dart-data-patterns

name	dart-data-patterns
description	Data handling patterns for Dart/Flutter. Use when implementing serialization, caching, layer separation, DTOs, repositories, offline-first, or data validation.

Data Patterns

Data Flow

• Repositories = single source of truth; reconcile local cache + remote; expose immutable data
• Abstract data sources (API, DB) via repository/service interfaces
• Consumers depend on abstractions, not implementations

Layer Boundaries

• Domain entities: no JSON, no SQL, no platform APIs
• DTOs encapsulate all mapping and JSON conversion; tables reference DTOs, not converters
• Data layer owns serialization/persistence; domain layer owns business rules

Serialization

• DTOs for data transfer between layers; separate from domain entities
• fromJson/toJson with explicit nullability and type handling
• Code generation (json_serializable, freezed) when appropriate
• copyWith for immutable updates; validate at boundaries before use

Streaming

• Stream<T> from repositories for reactive data (WebSocket, DB watch, polling)
• Broadcast streams when multiple listeners; handle errors and completion

Caching & Performance

• TTL-based caching with invalidation rules (on write, time, demand)
• Cursor-based or offset-based pagination for large datasets
• Isolates (compute()) for large payload parsing; minimize serialization in hot paths

Synchronization

• Sync strategy: last-write-wins, merge, or conflict resolution
• Optimistic updates: apply locally → reconcile with backend; roll back on failure
• Offline-first: serve from local cache, queue writes offline, flush when connected

Data Management

• Version schema; support upgrade paths; test migrations on real data
• Normalize where appropriate; denormalize for read performance
• Compress large payloads when beneficial (balance CPU vs bandwidth)

Security

• Encrypt at rest (flutter_secure_storage for credentials); HTTPS for network
• Support export/import for user data; handle corrupt data gracefully