sort-type-constituents
Enforce constituents of a type union/intersection to be sorted alphabetically.
Some problems reported by this rule are automatically fixable by the --fix ESLint command line option.
Some problems reported by this rule are manually fixable by editor suggestions.
This rule has been deprecated in favor of the perfectionist/sort-intersection-types and perfectionist/sort-union-types rules.
See Docs: Deprecate sort-type-constituents in favor of eslint-plugin-perfectionist and eslint-plugin: Feature freeze naming and sorting stylistic rules for more information.
Sorting union (|) and intersection (&) types can help:
- keep your codebase standardized
- find repeated types
- reduce diff churn
This rule reports on any types that aren't sorted alphabetically.
Types are sorted case-insensitively and treating numbers like a human would, falling back to character code sorting in case of ties.
- Flat Config
- Legacy Config
export default tseslint.config({
  rules: {
    "@typescript-eslint/sort-type-constituents": "error"
  }
});
module.exports = {
  "rules": {
    "@typescript-eslint/sort-type-constituents": "error"
  }
};
Try this rule in the playground ↗
Examples
- ❌ Incorrect
- ✅ Correct
type T1 = B | A;
type T2 = { b: string } & { a: string };
type T3 = [1, 2, 4] & [1, 2, 3];
type T4 =
  | [1, 2, 4]
  | [1, 2, 3]
  | { b: string }
  | { a: string }
  | (() => void)
  | (() => string)
  | 'b'
  | 'a'
  | 'b'
  | 'a'
  | readonly string[]
  | readonly number[]
  | string[]
  | number[]
  | B
  | A
  | string
  | any;
type T1 = A | B;
type T2 = { a: string } & { b: string };
type T3 = [1, 2, 3] & [1, 2, 4];
type T4 =
  | A
  | B
  | number[]
  | string[]
  | any
  | string
  | readonly number[]
  | readonly string[]
  | 'a'
  | 'a'
  | 'b'
  | 'b'
  | (() => string)
  | (() => void)
  | { a: string }
  | { b: string }
  | [1, 2, 3]
  | [1, 2, 4];
Options
This rule accepts the following options:
type Options = [
  {
    /** Whether to sort using case sensitive string comparisons. */
    caseSensitive?: boolean;
    /** Whether to check intersection types (`&`). */
    checkIntersections?: boolean;
    /** Whether to check union types (`|`). */
    checkUnions?: boolean;
    /** Ordering of the groups. */
    groupOrder?: (
      | 'conditional'
      | 'function'
      | 'import'
      | 'intersection'
      | 'keyword'
      | 'literal'
      | 'named'
      | 'nullish'
      | 'object'
      | 'operator'
      | 'tuple'
      | 'union'
    )[];
  },
];
const defaultOptions: Options = [
  {
    caseSensitive: false,
    checkIntersections: true,
    checkUnions: true,
    groupOrder: [
      'named',
      'keyword',
      'operator',
      'literal',
      'function',
      'import',
      'conditional',
      'object',
      'tuple',
      'intersection',
      'union',
      'nullish',
    ],
  },
];
caseSensitive
Whether to sort using case sensitive string comparisons. Default: false.
Examples of code with { "caseSensitive": true }:
- ❌ Incorrect
- ✅ Correct
type T = 'DeletedAt' | 'DeleteForever';
type T = 'DeleteForever' | 'DeletedAt';
checkIntersections
Whether to check intersection types (&). Default: true.
Examples of code with { "checkIntersections": true } (the default):
- ❌ Incorrect
- ✅ Correct
type ExampleIntersection = B & A;
type ExampleIntersection = A & B;
checkUnions
Whether to check union types (|). Default: true.
Examples of code with { "checkUnions": true } (the default):
- ❌ Incorrect
- ✅ Correct
type ExampleUnion = B | A;
type ExampleUnion = A | B;
groupOrder
Ordering of the groups. Default: ["named","keyword","operator","literal","function","import","conditional","object","tuple","intersection","union","nullish"].
Each constituent of the type is placed into a group, and then the rule sorts alphabetically within each group. The ordering of groups is determined by this option.
- conditional- Conditional types (- A extends B ? C : D)
- function- Function and constructor types (- () => void,- new () => type)
- import- Import types (- import('path'))
- intersection- Intersection types (- A & B)
- keyword- Keyword types (- any,- string, etc)
- literal- Literal types (- 1,- 'b',- true, etc)
- named- Named types (- A,- A['prop'],- B[],- Array<C>)
- object- Object types (- { a: string },- { [key: string]: number })
- operator- Operator types (- keyof A,- typeof B,- readonly C[])
- tuple- Tuple types (- [A, B, C])
- union- Union types (- A | B)
- nullish-- nulland- undefined
For example, configuring the rule with { "groupOrder": ["literal", "nullish" ]}:
- ❌ Incorrect
- ✅ Correct
type ExampleGroup = null | 123;
type ExampleGroup = 123 | null;
When Not To Use It
This rule is purely a stylistic rule for maintaining consistency in your project. You can turn it off if you don't want to keep a consistent, predictable order for intersection and union types. However, keep in mind that inconsistent style can harm readability in a project.