JavaScript RegExp v Flag: Set Notation & Properties

The JavaScript RegExp v (unicodeSets) flag modernizes regular expressions by upgrading the capabilities of the existing u flag to support advanced set operations and multi-code-point Unicode properties. Introduced in ECMAScript 2024, this feature allows developers to perform set intersections, subtractions, and unions directly within character classes, while also enabling character classes to match sequences of characters such as complex emojis. By introducing these features, the v flag eliminates the need for complex lookaheads and manual string splitting, making pattern matching in modern JavaScript more expressive, reliable, and readable.

Set Operations: Intersection, Subtraction, and Union

Prior to the v flag, character classes ([...]) were limited to simple unions. The v flag introduces explicit set operations directly inside character classes using double operators:

• Intersection (&&): Matches characters that exist in both specified sets.

  // Matches Greek letters that are also uppercase
  const greekUpper = /[\p{Script=Greek}&&\p{Uppercase}]/v;
  greekUpper.test('Ω'); // true
  greekUpper.test('ω'); // false

• Subtraction (--): Removes characters in the second set from the first set.

  // Matches any decimal digit except ASCII 0-9
  const nonAsciiDigits = /[\p{Decimal_Number}--[0-9]]/v;
  nonAsciiDigits.test('٣'); // true (Arabic-Indic digit 3)
  nonAsciiDigits.test('3'); // false

• Nested Character Classes: Classes can be nested within each other to combine different operations cleanly.

  // Matches ASCII letters, except vowels
  const consonants = /[[a-z]--[aeiou]]/iv;
  consonants.test('b'); // true
  consonants.test('e'); // false

Properties of Strings

Under the older u flag, Unicode property escapes (\p{...}) could only match individual code points. The v flag introduces “properties of strings,” allowing a single property escape inside a character class to match multi-code-point sequences.

This is particularly useful for emojis composed of multiple Unicode code points (such as flag sequences, skin tone modifiers, and zero-width joiner sequences):

// Matches any valid emoji, including complex multi-character sequences
const emojiRegex = /^\p{RGI_Emoji}$/v;

emojiRegex.test('⚽'); // true (single code point)
emojiRegex.test('👨‍👩‍👧‍👦'); // true (multi-code-point sequence joined by ZWJ)
emojiRegex.test('🇨🇦'); // true (flag sequence of two regional indicators)

Without the v flag, matching multi-character emoji sequences required verbose regular expressions or manual tokenization.

String Literals in Character Classes

The v flag allows explicit string literals inside character classes using the \q{...} syntax. This makes it possible to treat whole words or character sequences as single elements within a set:

// Matches standard digits or the written-out words "one", "two", "three"
const digitOrWord = /[\d|\q{one|two|three}]/v;

digitOrWord.test('5'); // true
digitOrWord.test('two'); // true

Stricter Parsing and Compatibility

The v flag enforces stricter syntax rules compared to older flags to avoid ambiguity with the new operator syntax:

• Characters with special meanings in set operations ((, ), [, ], {, }, /, -, \, &, |, !, #, $, %, *, +, ,, ., :, ;, <, =, >, ?, @, ^, `, ~) must be escaped when used as literals inside [...].
• The v flag is mutually exclusive with the u flag (/pattern/uv throws a SyntaxError). However, the v flag includes all existing u flag features by default, making it a complete replacement for u.