("w3SCHOOLS"));   // false: "W3" must match exactly

Only i, m, and s can be switched this way. Any other letter is an error: /(?g:a)/ is SyntaxError: Invalid regular expression: /(?g:a)/: Invalid group.

// Summary

// Quick Check

console.log("aaa-bbb".match(/(\w)+-/)[1]);

// Try It

▶ Open Playground

1. Click Open Playground. Extract the hours and minutes from "Meet at 14:35" with two capturing groups
2. Rewrite it with named groups and destructure .groups
3. Swap "first last" names to "last, first" with replace and $1 / $2
4. Use matchAll to list every #hashtag in a sentence without the #