Caution:

The library concept used in other language and projects.
BUT this golang library is an experimental implementation

Overview

es-go is an event sourcing library with simplied CQRS implementation

• Using Postgresql
• One database atomic transaction to store event and projection
• Event data as source of truth
• Considered database performance in mind
  • table indexing
  • projection for aggregate loader (load aggregate from another table BUT not aggregating from events for better performance)

Please check test/e2e for how it used

Problems to solve

• Current CQRS implementation mostly handle projection in another DB
  • event and projection are near real time consistence only
  • some use cases need strong consistence

For example,

es_go_local=# select * from events
        aggregate_id         | version |   event_type    |               payload               |         created_at       
  
-----------------------------+---------+-----------------+-------------------------------------+----------------------------
 2C11YkaaUdlHhAx0HRlhKIFk98u |       1 | created_event   | {"amount": 1.11, "currency": "BTC"} | 2022-07-16 15:08:46.831039
 2C11YkaaUdlHhAx0HRlhKIFk98u |       2 | completed_event | {"done_by": "marcusyip"}            | 2022-07-16 15:08:46.836044

projection

es_go_local=# select * from transaction_views
             id              | version |  status   | currency |        amount        |  done_by  |         created_at         |         updated_at         
-----------------------------+---------+-----------+----------+----------------------+-----------+----------------------------+----------------------------
 2C11YoSkPAuENcKtf9o8W1Hbb2R |       2 | completed | BTC      | 1.110000000000000000 | marcusyip | 2022-07-16 15:08:46.912425 | 2022-07-16 15:08:46.919962

Different data schema design aspects

Aspects	Examples
Diff based , Latest State	- Git Commit (Diff based) - Delta (Diff based - Rich Text Editing) - Database WAL (Diff based)
Data knowledge	Postgresql does not have knowledge on JSON or JSONB columns

es-go libraray

• the event data (Diff based data) as the source of truth. Transform and aggregate events to projection (Latest state data)
• Application System has the knowledge on data schema, but database does not have knowledge

How to run test

1. Run postgresql

docker-compose up

2. database schema migration

export POSTGRESQL_URL='postgres://postgres:postgres@localhost:5432/es_go_local?sslmode=disable'

migrate -database ${POSTGRESQL_URL} -path db/migrations up

go run test ./...

References

1. Mircosoft - apply-simplified-microservice-cqrs-ddd-patterns

• https://docs.microsoft.com/en-us/dotnet/architecture/microservices/microservice-ddd-cqrs-patterns/apply-simplified-microservice-cqrs-ddd-patterns