projection-patterns

You build read models by applying projection patterns that process event streams and materialize them into queryable structures. This supports CQRS workflows by separating write event stores from optimized read models for real-time analytics.

Projection types differ by timing: live projections update read models in real-time, catchup replays historical events to build state, and persistent projections maintain ongoing subscription checkpoints to ensure read models stay synchronized with event streams.

You create materialized views for dashboards by applying projection templates that aggregate multiple event streams into read models. These projections transform raw event data into structured formats suitable for real-time analytics and querying.

Yes, the projection architecture includes specific templates for Elasticsearch-backed search projections. These templates guide how event streams are projected into Elasticsearch indices to enable efficient querying and search capabilities.

Yes, you can use Python to implement event stream projections. The architecture provides concrete Python projection templates that wire your event streams to read models, facilitating the creation of CQRS workflows and materialized views.

Use inline projections when you need immediate read model updates during write operations, and use catchup projections when you need to replay historical event streams to rebuild or initialize read models from past events.