Skip to main content
Version: v2

AnyOf

Describes a union of types, i.e. a range of possible types:

import { anyOf } from 'dynamodb-toolbox/schema/anyOf'

const pokeTypeSchema = anyOf(
string().const('fire'),
string().const('grass'),
string().const('water')
)

type PokeType = FormattedValue<typeof pokeTypeSchema>
// => 'fire' | 'grass' | 'water'

In this example, an enum would have done the trick. However, anyOf becomes particularly powerful when used in conjunction with a map and the enum or const directives of a primitive attribute, to implement polymorphism:

const captureSchema = anyOf(
map({
status: string().const('caught'),
// 👇 captureState.trainerId exists if status is "caught"...
trainerId: string()
}),
// ...but not otherwise! 🙌
map({ status: string().const('wild') })
)

// Discriminate on string enum attributes for faster parsing! 🙌
const fasterSchema = captureSchema.discriminate('status')

type Capture = FormattedValue<typeof captureSchema>
// =>
// | { status: "caught"; trainerId: string }
// | { status: "wild" }
warning

In the absence of discriminating attribute, the parsing an anyOf schema value returns the parsed output of the first sub-schema it validates against.

This means the order of the sub-schemas matters: you should always start with the strictest schemas.

info

For the moment, anyOf properties can only be set by using methods.

AnyOf elements can have any type. However, they must respect some constraints:

  • They cannot be optional or always required
  • They cannot be hidden or key (tagging the anyOf itself as key is enough)
  • They cannot have default or links
// ❌ Raises a type AND a run-time error
const union = anyOf(number(), string().optional())
const union = anyOf(number(), string().hidden())
const union = anyOf(number(), string().key())
const union = anyOf(number(), string().default('foo'))

Properties

.required()

string | undefined

Tags schema values as required (within items or maps). Possible values are:

  • 'atLeastOnce' (default): Required (starting value)
  • 'always': Always required (including updates)
  • 'never': Optional
const pokeTypeSchema = anyOf(
string().const('fire'),
string().const('grass'),
string().const('water')
).required()

// shorthand for `.required('never')`
const pokeTypeSchema = anyOf(...).optional()

.hidden()

boolean | undefined

Omits schema values during formatting:

const pokeTypeSchema = anyOf(
string().const('fire'),
string().const('grass'),
string().const('water')
).hidden()

.key()

boolean | undefined

Tags schema values as a primary key attribute or linked to a primary key attribute:

// Note: The method also sets the `required` property to 'always'
// (it is often the case in practice, you can still use `.optional()` if needed)
const pokeTypeSchema = anyOf(
string().const('fire'),
string().const('grass'),
string().const('water')
).key()

.savedAs(...)

string

Renames schema values during the transformation step (within items or maps):

const pokeTypeSchema = anyOf(
string().const('fire'),
string().const('grass'),
string().const('water')
).savedAs('pkt')

.discriminate(...)

string

Leverages a specific attribute as a discriminator to efficiently match between different schema options. Optimizes performance during Parsing and Formatting:

const catSchema = map({
kind: string().enum('cat'),
... // Other cat attributes
})

const dogSchema = map({
kind: string().enum('dog'),
... // Other dog attributes
})

const petSchema = anyOf(catSchema, dogSchema)
.discriminate('kind')

You can retrieve a matching schema using the match method:

const matchingSchema = petSchema.match('dog') // => dogSchema

To be used as a discriminator, an attribute must meet all of the following conditions:

  • ✅ It must be present within a map attribute, either directly or as part of another anyOf schema.
  • ✅ It must be present in every schema option.
  • ✅ It must be of type string and use the enum property.
  • ✅ If renamed, the same savedAs value must be used across all options.
  • ❌ It must not be optional or transformed.

The following examples raises both type and runtime errors:

// ❌ 'age' is not a string with enum values
const petSchema = anyOf(
map({ age: number().enum(1, 2, 3) })
).discriminate('age')

// ❌ 'kind' is marked as optional
const petSchema = anyOf(
map({ kind: string().enum('cat').optional() })
).discriminate('kind')

// ❌ 'savedAs' is inconsistent across options
const petSchema = anyOf(
map({ kind: string().enum('cat').savedAs('kind') }),
map({ kind: string().enum('dog').savedAs('__kind__') })
).discriminate('kind')

.default(...)

ValueOrGetter<ATTRIBUTES>

Specifies default values. See Defaults and Links for more details:

Examples
const now = () => new Date().toISOString()

const hasUpdateSchema = anyOf(
map({ hasUpdate: boolean().const(false) }),
map({ hasUpdate: boolean().const(true), date: string() })
)
.default(() => ({ hasUpdate: false }))
.updateDefault(() => ({ hasUpdate: true, date: now() }))
// 👇 Similar to
const timestampsSchema = anyOf('...')
.putDefault(() => ({ hasUpdate: false }))
.updateDefault(() => ({ hasUpdate: true, date: now() }))

.link<Schema>(...)

Link<SCHEMA, ELEMENTS>

Similar to .default(...) but allows deriving the default value from other attributes. See Defaults and Links for more details:

const pokemonSchema = item({
name: string().optional(),
level: number()
}).and(prevSchema => ({
metadata: anyOf(string(), number()).link<
typeof prevSchema
>(
// 🙌 Correctly typed!
({ name, level }) => name ?? level
)
}))

.validate(...)

Validator<ELEMENTS>

Adds custom validation. See Custom Validation for more details:

Examples
const nonEmptyListSchema = anyOf(
list(string()),
list(number())
).validate(input => input.length > 0)
// 👇 Similar to
const nonEmptyListSchema = anyOf(
list(string()),
list(number())
).putValidate(input => input.length > 0)