Struggling with query performance in event-sourced systems? This skill provides proven patterns for building efficient read models and projections from event streams.
下載技能 ZIP
在 Claude 中上傳
前往 設定 → 功能 → 技能 → 上傳技能
開啟並開始使用
測試它
正在使用「projection-patterns」。 User asks for a basic projector template
預期結果:
Provides a Python Projection base class with abstract methods for name, handles(), and apply(). Includes a Projector runner with checkpoint management and batch processing. Shows async/await patterns for database operations.
正在使用「projection-patterns」。 User needs an order summary projection
預期結果:
Delivers OrderSummaryProjection class handling OrderCreated, OrderItemAdded, OrderShipped events. Each handler updates specific columns. Demonstrates parameterized SQL queries and connection pooling.
正在使用「projection-patterns」。 User wants to rebuild a corrupted projection
預期結果:
Explains rebuild process: delete checkpoint, optionally clear read model tables, re-run projection from position 0. Emphasizes idempotency requirement for safe replay.
安全審計
安全This skill contains documentation and Python code examples for CQRS projection patterns. Static analyzer flagged markdown code blocks as shell commands (false positive). URLs are documentation references, not network calls. No executable code, cryptographic operations, or system access detected. Safe for publication.
品質評分
你能建構什麼
E-commerce Order Projections
Build read models that aggregate order events into customer-facing summaries with real-time status updates.
Search Index from Events
Create Elasticsearch projections that keep product catalogs in sync with domain events for full-text search.
Analytics Dashboard Data
Implement aggregating projections that compute daily metrics like sales totals and user activity.
試試這些提示
I need to build a read model for order events. Show me how to create a basic projector that handles OrderCreated and OrderUpdated events, stores checkpoints, and updates a SQL table.
Create a projection that syncs product domain events to Elasticsearch. Include handling for ProductCreated, ProductUpdated, and ProductDeleted events with proper error handling.
Show me how to build an aggregating projection that computes daily sales metrics from OrderCompleted events. Include UPSERT logic for efficient updates.
I need a projection that updates multiple tables atomically when processing CustomerOrderPlaced events. Show transaction handling and rollback patterns.
最佳實務
- Make projections idempotent - safe to replay events multiple times without side effects
- Store checkpoints after each event to enable resume after failures
- Use database transactions for multi-table updates to maintain consistency
避免
- Coupling projections together - each projection should be independent and isolatable
- Skipping error handling - unhandled exceptions stop projection and cause data lag
- Processing events out of order - always respect global_position ordering