Skip to content

docs: add guide on directives #4401

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 4 commits into
base: 16.x.x
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
1 change: 1 addition & 0 deletions cspell.yml
Original file line number Diff line number Diff line change
Expand Up @@ -118,3 +118,4 @@ words:
- XXXF
- bfnrt
- wrds
- debuggable
1 change: 1 addition & 0 deletions website/pages/docs/_meta.ts
Original file line number Diff line number Diff line change
Expand Up @@ -26,6 +26,7 @@ const meta = {
'n1-dataloader': '',
'resolver-anatomy': '',
'graphql-errors': '',
'using-directives': '',
'-- 3': {
type: 'separator',
title: 'FAQ',
Expand Down
275 changes: 275 additions & 0 deletions website/pages/docs/using-directives.mdx
Original file line number Diff line number Diff line change
@@ -0,0 +1,275 @@
---
title: Using Directives in GraphQL.js
sidebarTitle: Using Directives
---

# Using Directives in GraphQL.js

Directives let you customize query execution at a fine-grained level. They act like
annotations in a GraphQL document, giving the server instructions about whether to
include a field, how to format a response, or how to apply custom behavior.

GraphQL.js supports built-in directives like `@include`, `@skip`, and `@deprecated` out
of the box. If you want to create your own directives and apply custom behavior, you'll
need to implement the logic yourself.

This guide covers how GraphQL.js handles built-in directives, how to define and apply
custom directives, and how to implement directive behavior during execution.

## How GraphQL.js handles built-in directives

GraphQL defines several built-in directives, each serving a specific purpose during
execution or in the schema. These include:

- `@include` and `@skip`: Used in the execution language to conditionally include or skip
fields and fragments at runtime.
- `@deprecated`: Used in Schema Definition Language (SDL) to mark fields or enum values as
deprecated, with an optional reason. It appears in introspection but doesn't affect query execution.

For example, the `@include` directive conditionally includes a field based on a Boolean variable:

```graphql
query($shouldInclude: Boolean!) {
greeting @include(if: $shouldInclude)
}
```

At runtime, GraphQL.js evaluates the `if` argument. If `shouldInclude` is `false`, the
`greeting` field in this example is skipped entirely and your resolver won't run.

```js
import { graphql, buildSchema } from 'graphql';

const schema = buildSchema(`
type Query {
greeting: String
}
`);

const rootValue = {
greeting: () => 'Hello!',
};

const query = `
query($shouldInclude: Boolean!) {
greeting @include(if: $shouldInclude)
}
`;

const variables = { shouldInclude: true };

const result = await graphql({
schema,
source: query,
rootValue,
variableValues: variables,
});

console.log(result);
// → { data: { greeting: 'Hello!' } }
```

If `shouldInclude` is `false`, the result would be `{ data: {} }`.

The `@deprecated` directive is used in the schema to indicate that a field or enum
value should no longer be used. It doesn't affect execution, but is included
in introspection output:

```graphql
{
__type(name: "MyType") {
fields {
name
isDeprecated
deprecationReason
}
}
}
```

GraphQL.js automatically includes deprecation metadata in introspection. Tools like
GraphiQL use this to show warnings, but GraphQL.js itself doesn't block or modify behavior.
You can still query deprecated fields unless you add validation rules yourself.

## Declaring custom directives in GraphQL.js

To use a custom directive, you first define it in your schema using the
`GraphQLDirective` class. This defines the directive's name, where it can
be applied, and any arguments it accepts.

A directive in GraphQL.js is just metadata. It doesn't perform any behavior on its own.

Here's a basic example that declares an `@uppercase` directive that can be applied to fields:

```js
import {
GraphQLDirective,
DirectiveLocation,
GraphQLNonNull,
GraphQLBoolean,
} from 'graphql';

const UppercaseDirective = new GraphQLDirective({
name: 'uppercase',
description: 'Converts the result of a field to uppercase.',
locations: [DirectiveLocation.FIELD],
args: {
enabled: {
type: GraphQLNonNull(GraphQLBoolean),
defaultValue: true,
description: 'Whether to apply the transformation.',
},
},
});
```

To make the directive available to your schema, you must explicitly include it:

```js
import { GraphQLSchema } from 'graphql';

const schema = new GraphQLSchema({
query: QueryType,
directives: [UppercaseDirective],
});
```

Once added, tools like validation and introspection will recognize it.

## Applying directives in queries

After defining and adding your directive to the schema, clients can apply it in queries using
the `@directiveName` syntax. Arguments are passed in parentheses, similar to field arguments.

You can apply directives to:

- Fields
- Fragment spreads
- Inline fragments

The following examples show how to apply directives:

```graphql
# Applied to a field
{
greeting @uppercase
}
```

```graphql
# Applied to a fragment spread
{
...userFields @include(if: true)
}
```

```graphql
# Applied to an inline fragment
{
... on User @skip(if: false) {
email
}
}
```

When a query is parsed, GraphQL.js includes directive nodes in the field's
Abstract Syntax Tree (AST). You can access these via `info.fieldNodes` inside
a resolver.

## Implementing custom directive behavior

GraphQL.js doesn't execute custom directive logic for you. You must handle it during
execution. There are two common approaches:

### 1. Handle directives in resolvers

Inside a resolver, use the `info` object to access AST nodes and inspect directives.
You can check whether a directive is present and change behavior accordingly.

```js
import {
graphql,
buildSchema,
getDirectiveValues,
} from 'graphql';

const schema = buildSchema(`
directive @uppercase(enabled: Boolean = true) on FIELD

type Query {
greeting: String
}
`);

const rootValue = {
greeting: (source, args, context, info) => {
const directive = getDirectiveValues(
schema.getDirective('uppercase'),
info.fieldNodes[0],
info.variableValues
);

const result = 'Hello, world';

if (directive?.enabled) {
return result.toUpperCase();
}

return result;
},
};

const query = `
query {
greeting @uppercase
}
`;

const result = await graphql({ schema, source: query, rootValue });
console.log(result);
// → { data: { greeting: 'HELLO, WORLD' } }
```

### 2. Use AST visitors or schema wrappers

For more complex logic, you can preprocess the schema or query using AST visitors or wrap
field resolvers. This lets you inject directive logic globally across
multiple types or fields.

This approach is useful for:

- Authorization
- Logging
- Schema transformations
- Feature flags

## Use cases for custom directives

Some common use cases for custom directives include:

- **Formatting**: `@uppercase`, `@date(format: "YYYY-MM-DD")`, `@currency`
- **Authorization**: `@auth(role: "admin")` to protect fields
- **Feature flags**: `@feature(name: "newHeader")` to expose experimental features
- **Observability**: `@log`, `@tag(name: "important")`, or `@metrics(id: "xyz")`
- **Execution control**: Mask or transform fields at runtime with schema visitors

## Best practices

When working with custom directives in GraphQL.js, keep the following best practices in mind:

- GraphQL.js doesn't have a directive middleware system. All custom directive logic must be implemented
manually.
- Weigh schema-driven logic against resolver logic. Directives can make queries more expressive, but they
may also hide behavior from developers reading the schema or resolvers.
- Keep directive behavior transparent and debuggable. Since directives are invisible at runtime unless
logged or documented, try to avoid magic behavior.
- Use directives when they offer real value. Avoid overusing directives to replace things that could be
handled more clearly in schema design or resolvers.
- Validate directive usage explicitly if needed. If your directive has rules or side effects, consider
writing custom validation rules to enforce correct usage.

## Additional resources

- [GraphQL Specification: Directives](https://spec.graphql.org/draft/#sec-Language.Directives)
- The Guild's guide on [Schema Directives](https://the-guild.dev/graphql/tools/docs/schema-directives)
- Apollo Server's guide on [Directives](https://www.apollographql.com/docs/apollo-server/schema/directives)
Loading