This section lists the FQL data types. FQL supports static typing.

Not all types are representable in JSON. FQL handles values similar to GraphQL where non-JSON types are downcast to a JSON equivalent, which can then be used as-is or recast client-side to a richer type.

Type representation summary:

  • JSON representable values are rendered verbatim.

  • Non-JSON scalar values are reduced to a JSON-compatible representation.

  • Document objects are reduced to JSON objects that include all the fields.

  • Nested documents aren’t expanded.

  • Sets are reduced to the first page of the set, which is equivalent to calling .paginate() without arguments. The query response includes before and after page references as applicable.

  • Singleton objects are reduced to string descriptions.

Persistable types

Persistable types have values that can be reliably stored, retrieved, and updated in a Fauna database. Documents can only store persistable values. Field definitions only allow persistable types.

Persistable types are:

Scalar types

The scalar types are JSON serializable types that hold a single value. Types, including String, Date, and Time, are objects that have built-in methods and properties.


A boolean data type has a boolean literal value of true or false.

Boolean variables and expressions can be compared explicitly or implicitly against a boolean literal. In the following example, the if statements are equivalent:

let a = true
if (a == true) {
  // expression

if (a) {
  // expression

Comparison operators return a boolean value:

let a = 10
let b = 20
a > b


A Base64-encoded string representing a byte array.


reference: Date

A date without a time zone in the ISO-8601 calendar system, such as 2022-11-03.

A Date value is encoded as a string in responses.


See Number.


A 64-bit integer represented as a decimal number that uniquely identifies a resource.


See Number.


See Number.


The Null type has the single value null.

Null is a marker used to indicate that a data value doesn’t exist. It is a representation of missing information, indicating a lack of a value, which differs from a value of zero.

To delete a document field, set the field value to null. The field is not saved after it is deleted. See update().


A marker used to indicate that a document doesn’t exist or is inaccessible. Testing a NullDoc against a value of null returns true. NullDoc is returned by the byId() or byName() queries and by links from another doc (id or coll) when the linked document doesn’t exist or is inaccessible. The Doc part of the type name is a collection name.

For example, calling the byId() method on a Users collection with an ID that doesn’t exist returns a null value of type NullUsers. Calling the byId() method on Token with an ID that doesn’t exist returns a null value of type NullToken.

The following lists the NullDoc types:

NullDoc type Description Value


Returned type when an AccessProvider document doesn’t exist.



Returned type when a CollectionDef document doesn’t exist.



Returned type when a Document doesn’t exist in the <Collection>.



Returned type when a Credential document doesn’t exist.



Returned type when a Key document doesn’t exist.



Returned type when a Role document doesn’t exist.



Returned type when a Token document doesn’t exist.



FQL has built-in types that represent numbers, which are summarized in the following table:

Type Size, bits Description Range Examples


double-precision IEEE-754 floating point




Signed two’s complement integer

-231 to 231-1



Signed two’s complement integer

-263 to 263-1


Underscores are permitted in numeric literals as separators to aid readability but have no other significance.


reference: String

String literals are single-quoted string values or double-quoted interpolated string values from the FQL-supported character set.

Single-quoted strings can also include the " and # characters.

Double-quoted strings can also include the ' character.

Use \ to escape a character in a string.

Interpolated strings

Double-quoted strings support the interpolation of expressions in a string. Encode the interpolation string using the #{<expr>} character sequence, where <expr> is an FQL expression:

"Alice is #{3 + 2} years old."

evaluates to:

"Alice is 5 years old."

Heredoc strings

You can use heredoc strings for multiple lines of text that behave similarly to double-quoted strings.

A heredoc string is delimited by tokens that start with the << character sequence followed by a user-defined character and a line break. Conventionally, the token is uppercase characters. The same token terminates a heredoc string and is on a line by itself:

A multiline

A heredoc string removes leading whitespace at the beginning of each line, removing the same number of characters in each line:

   A multiline string
     with leading

evaluates to:

  A multiline string
    with leading


reference: Time

The Time type is an instant in time expressed as a calendar date and time of day in UTC, in the range:

-999999999-01-01T00:00:00Z to 9999-12-31T23:59:59.999999999Z

A Time value can store nanosecond precision.

Times can be inserted with offsets but are converted to UTC and the offset component is lost.

A Time value is encoded as a string in responses.


reference: TransactionTime

The TransactionTime type is the query transaction expressed as a calendar date and time of day in UTC, in the range:

-999999999-01-01T00:00:00Z to 9999-12-31T23:59:59.999999999Z

A TransactionTime value can store nanosecond precision.

A Time value is encoded as a string in responses.


Universally Unique IDentifier (UUID) uniquely identifies a resource.

Data reference types

The data reference types represent FQL schema entities and have built-in methods and properties.


reference: Collection

A Collection groups documents in a database.


reference: Collection document definition

A Collection definition document.


A database can have Collections, Documents, User-defined functions, security elements such as Keys, Tokens, Credentials, Access Providers, and child databases.


reference: Database

A Database definition document.


reference: Document

A record in a Collection. You add documents to a collection as JSON-like objects. The document type is taken from the name of the collection of which the document is a member. For example, a document in the Product collection is of type Product, and if the requested document isn’t a member of the collection, the response type is NullProduct.

All documents have metadata fields.


reference: Function

A Function is an FQL expression stored in a database.


reference: Function document definition

A Function definition document.

Advanced types

The advanced types represent complex objects or types that can hold multiple values. The iterable Array and Set types have built-in properties and methods.


The Any type denotes a field that can be of any other type or might be not present.


reference: Array

The Array type is a comma-separated list of expressions enclosed by []. An array can hold mixed types, including functions:

["a", 1 + 1, if (true) "true" else "false", null]

evaluates to:

["a", 2, "true", null]

For map(), forEach(), and similar methods, array literals are evaluated left to right.

The Array type is an iterable data structure that has an ordered collection of typed values. An Array:

  • can hold values of different types.

  • can be accessed only using positive integers.

  • is zero-indexed.


The type of all iterable types:

  • Array type

  • Set type


The Never type represents a value that never occurs or the return value of a function that never returns.

See abort() for an example.


The Module type is a singleton object that represents a grouping of functionality. Examples include Math, Query, and Function, and Collections.

A module gets serialized as an @mod value in the tagged format. Use the isa operator for testing a module type.


The Object type is the type of all objects and represents a JSON-like object whose contents are a collection of key:value pairs. The keys must be strings and the values must be valid FQL data types. The value expressions are evaluated sequentially in the order defined, left to right. Objects evaluate to their contents. Objects can be combined to emulate abstract data types found in other functional languages.

Object subtypes:

Object types may have zero or more field types and an optional wildcard field type.


{ x: number, y: number } // object with two fields
{ kind: "dog" | "cat" | "bird", age: long }

The wildcard field type opens an object to arbitrary fields that correspond to the wildcard type:

// An object with one `long` field and any number of `string` fields
{ count: long, *: string }

// An object with any number of arbitrary fields
{ *: any }


reference: Set

A Set is an iterable group of values, typically representing documents in a collection.


Every primitive value is also a Singleton type.

  • Strings: "foo" "bar"

  • Integers: 1 2 3 -99

  • Booleans: true or false

  • null

Singletons can be combined to emulate abstract data types found in other functional languages.


reference: Event Streaming reference

A stream subscribes to tracked changes in a database. When a tracked change occurs, Fauna sends a related event to the stream.

Streams are represented by a stream token, a string used to subscribe to the stream.


A Struct is a plain Object that isn’t a Document. The value returned from projection on a document is a Struct.


Tuples are sequences of zero or more typed values:

[] // The empty tuple
[string, string] // tuple of two string values
[boolean] // tuple of one boolean value
[string, long, boolean] // tuple of a string, long, and boolean
[string[], long[]] // tuple of an array of strings and an array of longs


Tuple values are subtypes of arrays of the union of the tuple slot types:

[string, string] // string[]
[boolean] // boolean[]
[string, long, boolean] // (string | long | boolean)[]


Union type allows values of any constituent types:

boolean | number // all booleans and numbers
{ x: number } | string // all the object types and strings


Unions can be combined to emulate abstract data types found in other functional languages:

{ type: "point", x: long, y: long } | { type: "circle", x: long x: long, radius: long }

Security-related types

The security data types are used with Fauna authentication and authorization APIs.


reference: AccessProvider

An AccessProvider is a JSON object document subtype stored in a database AccessProvider collection that represents an AccessProvider.


reference: Credential

A Credential object represents a Credential in a Fauna database.


reference: Key

A Key is a JSON object document subtype stored in a database Key collection that represents an Key. A key ensures anonymous access to a database to execute operations permitted by the role associated with the key.


reference: Role

A Role is a JSON object document subtype stored in a database Role collection that represents an Role.


reference: Token

Tokens are defined as documents in the Token collection, and are used to control identity-based access to a database.

Is this article helpful? 

Tell Fauna how the article can be improved:
Visit Fauna's forums or email docs@fauna.com

Thank you for your feedback!