GraphQL

The human-readable syntax used to define the type system of a GraphQL API. SDL specifies object types, fields, relationships, and strict data shapes.

• Object Types: Define entities like Article or Author
• Scalar Types: Built-in primitives: String, Int, Float, Boolean, ID
• Non-Null: Appending ! to a type guarantees a value will always be returned
• Lists: Square brackets [Article] define one-to-many relationships

The schema acts as a contract between client and server, enabling static validation of queries before execution.

Functions on the server that populate data for each field in a schema. When a query requests specific fields, the resolver map executes the corresponding logic to fetch that data.

• Root Resolvers: Entry points for Query, Mutation, and Subscription operations
• Field-Level Resolvers: Granular functions that resolve individual object fields
• Default Resolvers: Automatically match field names to properties on the parent object
• Resolver Chains: Parent resolvers pass returned objects down to child field resolvers

Resolvers can fetch from databases, REST APIs, or other GraphQL services, enabling data federation.

Reusable units of field selection that prevent duplication in complex queries. Fragments let you define a set of fields once and spread them into multiple operations.

• Named Fragments: Defined with fragment Name on Type { ... } syntax
• Inline Fragments: Used directly in selections for conditional type-specific fields
• Fragment Composition: Fragments can reference other fragments for nested reuse

Essential for component-based UI frameworks like React, where each component declares its own data requirements via a co-located fragment.

A GraphQL operation type that establishes a persistent connection for real-time data push from server to client. Unlike queries and mutations, subscriptions maintain an open channel.

• Transport: Typically uses WebSockets or Server-Sent Events
• Event-Driven: Server pushes updates when specific events occur
• Filtering: Clients can subscribe with arguments to receive only relevant events
• Use Cases: Live chat, real-time dashboards, collaborative editing

Apollo Server and GraphQL-WS provide production-grade subscription implementations with connection management.

A built-in query capability that allows clients to interrogate the schema itself at runtime. By querying the __schema and __type meta-fields, tools can discover available types, fields, and directives.

• GraphiQL: The interactive IDE uses introspection for autocomplete and documentation
• Code Generation: Tools like GraphQL Code Generator read schemas to produce typed SDKs
• Security Note: Introspection is often disabled in production to prevent schema exposure

Introspection enables the rich developer experience that distinguishes GraphQL from REST APIs.

A utility library that batches and caches database requests within a single GraphQL query execution. It solves the N+1 problem where nested resolvers trigger redundant database round-trips.

• Batching: Collects individual loads during a single event-loop tick and dispatches them as one query
• Caching: Memoizes results per request so repeated loads for the same key return instantly
• Per-Request Scope: A new DataLoader instance is created for each incoming query to avoid stale data

Originally developed by Facebook, DataLoader is now a standard pattern in production GraphQL servers.