Working Draft – October 2016
Introduction
This is a Draft RFC Specification for GraphQL, a query language created by Facebook in 2012 for describing the capabilities and requirements of data models for client‐server applications. The development of this standard started in 2015. GraphQL is a new and evolving language and is not complete. Significant enhancement will continue in future editions of this specification.
Copyright notice
Copyright (c) 2015‐2016, Facebook, Inc. All rights reserved.
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
GraphQL is a query language designed to build client applications by providing an intuitive and flexible syntax and system for describing their data requirements and interactions.
For example, this GraphQL request will receive the name of the user with id 4 from the Facebook implementation of GraphQL.
{
user(id: 4) {
name
}
}
Which produces the resulting data (in JSON):
{
"user": {
"name": "Mark Zuckerberg"
}
}
GraphQL is not a programming language capable of arbitrary computation, but is instead a language used to query application servers that have capabilities defined in this specification. GraphQL does not mandate a particular programming language or storage system for application servers that implement it. Instead, application servers take their capabilities and map them to a uniform language, type system, and philosophy that GraphQL encodes. This provides a unified interface friendly to product development and a powerful platform for tool‐building.
GraphQL has a number of design principles:
Because of these principles, GraphQL is a powerful and productive environment for building client applications. Product developers and designers building applications against working GraphQL servers -- supported with quality tools -- can quickly become productive without reading extensive documentation and with little or no formal training. To enable that experience, there must be those that build those servers and tools.
The following formal specification serves as a reference for those builders. It describes the language and its grammar, the type system and the introspection system used to query it, and the execution and validation engines with the algorithms to power them. The goal of this specification is to provide a foundation and framework for an ecosystem of GraphQL tools, client libraries, and server implementations -- spanning both organizations and platforms -- that has yet to be built. We look forward to working with the community in order to do that.
Clients use the GraphQL query language to make requests to a GraphQL service. We refer to these request sources as documents. A document may contain operations (queries and mutations are both operations) as well as fragments, a common unit of composition allowing for query reuse.
A GraphQL document is defined as a syntactic grammar where terminal symbols are tokens (indivisible lexical units). These tokens are defined in a lexical grammar which matches patterns of source characters (defined by a double‐colon ::
).
GraphQL documents are expressed as a sequence of Unicode characters. However, with few exceptions, most of GraphQL is expressed only in the original non‐control ASCII range so as to be as widely compatible with as many existing tools, languages, and serialization formats as possible and avoid display issues in text editors and source control.
Non‐ASCII Unicode characters may freely appear within StringValue and Comment portions of GraphQL.
The “Byte Order Mark” is a special Unicode character which may appear at the beginning of a file containing Unicode which programs may use to determine the fact that the text stream is Unicode, what endianness the text stream is in, and which of several Unicode encodings to interpret.
White space is used to improve legibility of source text and act as separation between tokens, and any amount of white space may appear before or after any token. White space between tokens is not significant to the semantic meaning of a GraphQL query document, however white space characters may appear within a String or Comment token.
Like white space, line terminators are used to improve the legibility of source text, any amount may appear before or after any other token and have no significance to the semantic meaning of a GraphQL query document. Line terminators are not found within any other token.
GraphQL source documents may contain single‐line comments, starting with the # marker.
A comment can contain any Unicode code point except LineTerminator so a comment always consists of all code points starting with the # character up to but not including the line terminator.
Comments behave like white space and may appear after any token, or before a line terminator, and have no significance to the semantic meaning of a GraphQL query document.
Similar to white space and line terminators, commas (,) are used to improve the legibility of source text and separate lexical tokens but are otherwise syntactically and semantically insignificant within GraphQL query documents.
Non‐significant comma characters ensure that the absence or presence of a comma does not meaningfully alter the interpreted syntax of the document, as this can be a common user‐error in other languages. It also allows for the stylistic use of either trailing commas or line‐terminators as list delimiters which are both often desired for legibility and maintainability of source code.
A GraphQL document is comprised of several kinds of indivisible lexical tokens defined here in a lexical grammar by patterns of source Unicode characters.
Tokens are later used as terminal symbols in a GraphQL query document syntactic grammars.
Before and after every lexical token may be any amount of ignored tokens including WhiteSpace and Comment. No ignored regions of a source document are significant, however ignored source characters may appear within a lexical token in a significant way, for example a String may contain white space characters.
No characters are ignored while parsing a given token, as an example no white space characters are permitted between the characters defining a FloatValue.
! | $ | ( | ) | ... | : | = | @ | [ | ] | { | | | } |
GraphQL documents include punctuation in order to describe structure. GraphQL is a data description language and not a programming language, therefore GraphQL lacks the punctuation often used to describe mathematical expressions.
GraphQL query documents are full of named things: operations, fields, arguments, directives, fragments, and variables. All names must follow the same grammatical form.
Names in GraphQL are case‐sensitive. That is to say name
, Name
, and NAME
all refer to different names. Underscores are significant, which means other_name
and othername
are two different names.
Names in GraphQL are limited to this ASCII subset of possible characters to support interoperation with as many other systems as possible.
A GraphQL query document describes a complete file or request string received by a GraphQL service. A document contains multiple definitions of Operations and Fragments. GraphQL query documents are only executable by a server if they contain an operation. However documents which do not contain operations may still be parsed and validated to allow client to represent a single request across many documents.
If a document contains only one operation, that operation may be unnamed or represented in the shorthand form, which omits both the query keyword and operation name. Otherwise, if a GraphQL query document contains multiple operations, each operation must be named. When submitting a query document with multiple operations to a GraphQL service, the name of the desired operation to be executed must also be provided.
query | mutation |
There are two types of operations that GraphQL models:
Each operation is represented by an optional operation name and a selection set.
For example, this mutation operation might “like” a story and then retrieve the new number of likes:
mutation {
likeStory(storyID: 12345) {
story {
likeCount
}
}
}
Query shorthand
If a document contains only one query operation, and that query defines no variables and contains no directives, that operation may be represented in a short‐hand form which omits the query keyword and query name.
For example, this unnamed query operation is written via query shorthand.
{
field
}
An operation selects the set of information it needs, and will receive exactly that information and nothing more, avoiding over‐fetching and under‐fetching data.
{
id
firstName
lastName
}
In this query, the id
, firstName
, and lastName
fields form a selection set. Selection sets may also contain fragment references.
A selection set is primarily composed of fields. A field describes one discrete piece of information available to request within a selection set.
Some fields describe complex data or relationships to other data. In order to further explore this data, a field may itself contain a selection set, allowing for deeply nested requests. All GraphQL operations must specify their selections down to fields which return scalar values to ensure an unambiguously shaped response.
For example, this operation selects fields of complex data and relationships down to scalar values.
{
me {
id
firstName
lastName
birthday {
month
day
}
friends {
name
}
}
}
Fields in the top‐level selection set of an operation often represent some information that is globally accessible to your application and its current viewer. Some typical examples of these top fields include references to a current logged‐in viewer, or accessing certain types of data referenced by a unique identifier.
# `me` could represent the currently logged in viewer.
{
me {
name
}
}
# `user` represents one of many users in a graph of data, referred to by a
# unique identifier.
{
user(id: 4) {
name
}
}
Fields are conceptually functions which return values, and occasionally accept arguments which alter their behavior. These arguments often map directly to function arguments within a GraphQL server’s implementation.
In this example, we want to query a specific user (requested via the id
argument) and their profile picture of a specific size
:
{
user(id: 4) {
id
name
profilePic(size: 100)
}
}
Many arguments can exist for a given field:
{
user(id: 4) {
id
name
profilePic(width: 100, height: 50)
}
}
Arguments are unordered
Arguments may be provided in any syntactic order and maintain identical semantic meaning.
These two queries are semantically identical:
{
picture(width: 200, height: 100)
}
{
picture(height: 100, width: 200)
}
By default, the key in the response object will use the field name queried. However, you can define a different name by specifying an alias.
In this example, we can fetch two profile pictures of different sizes and ensure the resulting object will not have duplicate keys:
{
user(id: 4) {
id
name
smallPic: profilePic(size: 64)
bigPic: profilePic(size: 1024)
}
}
Which returns the result:
{
"user": {
"id": 4,
"name": "Mark Zuckerberg",
"smallPic": "https://cdn.site.io/pic-4-64.jpg",
"bigPic": "https://cdn.site.io/pic-4-1024.jpg"
}
}
Since the top level of a query is a field, it also can be given an alias:
{
zuck: user(id: 4) {
id
name
}
}
Returns the result:
{
"zuck": {
"id": 4,
"name": "Mark Zuckerberg"
}
}
A field’s response key is its alias if an alias is provided, and it is otherwise the field’s name.
Fragments are the primary unit of composition in GraphQL.
Fragments allow for the reuse of common repeated selections of fields, reducing duplicated text in the document. Inline Fragments can be used directly within a selection to condition upon a type condition when querying against an interface or union.
For example, if we wanted to fetch some common information about mutual friends as well as friends of some user:
query noFragments {
user(id: 4) {
friends(first: 10) {
id
name
profilePic(size: 50)
}
mutualFriends(first: 10) {
id
name
profilePic(size: 50)
}
}
}
The repeated fields could be extracted into a fragment and composed by a parent fragment or query.
query withFragments {
user(id: 4) {
friends(first: 10) {
...friendFields
}
mutualFriends(first: 10) {
...friendFields
}
}
}
fragment friendFields on User {
id
name
profilePic(size: 50)
}
Fragments are consumed by using the spread operator (...
). All fields selected by the fragment will be added to the query field selection at the same level as the fragment invocation. This happens through multiple levels of fragment spreads.
For example:
query withNestedFragments {
user(id: 4) {
friends(first: 10) {
...friendFields
}
mutualFriends(first: 10) {
...friendFields
}
}
}
fragment friendFields on User {
id
name
...standardProfilePic
}
fragment standardProfilePic on User {
profilePic(size: 50)
}
The queries noFragments
, withFragments
, and withNestedFragments
all produce the same response object.
Fragments must specify the type they apply to. In this example, friendFields
can be used in the context of querying a User
.
Fragments cannot be specified on any input value (scalar, enumeration, or input object).
Fragments can be specified on object types, interfaces, and unions.
Selections within fragments only return values when concrete type of the object it is operating on matches the type of the fragment.
For example in this query on the Facebook data model:
query FragmentTyping {
profiles(handles: ["zuck", "cocacola"]) {
handle
...userFragment
...pageFragment
}
}
fragment userFragment on User {
friends {
count
}
}
fragment pageFragment on Page {
likers {
count
}
}
The profiles
root field returns a list where each element could be a Page
or a User
. When the object in the profiles
result is a User
, friends
will be present and likers
will not. Conversely when the result is a Page
, likers
will be present and friends
will not.
{
"profiles": [
{
"handle": "zuck",
"friends": { "count" : 1234 }
},
{
"handle": "cocacola",
"likers": { "count" : 90234512 }
}
]
}
Fragments can be defined inline within a selection set. This is done to conditionally include fields based on their runtime type. This feature of standard fragment inclusion was demonstrated in the query FragmentTyping
example. We could accomplish the same thing using inline fragments.
query inlineFragmentTyping {
profiles(handles: ["zuck", "cocacola"]) {
handle
... on User {
friends {
count
}
}
... on Page {
likers {
count
}
}
}
}
Inline fragments may also be used to apply a directive to a group of fields. If the TypeCondition is omitted, an inline fragment is considered to be of the same type as the enclosing context.
query inlineFragmentNoType($expandedInfo: Boolean) {
user(handle: "zuck") {
id
name
... @include(if: $expandedInfo) {
firstName
lastName
birthday
}
}
}
Field and directive arguments accept input values of various literal primitives; input values can be scalars, enumeration values, lists, or input objects.
If not defined as constant (for example, in DefaultValue), input values can be specified as a variable. List and inputs objects may also contain variables (unless defined to be constant).
0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
An Int number is specified without a decimal point or exponent (ex. 1
).
+ | - |
A Float number includes either a decimal point (ex. 1.0
) or an exponent (ex. 1e50
) or both (ex. 6.0221413e23
).
true | false |
The two keywords true
and false
represent the two boolean values.
" | \ | / | b | f | n | r | t |
Strings are sequences of characters wrapped in double‐quotes ("
). (ex. "Hello World"
). White space and other otherwise‐ignored characters are significant within a string value.
Semantics
Escaped Character | Code Unit Value | Character Name |
---|---|---|
" | U+0022 | double quote |
\ | U+005C | reverse solidus (back slash) |
/ | U+002F | solidus (forward slash) |
b | U+0008 | backspace |
f | U+000C | form feed |
n | U+000A | line feed (new line) |
r | U+000D | carriage return |
t | U+0009 | horizontal tab |
Null values are represented as the keyword null.
GraphQL has two semantically different ways to represent the lack of a value:
For example, these two field calls are similar, but are not identical:
{
field(arg: null)
field
}
The first has explictly provided null to the argument “arg”, while the second has implicitly not provided a value to the argument “arg”. These two forms may be interpreted differently. For example, a mutation representing deleting a field vs not altering a field, respectively. Niether form may be used for an input expecting a Non‐Null type.
Enum values are represented as unquoted names (ex. MOBILE_WEB
). It is recommended that Enum values be “all caps”. Enum values are only used in contexts where the precise enumeration type is known. Therefore it’s not necessary to supply an enumeration type name in the literal.
Lists are ordered sequences of values wrapped in square‐brackets [ ]
. The values of a List literal may be any value literal or variable (ex. [1, 2, 3]
).
Commas are optional throughout GraphQL so trailing commas are allowed and repeated commas do not represent missing values.
Semantics
Input object literal values are unordered lists of keyed input values wrapped in curly‐braces { }
. The values of an object literal may be any input value literal or variable (ex. { name: "Hello world", score: 1.0 }
). We refer to literal representation of input objects as “object literals.”
Input object fields are unordered
Input object fields may be provided in any syntactic order and maintain identical semantic meaning.
These two queries are semantically identical:
{
nearestThing(location: { lon: 12.43, lat: -53.211 })
}
{
nearestThing(location: { lat: -53.211, lon: 12.43 })
}
Semantics
A GraphQL query can be parameterized with variables, maximizing query reuse, and avoiding costly string building in clients at runtime.
If not defined as constant (for example, in DefaultValue), a Variable can be supplied for an input value.
Variables must be defined at the top of an operation and are in scope throughout the execution of that operation.
In this example, we want to fetch a profile picture size based on the size of a particular device:
query getZuckProfile($devicePicSize: Int) {
user(id: 4) {
id
name
profilePic(size: $devicePicSize)
}
}
Values for those variables are provided to a GraphQL service along with a request so they may be substituted during execution. If providing JSON for the variables’ values, we could run this query and request profilePic of size 60
width:
{
"devicePicSize": 60
}
Variable use within Fragments
Query variables can be used within fragments. Query variables have global scope with a given operation, so a variable used within a fragment must be declared in any top‐level operation that transitively consumes that fragment. If a variable is referenced in a fragment and is included by an operation that does not define that variable, the operation cannot be executed.
GraphQL describes the types of data expected by query variables. Input types may be lists of another input type, or a non‐null variant of any other input type.
Semantics
Directives provide a way to describe alternate runtime execution and type validation behavior in a GraphQL document.
In some cases, you need to provide options to alter GraphQL’s execution behavior in ways field arguments will not suffice, such as conditionally including or skipping a field. Directives provide this by describing additional information to the executor.
Directives have a name along with a list of arguments which may accept values of any input type.
Directives can be used to describe additional information for fields, fragments, and operations.
As future versions of GraphQL adopt new configurable execution capabilities, they may be exposed via directives.
The GraphQL Type system describes the capabilities of a GraphQL server and is used to determine if a query is valid. The type system also describes the input types of query variables to determine if values provided at runtime are valid.
A GraphQL server’s capabilities are referred to as that server’s “schema”. A schema is defined in terms of the types and directives it supports.
A given GraphQL schema must itself be internally valid. This section describes the rules for this validation process where relevant.
A GraphQL schema is represented by a root type for each kind of operation: query and mutation; this determines the place in the type system where those operations begin.
All types within a GraphQL schema must have unique names. No two provided types may have the same name. No provided type may have a name which conflicts with any built in types (including Scalar and Introspection types).
All directives within a GraphQL schema must have unique names. A directive and a type may share the same name, since there is no ambiguity between them.
The fundamental unit of any GraphQL Schema is the type. There are eight kinds of types in GraphQL.
The most basic type is a Scalar
. A scalar represents a primitive value, like a string or an integer. Oftentimes, the possible responses for a scalar field are enumerable. GraphQL offers an Enum
type in those cases, where the type specifies the space of valid responses.
Scalars and Enums form the leaves in response trees; the intermediate levels are Object
types, which define a set of fields, where each field is another type in the system, allowing the definition of arbitrary type hierarchies.
GraphQL supports two abstract types: interfaces and unions.
An Interface
defines a list of fields; Object
types that implement that interface are guaranteed to implement those fields. Whenever the type system claims it will return an interface, it will return a valid implementing type.
A Union
defines a list of possible types; similar to interfaces, whenever the type system claims a union will be returned, one of the possible types will be returned.
All of the types so far are assumed to be both nullable and singular: e.g. a scalar string returns either null or a singular string. The type system might want to define that it returns a list of other types; the List
type is provided for this reason, and wraps another type. Similarly, the Non-Null
type wraps another type, and denotes that the result will never be null. These two types are referred to as “wrapping types”; non‐wrapping types are referred to as “base types”. A wrapping type has an underlying “base type”, found by continually unwrapping the type until a base type is found.
Finally, oftentimes it is useful to provide complex structs as inputs to GraphQL queries; the Input Object
type allows the schema to define exactly what data is expected from the client in these queries.
As expected by the name, a scalar represents a primitive value in GraphQL. GraphQL responses take the form of a hierarchical tree; the leaves on these trees are GraphQL scalars.
All GraphQL scalars are representable as strings, though depending on the response format being used, there may be a more appropriate primitive for the given scalar type, and server should use those types when appropriate.
GraphQL provides a number of built‐in scalars, but type systems can add additional scalars with semantic meaning. For example, a GraphQL system could define a scalar called Time
which, while serialized as a string, promises to conform to ISO‐8601. When querying a field of type Time
, you can then rely on the ability to parse the result with an ISO‐8601 parser and use a client‐specific primitive for time. Another example of a potentially useful custom scalar is Url
, which serializes as a string, but is guaranteed by the server to be a valid URL.
A server may omit any of the built‐in scalars from its schema, for example if a schema does not refer to a floating‐point number, then it will not include the Float
type. However, if a schema includes a type with the name of one of the types described here, it must adhere to the behavior described. As an example, a server must not include a type called Int
and use it to represent 128‐bit numbers, or internationalization information.
Result Coercion
A GraphQL server, when preparing a field of a given scalar type, must uphold the contract the scalar type describes, either by coercing the value or producing an error.
For example, a GraphQL server could be preparing a field with the scalar type Int
and encounter a floating‐point number. Since the server must not break the contract by yielding a non‐integer, the server should truncate the fractional value and only yield the integer value. If the server encountered a boolean true
value, it should return 1
. If the server encountered a string, it may attempt to parse the string for a base‐10 integer value. If the server encounters some value that cannot be reasonably coerced to an Int
, then it must raise a field error.
Since this coercion behavior is not observable to clients of the GraphQL server, the precise rules of coercion are left to the implementation. The only requirement is that the server must yield values which adhere to the expected Scalar type.
Input Coercion
If a GraphQL server expects a scalar type as input to an argument, coercion is observable and the rules must be well defined. If an input value does not match a coercion rule, a query error must be raised.
GraphQL has different constant literals to represent integer and floating‐point input values, and coercion rules may apply differently depending on which type of input value is encountered. GraphQL may be parameterized by query variables, the values of which are often serialized when sent over a transport like HTTP. Since some common serializations (ex. JSON) do not discriminate between integer and floating‐point values, they are interpreted as an integer input value if they have an empty fractional part (ex. 1.0
) and otherwise as floating‐point input value.
For all types below, with the exception of Non‐Null, if the explicit value null is provided, then the result of input coercion is null.
Built‐in Scalars
GraphQL provides a basic set of well‐defined Scalar types. A GraphQL server should support all of these types, and a GraphQL server which provide a type by these names must adhere to the behavior described below.
The Int scalar type represents a signed 32‐bit numeric non‐fractional value. Response formats that support a 32‐bit integer or a number type should use that type to represent this scalar.
Result Coercion
GraphQL servers should coerce non‐int raw values to Int when possible otherwise they must raise a field error. Examples of this may include returning 1
for the floating‐point number 1.0
, or 2
for the string "2"
.
Input Coercion
When expected as an input type, only integer input values are accepted. All other input values, including strings with numeric content, must raise a query error indicating an incorrect type. If the integer input value represents a value less than -231 or greater than or equal to 231, a query error should be raised.
The Float scalar type represents signed double‐precision fractional values as specified by IEEE 754. Response formats that support an appropriate double‐precision number type should use that type to represent this scalar.
Result Coercion
GraphQL servers should coerce non‐floating‐point raw values to Float when possible otherwise they must raise a field error. Examples of this may include returning 1.0
for the integer number 1
, or 2.0
for the string "2"
.
Input Coercion
When expected as an input type, both integer and float input values are accepted. Integer input values are coerced to Float by adding an empty fractional part, for example 1.0
for the integer input value 1
. All other input values, including strings with numeric content, must raise a query error indicating an incorrect type. If the integer input value represents a value not representable by IEEE 754, a query error should be raised.
The String scalar type represents textual data, represented as UTF‐8 character sequences. The String type is most often used by GraphQL to represent free‐form human‐readable text. All response formats must support string representations, and that representation must be used here.
Result Coercion
GraphQL servers should coerce non‐string raw values to String when possible otherwise they must raise a field error. Examples of this may include returning the string "true"
for a boolean true value, or the string "1"
for the integer 1
.
Input Coercion
When expected as an input type, only valid UTF‐8 string input values are accepted. All other input values must raise a query error indicating an incorrect type.
The Boolean scalar type represents true
or false
. Response formats should use a built‐in boolean type if supported; otherwise, they should use their representation of the integers 1
and 0
.
Result Coercion
GraphQL servers should coerce non‐boolean raw values to Boolean when possible otherwise they must raise a field error. Examples of this may include returning true
for any non‐zero number.
Input Coercion
When expected as an input type, only boolean input values are accepted. All other input values must raise a query error indicating an incorrect type.
The ID scalar type represents a unique identifier, often used to refetch an object or as the key for a cache. The ID type is serialized in the same way as a String
; however, it is not intended to be human‐readable. While it is often numeric, it should always serialize as a String
.
Result Coercion
GraphQL is agnostic to ID format, and serializes to string to ensure consistency across many formats ID could represent, from small auto‐increment numbers, to large 128‐bit random numbers, to base64 encoded values, or string values of a format like GUID.
GraphQL servers should coerce as appropriate given the ID formats they expect. When coercion is not possible they must raise a field error.
Input Coercion
When expected as an input type, any string (such as "4"
) or integer (such as 4
) input value should be coerced to ID as appropriate for the ID formats a given GraphQL server expects. Any other input value, including float input values (such as 4.0
), must raise a query error indicating an incorrect type.
GraphQL queries are hierarchical and composed, describing a tree of information. While Scalar types describe the leaf values of these hierarchical queries, Objects describe the intermediate levels.
GraphQL Objects represent a list of named fields, each of which yield a value of a specific type. Object values should be serialized as ordered maps, where the queried field names (or aliases) are the keys and the result of evaluating the field is the value, ordered by the order in which they appear in the query.
For example, a type Person
could be described as:
type Person {
name: String
age: Int
picture: Url
}
Where name
is a field that will yield a String
value, and age
is a field that will yield an Int
value, and picture
is a field that will yield a Url
value.
A query of an object value must select at least one field. This selection of fields will yield an ordered map containing exactly the subset of the object queried, which should be represented in the order in which they were queried. Only fields that are declared on the object type may validly be queried on that object.
For example, selecting all the fields of Person
:
{
name
age
picture
}
Would yield the object:
{
"name": "Mark Zuckerberg",
"age": 30,
"picture": "http://some.cdn/picture.jpg"
}
While selecting a subset of fields:
{
age
name
}
Must only yield exactly that subset:
{
"age": 30,
"name": "Mark Zuckerberg"
}
A field of an Object type may be a Scalar, Enum, another Object type, an Interface, or a Union. Additionally, it may be any wrapping type whose underlying base type is one of those five.
For example, the Person
type might include a relationship
:
type Person {
name: String
age: Int
picture: Url
relationship: Person
}
Valid queries must supply a nested field set for a field that returns an object, so this query is not valid:
{
name
relationship
}
However, this example is valid:
{
name
relationship {
name
}
}
And will yield the subset of each object type queried:
{
"name": "Mark Zuckerberg",
"relationship": {
"name": "Priscilla Chan"
}
}
Field Ordering
When querying an Object, the resulting mapping of fields are conceptually ordered in the same order in which they were encountered during query execution, excluding fragments for which the type does not apply and fields or fragments that are skipped via @skip
or @include
directives. This ordering is correctly produced when using the CollectFields() algorithm.
Response serialization formats capable of representing ordered maps should maintain this ordering. Serialization formats which can only represent unordered maps should retain this order grammatically (such as JSON).
Producing a response where fields are represented in the same order in which they appear in the request improves human readability during debugging and enables more efficient parsing of responses if the order of properties can be anticipated.
If a fragment is spread before other fields, the fields that fragment specifies occur in the response before the following fields.
{
foo
...Frag
qux
}
fragment Frag on Query {
bar
baz
}
Produces the ordered result:
{
"foo": 1,
"bar": 2,
"baz": 3,
"qux": 4
}
If a field is queried multiple times in a selection, it is ordered by the first time it is encountered. However fragments for which the type does not apply does not affect ordering.
{
foo
...Ignored
...Matching
bar
}
fragment Ignored on UnknownType {
qux
baz
}
fragment Matching on Query {
bar
qux
foo
}
Produces the ordered result:
{
"foo": 1,
"bar": 2,
"qux": 3
}
Also, if directives result in fields being excluded, they are not considered in the ordering of fields.
{
foo @skip(if: true)
bar
foo
}
Produces the ordered result:
{
"bar": 1,
"foo": 2
}
Result Coercion
Determining the result of coercing an object is the heart of the GraphQL executor, so this is covered in that section of the spec.
Input Coercion
Objects are never valid inputs.
Object fields are conceptually functions which yield values. Occasionally object fields can accept arguments to further specify the return value. Object field arguments are defined as a list of all possible argument names and their expected input types.
For example, a Person
type with a picture
field could accept an argument to determine what size of an image to return.
type Person {
name: String
picture(size: Int): Url
}
GraphQL queries can optionally specify arguments to their fields to provide these arguments.
This example query:
{
name
picture(size: 600)
}
May yield the result:
{
"name": "Mark Zuckerberg",
"picture": "http://some.cdn/picture_600.jpg"
}
The type of an object field argument can be any Input type.
Fields in an object may be marked as deprecated as deemed necessary by the application. It is still legal to query for these fields (to ensure existing clients are not broken by the change), but the fields should be appropriately treated in documentation and tooling.
Object types have the potential to be invalid if incorrectly defined. This set of rules must be adhered to by every Object type in a GraphQL schema.
GraphQL Interfaces represent a list of named fields and their arguments. GraphQL objects can then implement an interface, which guarantees that they will contain the specified fields.
Fields on a GraphQL interface have the same rules as fields on a GraphQL object; their type can be Scalar, Object, Enum, Interface, or Union, or any wrapping type whose base type is one of those five.
For example, an interface may describe a required field and types such as Person
or Business
may then implement this interface.
interface NamedEntity {
name: String
}
type Person implements NamedEntity {
name: String
age: Int
}
type Business implements NamedEntity {
name: String
employeeCount: Int
}
Fields which yield an interface are useful when one of many Object types are expected, but some fields should be guaranteed.
To continue the example, a Contact
might refer to NamedEntity
.
type Contact {
entity: NamedEntity
phoneNumber: String
address: String
}
This allows us to write a query for a Contact
that can select the common fields.
{
entity {
name
}
phoneNumber
}
When querying for fields on an interface type, only those fields declared on the interface may be queried. In the above example, entity
returns a NamedEntity
, and name
is defined on NamedEntity
, so it is valid. However, the following would not be a valid query:
{
entity {
name
age
}
phoneNumber
}
because entity
refers to a NamedEntity
, and age
is not defined on that interface. Querying for age
is only valid when the result of entity
is a Person
; the query can express this using a fragment or an inline fragment:
{
entity {
name
... on Person {
age
}
},
phoneNumber
}
Result Coercion
The interface type should have some way of determining which object a given result corresponds to. Once it has done so, the result coercion of the interface is the same as the result coercion of the object.
Input Coercion
Interfaces are never valid inputs.
Interface types have the potential to be invalid if incorrectly defined.
GraphQL Unions represent an object that could be one of a list of GraphQL Object types, but provides for no guaranteed fields between those types. They also differ from interfaces in that Object types declare what interfaces they implement, but are not aware of what unions contain them.
With interfaces and objects, only those fields defined on the type can be queried directly; to query other fields on an interface, typed fragments must be used. This is the same as for unions, but unions do not define any fields, so no fields may be queried on this type without the use of typed fragments.
For example, we might have the following type system:
union SearchResult = Photo | Person
type Person {
name: String
age: Int
}
type Photo {
height: Int
width: Int
}
type SearchQuery {
firstSearchResult: SearchResult
}
When querying the firstSearchResult
field of type SearchQuery
, the query would ask for all fields inside of a fragment indicating the appropriate type. If the query wanted the name if the result was a Person, and the height if it was a photo, the following query is invalid, because the union itself defines no fields:
{
firstSearchResult {
name
height
}
}
Instead, the query would be:
{
firstSearchResult {
... on Person {
name
}
... on Photo {
height
}
}
}
Result Coercion
The union type should have some way of determining which object a given result corresponds to. Once it has done so, the result coercion of the union is the same as the result coercion of the object.
Input Coercion
Unions are never valid inputs.
Union types have the potential to be invalid if incorrectly defined.
GraphQL Enums are a variant on the Scalar type, which represents one of a finite set of possible values.
GraphQL Enums are not references for a numeric value, but are unique values in their own right. They serialize as a string: the name of the represented value.
Result Coercion
GraphQL servers must return one of the defined set of possible values. If a reasonable coercion is not possible they must raise a field error.
Input Coercion
GraphQL has a constant literal to represent enum input values. GraphQL string literals must not be accepted as an enum input and instead raise a query error.
Query variable transport serializations which have a different representation for non‐string symbolic values (for example, EDN) should only allow such values as enum input values. Otherwise, for most transport serializations that do not, strings may be interpreted as the enum input value with the same name.
Fields can define arguments that the client passes up with the query, to configure their behavior. These inputs can be Strings or Enums, but they sometimes need to be more complex than this.
The Object
type defined above is inappropriate for re‐use here, because Object
s can contain fields that express circular references or references to interfaces and unions, neither of which is appropriate for use as an input argument. For this reason, input objects have a separate type in the system.
An Input Object
defines a set of input fields; the input fields are either scalars, enums, or other input objects. This allows arguments to accept arbitrarily complex structs.
Result Coercion
An input object is never a valid result.
Input Coercion
The value for an input object should be an input object literal or an unordered map, otherwise an error should be thrown. This unordered map should not contain any entries with names not defined by a field of this input object type, otherwise an error should be thrown.
If any non‐nullable fields defined by the input object do not have corresponding entries in the original value, were provided a variable for which a value was not provided, or for which the value null was provided, an error should be thrown.
The result of coercion is an environment‐specific unordered map defining slots for each field both defined by the input object type and provided by the original value.
For each field of the input object type, if the original value has an entry with the same name, and the value at that entry is a literal value or a variable which was provided a runtime value, an entry is added to the result with the name of the field.
The value of that entry in the result is the outcome of input coercing the original entry value according to the input coercion rules of the type declared by the input field.
Following are examples of Input Object coercion for the type:
input ExampleInputObject {
a: String
b: Int!
}
Original Value | Variables | Coerced Value |
---|---|---|
{ a: "abc", b: 123 } | null | { a: "abc", b: 123 } |
{ a: 123, b: "123" } | null | { a: "123", b: 123 } |
{ a: "abc" } | null | Error: Missing required field b |
{ a: "abc", b: null } | null | Error: b must be non‐null. |
{ a: null, b: 1 } | null | { a: null, b: 1 } |
{ b: $var } | { var: 123 } | { b: 123 } |
{ b: $var } | {} | Error: Missing required field b. |
{ b: $var } | { var: null } | Error: b must be non‐null. |
{ a: $var, b: 1 } | { var: null } | { a: null, b: 1 } |
{ a: $var, b: 1 } | {} | { b: 1 } |
A GraphQL list is a special collection type which declares the type of each item in the List (referred to as the item type of the list). List values are serialized as ordered lists, where each item in the list is serialized as per the item type. To denote that a field uses a List type the item type is wrapped in square brackets like this: pets: [Pet]
.
Result Coercion
GraphQL servers must return an ordered list as the result of a list type. Each item in the list must be the result of a result coercion of the item type. If a reasonable coercion is not possible they must raise a field error. In particular, if a non‐list is returned, the coercion should fail, as this indicates a mismatch in expectations between the type system and the implementation.
Input Coercion
When expected as an input, list values are accepted only when each item in the list can be accepted by the list’s item type.
If the value passed as an input to a list type is not a list and not the null value, it should be coerced as though the input was a list of size one, where the value passed is the only item in the list. This is to allow inputs that accept a “var args” to declare their input type as a list; if only one argument is passed (a common case), the client can just pass that value rather than constructing the list.
By default, all types in GraphQL are nullable; the null value is a valid response for all of the above types. To declare a type that disallows null, the GraphQL Non‐Null type can be used. This type wraps an underlying type, and this type acts identically to that wrapped type, with the exception that null is not a valid response for the wrapping type. A trailing exclamation mark is used to denote a field that uses a Non‐Null type like this: name: String!
.
Nullable vs. Optional
Fields are always optional within the context of a query, a field may be omitted and the query is still valid. However fields that return Non‐Null types will never return the value null if queried.
Inputs (such as field arguments), are always optional by default. However a non‐null input type is required. In addition to not accepting the value null, it also does not accept omission. For the sake of simplicity nullable types are always optional and non‐null types are always required.
Result Coercion
In all of the above result coercions, null was considered a valid value. To coerce the result of a Non‐Null type, the coercion of the wrapped type should be performed. If that result was not null, then the result of coercing the Non‐Null type is that result. If that result was null, then a field error must be raised.
Input Coercion
If an argument or input‐object field of a Non‐Null type is not provided, is provided with the literal value null, or is provided with a variable that was either not provided a value at runtime, or was provided the value null, then a query error must be raised.
If the value provided to the Non‐Null type is provided with a literal value other than null, or a Non‐Null variable value, it is coerced using the input coercion for the wrapped type.
{
fieldWithNonNullArg
}
{
fieldWithNonNullArg(nonNullArg: null)
}
query withNullableVariable($var: String) {
fieldWithNonNullArg(nonNullArg: $var)
}
Non‐Null type validation
A GraphQL schema includes a list of the directives the execution engine supports.
GraphQL implementations should provide the @skip
and @include
directives.
The @skip
directive may be provided for fields, fragment spreads, and inline fragments, and allows for conditional exclusion during execution as described by the if argument.
In this example experimentalField
will be queried only if the $someTest
is provided a false
value.
query myQuery($someTest: Boolean) {
experimentalField @skip(if: $someTest)
}
The @include
directive may be provided for fields, fragment spreads, and inline fragments, and allows for conditional inclusion during execution as described by the if argument.
In this example experimentalField
will be queried only if the $someTest
is provided a true
value.
query myQuery($someTest: Boolean) {
experimentalField @include(if: $someTest)
}
@skip
nor @include
has precedence over the other. In the case that both the @skip
and @include
directives are provided in on the same the field or fragment, it must be queried only if the @skip
condition is false and the @include
condition is true. Stated conversely, the field or fragment must not be queried if either the @skip
condition is true or the @include
condition is false.A GraphQL schema includes types, indicating where query and mutation operations start. This provides the initial entry points into the type system. The query type must always be provided, and is an Object base type. The mutation type is optional; if it is null, that means the system does not support mutations. If it is provided, it must be an object base type.
The fields on the query type indicate what fields are available at the top level of a GraphQL query. For example, a basic GraphQL query like this one:
query getMe {
me
}
Is valid when the type provided for the query starting type has a field named “me”. Similarly
mutation setName {
setName(name: "Zuck") {
newName
}
}
Is valid when the type provided for the mutation starting type is not null, and has a field named “setName” with a string argument named “name”.
A GraphQL server supports introspection over its schema. This schema is queried using GraphQL itself, creating a powerful platform for tool‐building.
Take an example query for a trivial app. In this case there is a User type with three fields: id, name, and birthday.
For example, given a server with the following type definition:
type User {
id: String
name: String
birthday: Date
}
The query
{
__type(name: "User") {
name
fields {
name
type {
name
}
}
}
}
would return
{
"__type": {
"name": "User",
"fields": [
{
"name": "id",
"type": { "name": "String" }
},
{
"name": "name",
"type": { "name": "String" }
},
{
"name": "birthday",
"type": { "name": "Date" }
},
]
}
}
Types and fields required by the GraphQL introspection system that are used in the same context as user‐defined types and fields are prefixed with two underscores. This in order to avoid naming collisions with user‐defined GraphQL types. Conversely, GraphQL type system authors must not define any types, fields, arguments, or any other type system artifact with two leading underscores.
All types in the introspection system provide a description
field of type String
to allow type designers to publish documentation in addition to capabilities. A GraphQL server may return the description
field using Markdown syntax. Therefore it is recommended that any tool that displays description use a Markdown renderer.
To support the management of backwards compatibility, GraphQL fields and enum values can indicate whether or not they are deprecated (isDeprecated: Boolean
) and a description of why it is deprecated (deprecationReason: String
).
Tools built using GraphQL introspection should respect deprecation by discouraging deprecated use through information hiding or developer‐facing warnings.
GraphQL supports type name introspection at any point within a query by the meta field __typename: String!
when querying against any Object, Interface, or Union. It returns the name of the object type currently being queried.
This is most often used when querying against Interface or Union types to identify which actual type of the possible types has been returned.
This field is implicit and does not appear in the fields list in any defined type.
The schema introspection system is accessible from the meta‐fields __schema
and __type
which are accessible from the type of the root of a query operation.
__schema: __Schema!
__type(name: String!): __Type
These fields are implicit and do not appear in the fields list in the root type of the query operation.
The schema of the GraphQL schema introspection system:
type __Schema {
types: [__Type!]!
queryType: __Type!
mutationType: __Type
directives: [__Directive!]!
}
type __Type {
kind: __TypeKind!
name: String
description: String
# OBJECT and INTERFACE only
fields(includeDeprecated: Boolean = false): [__Field!]
# OBJECT only
interfaces: [__Type!]
# INTERFACE and UNION only
possibleTypes: [__Type!]
# ENUM only
enumValues(includeDeprecated: Boolean = false): [__EnumValue!]
# INPUT_OBJECT only
inputFields: [__InputValue!]
# NON_NULL and LIST only
ofType: __Type
}
type __Field {
name: String!
description: String
args: [__InputValue!]!
type: __Type!
isDeprecated: Boolean!
deprecationReason: String
}
type __InputValue {
name: String!
description: String
type: __Type!
defaultValue: String
}
type __EnumValue {
name: String!
description: String
isDeprecated: Boolean!
deprecationReason: String
}
enum __TypeKind {
SCALAR
OBJECT
INTERFACE
UNION
ENUM
INPUT_OBJECT
LIST
NON_NULL
}
type __Directive {
name: String!
description: String
locations: [__DirectiveLocation!]!
args: [__InputValue!]!
}
enum __DirectiveLocation {
QUERY
MUTATION
FIELD
FRAGMENT_DEFINITION
FRAGMENT_SPREAD
INLINE_FRAGMENT
}
__Type
is at the core of the type introspection system. It represents scalars, interfaces, object types, unions, enums in the system.
__Type
also represents type modifiers, which are used to modify a type that it refers to (ofType: __Type
). This is how we represent lists, non‐nullable types, and the combinations thereof.
There are several different kinds of type. In each kind, different fields are actually valid. These kinds are listed in the __TypeKind
enumeration.
Represents scalar types such as Int, String, and Boolean. Scalars cannot have fields.
A GraphQL type designer should describe the data format and scalar coercion rules in the description field of any scalar.
Fields
kind
must return __TypeKind.SCALAR
.name
must return a String.description
may return a String or null.Object types represent concrete instantiations of sets of fields. The introspection types (e.g. __Type
, __Field
, etc) are examples of objects.
Fields
kind
must return __TypeKind.OBJECT
.name
must return a String.description
may return a String or null.fields
: The set of fields query‐able on this type.includeDeprecated
which defaults to false. If true, deprecated fields are also returned.interfaces
: The set of interfaces that an object implements.Unions are an abstract type where no common fields are declared. The possible types of a union are explicitly listed out in possibleTypes
. Types can be made parts of unions without modification of that type.
Fields
kind
must return __TypeKind.UNION
.name
must return a String.description
may return a String or null.possibleTypes
returns the list of types that can be represented within this union. They must be object types.Interfaces are an abstract type where there are common fields declared. Any type that implements an interface must define all the fields with names and types exactly matching. The implementations of this interface are explicitly listed out in possibleTypes
.
Fields
kind
must return __TypeKind.INTERFACE
.name
must return a String.description
may return a String or null.fields
: The set of fields required by this interface.includeDeprecated
which defaults to false. If true, deprecated fields are also returned.possibleTypes
returns the list of types that implement this interface. They must be object types.Enums are special scalars that can only have a defined set of values.
Fields
kind
must return __TypeKind.ENUM
.name
must return a String.description
may return a String or null.enumValues
: The list of EnumValue
. There must be at least one and they must have unique names.includeDeprecated
which defaults to false. If true, deprecated enum values are also returned.Input objects are composite types used as inputs into queries defined as a list of named input values.
For example the input object Point
could be defined as:
type Point {
x: Int
y: Int
}
Fields
kind
must return __TypeKind.INPUT_OBJECT
.name
must return a String.description
may return a String or null.inputFields
: a list of InputValue
.Lists represent sequences of values in GraphQL. A List type is a type modifier: it wraps another type instance in the ofType
field, which defines the type of each item in the list.
Fields
kind
must return __TypeKind.LIST
.ofType
: Any type.GraphQL types are nullable. The value null is a valid response for field type.
A Non‐null type is a type modifier: it wraps another type instance in the ofType
field. Non‐null types do not allow null as a response, and indicate required inputs for arguments and input object fields.
kind
must return __TypeKind.NON_NULL
.ofType
: Any type except Non‐null.List and Non‐Null can compose, representing more complex types.
If the modified type of a List is Non‐Null, then that List may not contain any null items.
If the modified type of a Non‐Null is List, then null is not accepted, however an empty list is accepted.
If the modified type of a List is a List, then each item in the first List is another List of the second List’s type.
A Non‐Null type cannot modify another Non‐Null type.
The __Field
type represents each field in an Object or Interface type.
Fields
name
must return a Stringdescription
may return a String or nullargs
returns a List of __InputValue
representing the arguments this field accepts.type
must return a __Type
that represents the type of value returned by this field.isDeprecated
returns true if this field should no longer be used, otherwise false.deprecationReason
optionally provides a reason why this field is deprecated.The __InputValue
type represents field and directive arguments as well as the inputFields
of an input object.
Fields
name
must return a Stringdescription
may return a String or nulltype
must return a __Type
that represents the type this input value expects.defaultValue
may return a String encoding (using the GraphQL language) of the default value used by this input value in the condition a value is not provided at runtime. If this input value has no default value, returns null.The __Directive
type represents a Directive that a server supports.
Fields
name
must return a Stringdescription
may return a String or nulllocations
returns a List of __DirectiveLocation
representing the valid locations this directive may be placed.args
returns a List of __InputValue
representing the arguments this directive accepts. GraphQL does not just verify if a request is syntactically correct, but also ensures that it is unambiguous and mistake‐free in the context of a given GraphQL schema.
An invalid request is still technically executable, and will always produce a stable result as defined by the procedures in the Execution section, however that result may be ambiguous, surprising, or unexpected relative to the request containing validation errors, so execution should not occur for invalid requests.
Typically validation is performed in the context of a request immediately before execution, however a GraphQL service may execute a request without explicitly validating it if that exact same request is known to have been validated before. For example: the request may be validated during development, provided it does not later change, or a service may validate a request once and memoize the result to avoid validating the same request again in the future. Any client‐side or development‐time tool should report validation errors and not allow the formulation or execution of requests known to be invalid at that given point in time.
Type system evolution
As GraphQL type system schema evolve over time by adding new types and new fields, it is possible that a request which was previously valid could later become invalid. Any change that can cause a previously valid request to become invalid is considered a breaking change. GraphQL services and schema maintainers are encouraged to avoid breaking changes, however in order to be more resilient to these breaking changes, sophisticated GraphQL systems may still allow for the execution of requests which at some point were known to be free of any validation errors, and have not changed since.
Examples
For this section of this schema, we will assume the following type system in order to demonstrate examples:
enum DogCommand { SIT, DOWN, HEEL }
type Dog implements Pet {
name: String!
nickname: String
barkVolume: Int
doesKnowCommand(dogCommand: DogCommand!): Boolean!
isHousetrained(atOtherHomes: Boolean): Boolean!
owner: Human
}
interface Sentient {
name: String!
}
interface Pet {
name: String!
}
type Alien implements Sentient {
name: String!
homePlanet: String
}
type Human implements Sentient {
name: String!
}
enum CatCommand { JUMP }
type Cat implements Pet {
name: String!
nickname: String
doesKnowCommand(catCommand: CatCommand!): Boolean!
meowVolume: Int
}
union CatOrDog = Cat | Dog
union DogOrHuman = Dog | Human
union HumanOrAlien = Human | Alien
type QueryRoot {
dog: Dog
}
Formal Specification
Explanatory Text
Each named operation definition must be unique within a document when referred to by its name.
For example the following document is valid:
query getDogName {
dog {
name
}
}
query getOwnerName {
dog {
owner {
name
}
}
}
While this document is invalid:
query getName {
dog {
name
}
}
query getName {
dog {
owner {
name
}
}
}
It is invalid even if the type of each operation is different:
query dogOperation {
dog {
name
}
}
mutation dogOperation {
mutateDog {
id
}
}
Formal Specification
Explanatory Text
GraphQL allows a short‐hand form for defining query operations when only that one operation exists in the document.
For example the following document is valid:
{
dog {
name
}
}
While this document is invalid:
{
dog {
name
}
}
query getName {
dog {
owner {
name
}
}
}
Formal Specification
Explanatory Text
The target field of a field selection must be defined on the scoped type of the selection set. There are no limitations on alias names.
For example the following fragment would not pass validation:
fragment fieldNotDefined on Dog {
meowVolume
}
fragment aliasedLyingFieldTargetNotDefined on Dog {
barkVolume: kawVolume
}
For interfaces, direct field selection can only be done on fields. Fields of concrete implementors are not relevant to the validity of the given interface‐typed selection set.
For example, the following is valid:
fragment interfaceFieldSelection on Pet {
name
}
and the following is invalid:
fragment definedOnImplementorsButNotInterface on Pet {
nickname
}
Because unions do not define fields, fields may not be directly selected from a union‐typed selection set, with the exception of the meta‐field __typename. Fields from a union‐typed selection set must only be queried indirectly via a fragment.
For example the following is valid:
fragment inDirectFieldSelectionOnUnion on CatOrDog {
__typename
... on Pet {
name
}
... on Dog {
barkVolume
}
}
But the following is invalid:
fragment directFieldSelectionOnUnion on CatOrDog {
name
barkVolume
}
Formal Specification
Explanatory Text
If multiple field selections with the same response names are encountered during execution, the field and arguments to execute and the resulting value should be unambiguous. Therefore any two field selections which might both be encountered for the same object are only valid if they are equivalent.
For simple hand‐written GraphQL, this rule is obviously a clear developer error, however nested fragments can make this difficult to detect manually.
The following selections correctly merge:
fragment mergeIdenticalFields on Dog {
name
name
}
fragment mergeIdenticalAliasesAndFields on Dog {
otherName: name
otherName: name
}
The following is not able to merge:
fragment conflictingBecauseAlias on Dog {
name: nickname
name
}
Identical arguments are also merged if they have identical arguments. Both values and variables can be correctly merged.
For example the following correctly merge:
fragment mergeIdenticalFieldsWithIdenticalArgs on Dog {
doesKnowCommand(dogCommand: SIT)
doesKnowCommand(dogCommand: SIT)
}
fragment mergeIdenticalFieldsWithIdenticalValues on Dog {
doesKnowCommand(dogCommand: $dogCommand)
doesKnowCommand(dogCommand: $dogCommand)
}
The following do not correctly merge:
fragment conflictingArgsOnValues on Dog {
doesKnowCommand(dogCommand: SIT)
doesKnowCommand(dogCommand: HEEL)
}
fragment conflictingArgsValueAndVar on Dog {
doesKnowCommand(dogCommand: SIT)
doesKnowCommand(dogCommand: $dogCommand)
}
fragment conflictingArgsWithVars on Dog {
doesKnowCommand(dogCommand: $varOne)
doesKnowCommand(dogCommand: $varTwo)
}
fragment differingArgs on Dog {
doesKnowCommand(dogCommand: SIT)
doesKnowCommand
}
The following fields would not merge together, however both cannot be encountered against the same object, so they are safe:
fragment safeDifferingFields on Pet {
... on Dog {
volume: barkVolume
}
... on Cat {
volume: meowVolume
}
}
fragment safeDifferingArgs on Pet {
... on Dog {
doesKnowCommand(dogCommand: SIT)
}
... on Cat {
doesKnowCommand(catCommand: JUMP)
}
}
However, the field responses must be shapes which can be merged. For example, scalar values must not differ. In this example, someValue
might be a String
or an Int
:
fragment conflictingDifferingResponses on Pet {
... on Dog {
someValue: nickname
}
... on Cat {
someValue: meowVolume
}
}
Formal Specification
Explanatory Text
Field selections on scalars are never allowed: scalars are the leaf nodes of any GraphQL query.
The following is valid.
fragment scalarSelection on Dog {
barkVolume
}
The following is invalid.
fragment scalarSelectionsNotAllowedOnBoolean on Dog {
barkVolume {
sinceWhen
}
}
Conversely the leaf field selections of GraphQL queries must be scalars. Leaf selections on objects, interfaces, and unions without subfields are disallowed.
Let’s assume the following additions to the query root type of the schema:
extend type QueryRoot {
human: Human
pet: Pet
catOrDog: CatOrDog
}
The following examples are invalid
query directQueryOnObjectWithoutSubFields {
human
}
query directQueryOnInterfaceWithoutSubFields {
pet
}
query directQueryOnUnionWithoutSubFields {
catOrDog
}
Arguments are provided to both fields and directives. The following validation rules apply in both cases.
Formal Specification
Explanatory Text
Every argument provided to a field or directive must be defined in the set of possible arguments of that field or directive.
For example the following are valid:
fragment argOnRequiredArg on Dog {
doesKnowCommand(dogCommand: SIT)
}
fragment argOnOptional on Dog {
isHousetrained(atOtherHomes: true) @include(if: true)
}
the following is invalid since command
is not defined on DogCommand
.
fragment invalidArgName on Dog {
doesKnowCommand(command: CLEAN_UP_HOUSE)
}
and this is also invalid as unless
is not defined on @include
.
fragment invalidArgName on Dog {
isHousetrained(atOtherHomes: true) @include(unless: false)
}
In order to explore more complicated argument examples, let’s add the following to our type system:
type Arguments {
multipleReqs(x: Int!, y: Int!): Int!
booleanArgField(booleanArg: Boolean): Boolean
floatArgField(floatArg: Float): Float
intArgField(intArg: Int): Int
nonNullBooleanArgField(nonNullBooleanArg: Boolean!): Boolean!
booleanListArgField(booleanListArg: [Boolean]!): [Boolean]
}
extend type QueryRoot {
arguments: Arguments
}
Order does not matter in arguments. Therefore both the following example are valid.
fragment multipleArgs on Arguments {
multipleReqs(x: 1, y: 2)
}
fragment multipleArgsReverseOrder on Arguments {
multipleReqs(y: 1, x: 2)
}
Fields and directives treat arguments as a mapping of argument name to value. More than one argument with the same name in an argument set is ambiguous and invalid.
Formal Specification
Formal Specification
Explanatory Text
Literal values must be compatible with the type defined by the argument they are being provided to, as per the coercion rules defined in the Type System chapter.
For example, an Int can be coerced into a Float.
fragment goodBooleanArg on Arguments {
booleanArgField(booleanArg: true)
}
fragment coercedIntIntoFloatArg on Arguments {
floatArgField(floatArg: 1)
}
An incoercible conversion, is string to int. Therefore, the following example is invalid.
fragment stringIntoInt on Arguments {
intArgField(intArg: "3")
}
Explanatory Text
Arguments can be required. Arguments are required if the type of the argument is non‐null. If it is not non‐null, the argument is optional. When an argument type is non‐null, and is required, the explicit value null may also not be provided.
For example the following are valid:
fragment goodBooleanArg on Arguments {
booleanArgField(booleanArg: true)
}
fragment goodNonNullArg on Arguments {
nonNullBooleanArgField(nonNullBooleanArg: true)
}
The argument can be omitted from a field with a nullable argument.
Therefore the following query is valid:
fragment goodBooleanArgDefault on Arguments {
booleanArgField
}
but this is not valid on a non‐null argument.
fragment missingRequiredArg on Arguments {
nonNullBooleanArgField
}
Providing the explicit value null is also not valid.
fragment missingRequiredArg on Arguments {
notNullBooleanArgField(nonNullBooleanArg: null)
}
Formal Specification
Explanatory Text
Fragment definitions are referenced in fragment spreads by name. To avoid ambiguity, each fragment’s name must be unique within a document.
Inline fragments are not considered fragment definitions, and are unaffected by this validation rule.
For example the following document is valid:
{
dog {
...fragmentOne
...fragmentTwo
}
}
fragment fragmentOne on Dog {
name
}
fragment fragmentTwo on Dog {
owner {
name
}
}
While this document is invalid:
{
dog {
...fragmentOne
}
}
fragment fragmentOne on Dog {
name
}
fragment fragmentOne on Dog {
owner {
name
}
}
Formal Specification
Explanatory Text
Fragments must be specified on types that exist in the schema. This applies for both named and inline fragments. If they are not defined in the schema, the query does not validate.
For example the following fragments are valid:
fragment correctType on Dog {
name
}
fragment inlineFragment on Dog {
... on Dog {
name
}
}
fragment inlineFragment2 on Dog {
... @include(if: true) {
name
}
}
and the following do not validate:
fragment notOnExistingType on NotInSchema {
name
}
fragment inlineNotExistingType on Dog {
... on NotInSchema {
name
}
}
Formal Specification
Explanatory Text
Fragments can only be declared on unions, interfaces, and objects. They are invalid on scalars. They can only be applied on non‐leaf fields. This rule applies to both inline and named fragments.
The following fragment declarations are valid:
fragment fragOnObject on Dog {
name
}
fragment fragOnInterface on Pet {
name
}
fragment fragOnUnion on CatOrDog {
... on Dog {
name
}
}
and the following are invalid:
fragment fragOnScalar on Int {
something
}
fragment inlineFragOnScalar on Dog {
... on Boolean {
somethingElse
}
}
Formal Specification
Explanatory Text
Defined fragments must be used within a query document.
For example the following is an invalid query document:
fragment nameFragment on Dog { # unused
name
}
{
dog {
name
}
}
Field selection is also determined by spreading fragments into one another. The selection set of the target fragment is unioned with the selection set at the level at which the target fragment is referenced.
Formal Specification
Explanatory Text
Named fragment spreads must refer to fragments defined within the document. If the target of a spread is not defined, this is an error:
{
dog {
...undefinedFragment
}
}
Formal Specification
DetectCycles(fragmentDefinition, visited) :
Explanatory Text
The graph of fragment spreads must not form any cycles including spreading itself. Otherwise an operation could infinitely spread or infinitely execute on cycles in the underlying data.
This invalidates fragments that would result in an infinite spread:
{
dog {
...nameFragment
}
}
fragment nameFragment on Dog {
name
...barkVolumeFragment
}
fragment barkVolumeFragment on Dog {
barkVolume
...nameFragment
}
If the above fragments were inlined, this would result in the infinitely large:
{
dog {
name
barkVolume
name
barkVolume
name
barkVolume
name
# forever...
}
}
This also invalidates fragments that would result in an infinite recursion when executed against cyclic data:
{
dog {
...dogFragment
}
}
fragment dogFragment on Dog {
name
owner {
...ownerFragment
}
}
fragment ownerFragment on Dog {
name
pets {
...dogFragment
}
}
Formal Specification
Explanatory Text
Fragments are declared on a type and will only apply when the runtime object type matches the type condition. They also are spread within the context of a parent type. A fragment spread is only valid if its type condition could ever apply within the parent type.
In the scope of an object type, the only valid object type fragment spread is one that applies to the same type that is in scope.
For example
fragment dogFragment on Dog {
... on Dog {
barkVolume
}
}
and the following is invalid
fragment catInDogFragmentInvalid on Dog {
... on Cat {
meowVolume
}
}
In scope of an object type, unions or interface spreads can be used if the object type implements the interface or is a member of the union.
For example
fragment petNameFragment on Pet {
name
}
fragment interfaceWithinObjectFragment on Dog {
...petNameFragment
}
is valid because Dog implements Pet.
Likewise
fragment catOrDogNameFragment on CatOrDog {
... on Cat {
meowVolume
}
}
fragment unionWithObjectFragment on Dog {
...catOrDogNameFragment
}
is valid because Dog is a member of the CatOrDog union. It is worth noting that if one inspected the contents of the CatOrDogNameFragment you could note that no valid results would ever be returned. However we do not specify this as invalid because we only consider the fragment declaration, not its body.
Union or interface spreads can be used within the context of an object type fragment, but only if the object type is one of the possible types of that interface or union.
For example, the following fragments are valid:
fragment petFragment on Pet {
name
... on Dog {
barkVolume
}
}
fragment catOrDogFragment on CatOrDog {
... on Cat {
meowVolume
}
}
petFragment is valid because Dog implements the interface Pet. catOrDogFragment is valid because Cat is a member of the CatOrDog union.
By contrast the following fragments are invalid:
fragment sentientFragment on Sentient {
... on Dog {
barkVolume
}
}
fragment humanOrAlienFragment on HumanOrAlien {
... on Cat {
meowVolume
}
}
Dog does not implement the interface Sentient and therefore sentientFragment can never return meaningful results. Therefore the fragment is invalid. Likewise Cat is not a member of the union HumanOrAlien, and it can also never return meaningful results, making it invalid.
Union or interfaces fragments can be used within each other. As long as there exists at least one object type that exists in the intersection of the possible types of the scope and the spread, the spread is considered valid.
So for example
fragment unionWithInterface on Pet {
...dogOrHumanFragment
}
fragment dogOrHumanFragment on DogOrHuman {
... on Dog {
barkVolume
}
}
is consider valid because Dog implements interface Pet and is a member of DogOrHuman.
However
fragment nonIntersectingInterfaces on Pet {
...sentientFragment
}
fragment sentientFragment on Sentient {
name
}
is not valid because there exists no type that implements both Pet and Sentient.
Formal Specification
Explanatory Text
Input objects must not contain more than one field of the same name, otherwise an ambiguity would exist which includes an ignored portion of syntax.
For example the following query will not pass validation.
{
field(arg: { field: true, field: false })
}
Formal Specification
Explanatory Text
GraphQL servers define what directives they support. For each usage of a directive, the directive must be available on that server.
Formal Specification
Explanatory Text
GraphQL servers define what directives they support and where they support them. For each usage of a directive, the directive must be used in a location that the server has declared support for.
For example the following query will not pass validation because @skip
does not provide QUERY
as a valid location.
query @skip(if: $foo) {
field
}
Formal Specification
Explanatory Text
Directives are used to describe some metadata or behavioral change on the definition they apply to. When more than one directive of the same name is used, the expected metadata or behavior becomes ambiguous, therefore only one of each directive is allowed per location.
For example, the following query will not pass validation because @skip
has been used twice for the same field:
query ($foo: Boolean = true, $bar: Boolean = false) {
field @skip(if: $foo) @skip(if: $bar)
}
However the following example is valid because @skip
has been used only once per location, despite being used twice in the query and on the same named field:
query ($foo: Boolean = true, $bar: Boolean = false) {
field @skip(if: $foo) {
subfieldA
}
field @skip(if: $bar) {
subfieldB
}
}
Formal Specification
Explanatory Text
If any operation defines more than one variable with the same name, it is ambiguous and invalid. It is invalid even if the type of the duplicate variable is the same.
query houseTrainedQuery($atOtherHomes: Boolean, $atOtherHomes: Boolean) {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
It is valid for multiple operations to define a variable with the same name. If two operations reference the same fragment, it might actually be necessary:
query A($atOtherHomes: Boolean) {
...HouseTrainedFragment
}
query B($atOtherHomes: Boolean) {
...HouseTrainedFragment
}
fragment HouseTrainedFragment {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
Formal Specification
Explanatory Text
Variables defined by operations are allowed to define default values if the type of that variable is not non‐null.
For example the following query will pass validation.
query houseTrainedQuery($atOtherHomes: Boolean = true) {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
However if the variable is defined as non‐null, default values are unreachable. Therefore queries such as the following fail validation
query houseTrainedQuery($atOtherHomes: Boolean! = true) {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
Default values must be compatible with the types of variables. Types must match or they must be coercible to the type.
Non‐matching types fail, such as in the following example:
query houseTrainedQuery($atOtherHomes: Boolean = "true") {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
However if a type is coercible the query will pass validation.
For example:
query intToFloatQuery($floatVar: Float = 1) {
arguments {
floatArgField(floatArg: $floatVar)
}
}
Formal Specification
Explanatory Text
Variables can only be scalars, enums, input objects, or lists and non‐null variants of those types. These are known as input types. Objects, unions, and interfaces cannot be used as inputs.
For these examples, consider the following typesystem additions:
input ComplexInput { name: String, owner: String }
extend type QueryRoot {
findDog(complex: ComplexInput): Dog
booleanList(booleanListArg: [Boolean!]): Boolean
}
The following queries are valid:
query takesBoolean($atOtherHomes: Boolean) {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
query takesComplexInput($complexInput: ComplexInput) {
findDog(complex: $complexInput) {
name
}
}
query TakesListOfBooleanBang($booleans: [Boolean!]) {
booleanList(booleanListArg: $booleans)
}
The following queries are invalid:
query takesCat($cat: Cat) {
# ...
}
query takesDogBang($dog: Dog!) {
# ...
}
query takesListOfPet($pets: [Pet]) {
# ...
}
query takesCatOrDog($catOrDog: CatOrDog) {
# ...
}
Formal Specification
Explanatory Text
Variables are scoped on a per‐operation basis. That means that any variable used within the context of an operation must be defined at the top level of that operation
For example:
query variableIsDefined($atOtherHomes: Boolean) {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
is valid. $atOtherHomes is defined by the operation.
By contrast the following query is invalid:
query variableIsNotDefined {
dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
}
$atOtherHomes is not defined by the operation.
Fragments complicate this rule. Any fragment transitively included by an operation has access to the variables defined by that operation. Fragments can appear within multiple operations and therefore variable usages must correspond to variable definitions in all of those operations.
For example the following is valid:
query variableIsDefinedUsedInSingleFragment($atOtherHomes: Boolean) {
dog {
...isHousetrainedFragment
}
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
since isHousetrainedFragment is used within the context of the operation variableIsDefinedUsedInSingleFragment and the variable is defined by that operation.
On the other hand, if a fragment is included within an operation that does not define a referenced variable, the query is invalid.
query variableIsNotDefinedUsedInSingleFragment {
dog {
...isHousetrainedFragment
}
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
This applies transitively as well, so the following also fails:
query variableIsNotDefinedUsedInNestedFragment {
dog {
...outerHousetrainedFragment
}
}
fragment outerHousetrainedFragment on Dog {
...isHousetrainedFragment
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
Variables must be defined in all operations in which a fragment is used.
query housetrainedQueryOne($atOtherHomes: Boolean) {
dog {
...isHousetrainedFragment
}
}
query housetrainedQueryTwo($atOtherHomes: Boolean) {
dog {
...isHousetrainedFragment
}
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
However the following does not validate:
query housetrainedQueryOne($atOtherHomes: Boolean) {
dog {
...isHousetrainedFragment
}
}
query housetrainedQueryTwoNotDefined {
dog {
...isHousetrainedFragment
}
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
This is because housetrainedQueryTwoNotDefined does not define a variable $atOtherHomes but that variable is used by isHousetrainedFragment which is included in that operation.
Formal Specification
Explanatory Text
All variables defined by an operation must be used in that operation or a fragment transitively included by that operation. Unused variables cause a validation error.
For example the following is invalid:
query variableUnused($atOtherHomes: Boolean) {
dog {
isHousetrained
}
}
because $atOtherHomes is not referenced.
These rules apply to transitive fragment spreads as well:
query variableUsedInFragment($atOtherHomes: Boolean) {
dog {
...isHousetrainedFragment
}
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
The above is valid since $atOtherHomes is used in isHousetrainedFragment which is included by variableUsedInFragment.
If that fragment did not have a reference to $atOtherHomes it would be not valid:
query variableNotUsedWithinFragment($atOtherHomes: Boolean) {
...isHousetrainedWithoutVariableFragment
}
fragment isHousetrainedWithoutVariableFragment on Dog {
isHousetrained
}
All operations in a document must use all of their variables.
As a result, the following document does not validate.
query queryWithUsedVar($atOtherHomes: Boolean) {
dog {
...isHousetrainedFragment
}
}
query queryWithExtraVar($atOtherHomes: Boolean, $extra: Int) {
dog {
...isHousetrainedFragment
}
}
fragment isHousetrainedFragment on Dog {
isHousetrained(atOtherHomes: $atOtherHomes)
}
This document is not valid because queryWithExtraVar defines an extraneous variable.
Formal Specification
Explanatory Text
Variable usages must be compatible with the arguments they are passed to.
Validation failures occur when variables are used in the context of types that are complete mismatches, or if a nullable type in a variable is passed to a non‐null argument type.
Types must match:
query intCannotGoIntoBoolean($intArg: Int) {
arguments {
booleanArgField(booleanArg: $intArg)
}
}
$intArg typed as Int cannot be used as a argument to booleanArg, typed as Boolean.
List cardinality must also be the same. For example, lists cannot be passed into singular values.
query booleanListCannotGoIntoBoolean($booleanListArg: [Boolean]) {
arguments {
booleanArgField(booleanArg: $booleanListArg)
}
}
Nullability must also be respected. In general a nullable variable cannot be passed to a non‐null argument.
query booleanArgQuery($booleanArg: Boolean) {
arguments {
nonNullBooleanArgField(nonNullBooleanArg: $booleanArg)
}
}
A notable exception is when default arguments are provided. They are, in effect, treated as non‐nulls.
query booleanArgQueryWithDefault($booleanArg: Boolean = true) {
arguments {
nonNullBooleanArgField(nonNullBooleanArg: $booleanArg)
}
}
For list types, the same rules around nullability apply to both outer types and inner types. A nullable list cannot be passed to a non‐null list, and a list of nullable values cannot be passed to a list of non‐null values. The following is valid:
query nonNullListToList($nonNullBooleanList: [Boolean]!) {
arguments {
booleanListArgField(booleanListArg: $nonNullBooleanList)
}
}
However, a nullable list cannot be passed to a non‐null list:
query listToNonNullList($booleanList: [Boolean]) {
arguments {
nonNullBooleanListField(nonNullBooleanListArg: $booleanList)
}
}
This would fail validation because a [T]
cannot be passed to a [T]!
.
Similarly a [T]
cannot be passed to a [T!]
.
GraphQL generates a response from a request via execution.
A request for execution consists of a few pieces of information:
Given this information, the result of ExecuteRequest() produces the response, to be formatted according to the Response section below.
To execute a request, the executor must have a parsed Document
(as defined in the “Query Language” part of this spec) and a selected operation name to run if the document defines multiple operations, otherwise the document is expected to only contain a single operation. The result of the request is determined by the result of executing this operation according to the “Executing Operations” section below.
As explained in the Validation section, only requests which pass all validation rules should be executed. If validation errors are known, they should be reported in the list of “errors” in the response and the request must fail without execution.
Typically validation is performed in the context of a request immediately before execution, however a GraphQL service may execute a request without immediately validating it if that exact same request is known to have been validated before. A GraphQL service should only execute requests which at some point were known to be free of any validation errors, and have since not changed.
For example: the request may be validated during development, provided it does not later change, or a service may validate a request once and memoize the result to avoid validating the same request again in the future.
If the operation has defined any variables, then the values for those variables need to be coerced using the input coercion rules of variable’s declared type. If a query error is encountered during input coercion of variable values, then the operation fails without execution.
The type system, as described in the “Type System” section of the spec, must provide a query root object type. If mutations are supported, it must also provide a mutation root object type.
If the operation is a query, the result of the operation is the result of executing the query’s top level selection set with the query root object type.
An initial value may be provided when executing a query.
If the operation is a mutation, the result of the operation is the result of executing the mutation’s top level selection set on the mutation root object type. This selection set should be executed serially.
It is expected that the top level fields in a mutation operation perform side‐effects on the underlying data system. Serial execution of the provided mutations ensures against race conditions during these side‐effects.
To execute a selection set, the object value being evaluated and the object type need to be known, as well as whether it must be executed serially, or may be executed in parallel.
First, the selection set is turned into a grouped field set; then, each represented field in the grouped field set produces an entry into a response map.
Normally the executor can execute the entries in a grouped field set in whatever order it chooses (often in parallel). Because the resolution of fields other than top‐level mutation fields must always be side effect‐free and idempotent, the execution order must not affect the result, and hence the server has the freedom to execute the field entries in whatever order it deems optimal.
For example, given the following grouped field set to be executed normally:
{
birthday {
month
}
address {
street
}
}
A valid GraphQL executor can resolve the four fields in whatever order it chose (however of course birthday
must be resolved before month
, and address
before street
).
When executing a mutation, the selections in the top most selection set will be executed in serial order.
When executing a grouped field set serially, the executor must consider each entry from the grouped field set in the order provided in the grouped field set. It must determine the corresponding entry in the result map for each item to completion before it continues on to the next item in the grouped field set:
For example, given the following selection set to be executed serially:
{
changeBirthday(birthday: $newBirthday) {
month
}
changeAddress(address: $newAddress) {
street
}
}
The executor must, in serial:
changeBirthday
, which during CompleteValue() will execute the { month }
sub‐selection set normally.changeAddress
, which during CompleteValue() will execute the { street }
sub‐selection set normally.As an illustrative example, let’s assume we have a mutation field changeTheNumber
that returns an object containing one field, theNumber
. If we execute the following selection set serially:
{
first: changeTheNumber(newNumber: 1) {
theNumber
}
second: changeTheNumber(newNumber: 3) {
theNumber
}
third: changeTheNumber(newNumber: 2) {
theNumber
}
}
The executor will execute the following serially:
changeTheNumber(newNumber: 1)
field{ theNumber }
sub‐selection set of first
normallychangeTheNumber(newNumber: 3)
field{ theNumber }
sub‐selection set of second
normallychangeTheNumber(newNumber: 2)
field{ theNumber }
sub‐selection set of third
normallyA correct executor must generate the following result for that selection set:
{
"first": {
"theNumber": 1
},
"second": {
"theNumber": 3
},
"third": {
"theNumber": 2
}
}
Before execution, the selection set is converted to a grouped field set by calling CollectFields(). Each entry in the grouped field set is a list of fields that share a response key. This ensures all fields with the same response key (alias or field name) included via referenced fragments are executed at the same time.
As an example, collecting the fields of this selection set would collect two instances of the field a
and one of field b
:
{
a {
subfield1
}
...ExampleFragment
}
fragment ExampleFragment on Query {
a {
subfield2
}
b
}
The depth‐first‐search order of the field groups produced by CollectFields() is maintained through execution, ensuring that fields appear in the executed response in a stable and predictable order.
@skip
, let skipDirective be that directive.@include
, let includeDirective be that directive.Each field requested in the grouped field set that is defined on the selected objectType will result in an entry in the response map. Field execution first coerces any provided argument values, then resolves a value for the field, and finally completes that value either by recursively executing another selection set or coercing a scalar value.
Fields may include arguments which are provided to the underlying runtime in order to correctly produce a value. These arguments are defined by the field in the type system to have a specific input type: Scalars, Enum, Input Object, or List or Non‐Null wrapped variations of these three.
At each argument position in a query may be a literal value or a variable to be provided at runtime.
While nearly all of GraphQL execution can be described generically, ultimately the internal system exposing the GraphQL interface must provide values. This is exposed via ResolveFieldValue, which produces a value for a given field on a type for a real value.
As an example, this might accept the objectType Person
, the field "soulMate", and the objectValue representing John Lennon. It would be expected to yield the value representing Yoko Ono.
After resolving the value for a field, it is completed by ensuring it adheres to the expected return type. If the return type is another Object type, then the field execution process continues recursively.
Resolving Abstract Types
When completing a field with an abstract return type, that is an Interface or Union return type, first the abstract type must be resolved to a relevant Object type. This determination is made by the internal system using whatever means appropriate.
Merging Selection Sets
When more than one fields of the same name are executed in parallel, their selection sets are merged together when completing the value in order to continue execution of the sub‐selection sets.
An example query illustrating parallel fields with the same name with sub‐selections.
{
me {
firstName
}
me {
lastName
}
}
After resolving the value for me
, the selection sets are merged together so firstName
and lastName
can be resolved for one value.
If an error is thrown while resolving a field, it should be treated as though the field returned null, and an error must be added to the "errors" list in the response.
If the result of resolving a field is null (either because the function to resolve the field returned null or because an error occurred), and that field is of a Non-Null
type, then a field error is thrown. The error must be added to the "errors" list in the response.
If the field returns null because of an error which has already been added to the "errors" list in the response, the "errors" list must not be further affected. That is, only one error should be added to the errors list per field.
Since Non-Null
type fields cannot be null, field errors are propagated to be handled by the parent field. If the parent field may be null then it resolves to null, otherwise if it is a Non-Null
type, the field error is further propagated to it’s parent field.
If all fields from the root of the request to the source of the error return Non-Null
types, then the "data" entry in the response should be null.
When a GraphQL server receives a request, it must return a well‐formed response. The server’s response describes the result of executing the requested operation if successful, and describes any errors encountered during the request.
A response may contain both a partial response as well as encountered errors in the case that an error occurred on a field which was replaced with null.
GraphQL does not require a specific serialization format. However, clients should use a serialization format that supports the major primitives in the GraphQL response. In particular, the serialization format must support representations of the following four primitives:
Serialization formats which can represent an ordered map should preserve the order of requested fields as defined by CollectFields() in the Execution section. Serialization formats which can only represent unordered maps should retain this order grammatically (such as JSON).
Producing a response where fields are represented in the same order in which they appear in the request improves human readability during debugging and enables more efficient parsing of responses if the order of properties can be anticipated.
A serialization format may support the following primitives, however, strings may be used as a substitute for those primitives.
JSON is the preferred serialization format for GraphQL, though as noted above, GraphQL does not require a specific serialization format. For consistency and ease of notation, examples of the response are given in JSON throughout the spec. In particular, in our JSON examples, we will represent primitives using the following JSON concepts:
GraphQL Value | JSON Value |
---|---|
Map | Object |
List | Array |
Null | null |
String | String |
Boolean | true or false |
Int | Number |
Float | Number |
Enum Value | String |
Object Property Ordering
While JSON Objects are specified as an unordered collection of key‐value pairs the pairs are represented in an ordered manner. In other words, while the JSON strings { "name": "Mark", "age": 30 }
and { "age": 30, "name": "Mark" }
encode the same value, they also have observably different property orderings.
Since the result of evaluating a selection set is ordered, the JSON object serialized should preserve this order by writing the object properties in the same order as those fields were requested as defined by query execution.
For example, if the query was { name, age }
, a GraphQL server responding in JSON should respond with { "name": "Mark", "age": 30 }
and should not respond with { "age": 30, "name": "Mark" }
.
A response to a GraphQL operation must be a map.
If the operation included execution, the response map must contain a first entry with key data
. The value of this entry is described in the “Data” section. If the operation failed before execution, due to a syntax error, missing information, or validation error, this entry must not be present.
If the operation encountered any errors, the response map must contain a next entry with key errors
. The value of this entry is described in the “Errors” section. If the operation completed without encountering any errors, this entry must not be present.
The response map may also contain an entry with key extensions
. This entry, if set, must have a map as its value. This entry is reserved for implementors to extend the protocol however they see fit, and hence there are no additional restrictions on its contents.
To ensure future changes to the protocol do not break existing servers and clients, the top level response map must not contain any entries other than the three described above.
The data
entry in the response will be the result of the execution of the requested operation. If the operation was a query, this output will be an object of the schema’s query root type; if the operation was a mutation, this output will be an object of the schema’s mutation root type.
If an error was encountered before execution begins, the data
entry should not be present in the result.
If an error was encountered during the execution that prevented a valid response, the data
entry in the response should be null
.
The errors
entry in the response is a non‐empty list of errors, where each error is a map.
If no errors were encountered during the requested operation, the errors
entry should not be present in the result.
Every error must contain an entry with the key message
with a string description of the error intended for the developer as a guide to understand and correct the error.
If an error can be associated to a particular point in the requested GraphQL document, it should contain an entry with the key locations
with a list of locations, where each location is a map with the keys line
and column
, both positive numbers starting from 1
which describe the beginning of an associated syntax element.
GraphQL servers may provide additional entries to error as they choose to produce more helpful or machine‐readable errors, however future versions of the spec may describe additional entries to errors.
If the data
entry in the response is null
or not present, the errors
entry in the response must not be empty. It must contain at least one error. The errors it contains should indicate why no data was able to be returned.
If the data
entry in the response is not null
, the errors
entry in the response may contain any errors that occurred during execution. If errors occurred during execution, it should contain those errors.
This specification document contains a number of notation conventions used to describe technical concepts such as language grammar and semantics as well as runtime algorithms.
This appendix seeks to explain these notations in greater detail to avoid ambiguity.
A context‐free grammar consists of a number of productions. Each production has an abstract symbol called a “non‐terminal” as its left‐hand side, and zero or more possible sequences of non‐terminal symbols and or terminal characters as its right‐hand side.
Starting from a single goal non‐terminal symbol, a context‐free grammar describes a language: the set of possible sequences of characters that can be described by repeatedly replacing any non‐terminal in the goal sequence with one of the sequences it is defined by, until all non‐terminal symbols have been replaced by terminal characters.
Terminals are represented in this document in a monospace font in two forms: a specific Unicode character or sequence of Unicode characters (ex. = or terminal), and a pattern of Unicode characters defined by a regular expression (ex /[0-9]+/).
Non‐terminal production rules are represented in this document using the following notation for a non‐terminal with a single definition:
While using the following notation for a production with a list of definitions:
A definition may refer to itself, which describes repetitive sequences, for example:
The GraphQL language is defined in a syntactic grammar where terminal symbols are tokens. Tokens are defined in a lexical grammar which matches patterns of source characters. The result of parsing a sequence of source Unicode characters produces a GraphQL AST.
A Lexical grammar production describes non‐terminal “tokens” by patterns of terminal Unicode characters. No “whitespace” or other ignored characters may appear between any terminal Unicode characters in the lexical grammar production. A lexical grammar production is distinguished by a two colon ::
definition.
A Syntactical grammar production describes non‐terminal “rules” by patterns of terminal Tokens. Whitespace and other ignored characters may appear before or after any terminal Token. A syntactical grammar production is distinguished by a one colon :
definition.
This specification uses some additional notation to describe common patterns, such as optional or repeated patterns, or parameterized alterations of the definition of a non‐terminal. This section explains these short‐hand notations and their expanded definitions in the context‐free grammar.
Constraints
A grammar production may specify that certain expansions are not permitted by using the phrase “but not” and then indicating the expansions to be excluded.
For example, the production:
means that the nonterminal SafeName may be replaced by any sequence of characters that could replace Name provided that the same sequence of characters could not replace SevenCarlinWords.
A grammar may also list a number of restrictions after “but not” separated by “or”.
For example:
Optionality and Lists
A subscript suffix “Symbolopt” is shorthand for two possible sequences, one including that symbol and one excluding it.
As an example:
is shorthand for
A subscript suffix “Symbollist” is shorthand for a list of one or more of that symbol.
As an example:
is shorthand for
Parameterized Grammar Productions
A symbol definition subscript suffix parameter in braces “SymbolParam” is shorthand for two symbol definitions, one appended with that parameter name, the other without. The same subscript suffix on a symbol is shorthand for that variant of the definition. If the parameter starts with “?”, that form of the symbol is used if in a symbol definition with the same parameter. Some possible sequences can be included or excluded conditionally when respectively prefixed with “[+Param]” and “[~Param]”.
As an example:
is shorthand for
This specification describes the semantic value of many grammar productions in the form of a list of algorithmic steps.
For example, this describes how a parser should interpret a string literal:
This specification describes some algorithms used by the static and runtime semantics, they’re defined in the form of a function‐like syntax along with a list of algorithmic steps to take.
For example, this describes if a fragment should be spread into place given a runtime objectType and the fragment’s fragmentType:
! | $ | ( | ) | ... | : | = | @ | [ | ] | { | | | } |
0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
+ | - |
" | \ | / | b | f | n | r | t |
query | mutation |
true | false |