Projections in Event Sourcing: Build ANY model you want!

Projections in Event Sourcing are a way to derive the current state from an event stream. This can be done asynchronously as events are persisted to an event stream which can update a projection. You don’t need to replay all the events in an event stream to get to the current state for a UI every time you need to display data. This could be very inefficient if you have a lot of events in a stream. Rather, create a projection that represents the current state and keep it updated as events occur.

YouTube

Check out my YouTube channel where I post all kinds of content that accompanies my posts including this video showing everything that is in this post.

Projections in Event Sourcing

Projections are a primary pattern you’ll likely use with Event Sourcing. It’s the answer to the most common question I get about event sourcing which is: How do you build out UI? If you have to replay all the events in an event stream to get to the current state, isn’t that very inefficient when it comes to displaying in a UI or reporting?

The answer, depending on your situation, could be yes. It could be very inefficient to have to replay an entire event stream to get to the current state.

This is where projections come in.

Data Transformation

Really a projection is transforming an event stream into another model. That other model could be almost anything depending on the events in your stream.

This is a stream of events for a specific product in a warehouse. Our event streams are per unique aggregate. In this case it’s for the product that’s identified with a SKU of ABC123

We can turn these series of events into a model that could be used for display purposes. The most obvious is probably to show users the current quantity on hand.

If we process these events we can derive these events into a current state that looks like this:

Our current state for the quantity is 59. If we process each event and keep track of the quantity (10 + 5 – 6 + 50) we would come to this final state.

The beauty of event sourcing is that you can create many different models. For example, we could also derive the event stream into this state:

Projections in Event Sourcing

In the above, we’re simply breaking out by keeping track of Received, Shipped, and Adjusted all separately.

As another example, we could keep track of product aging. Meaning how long is the oldest product in the warehouse from when it was received.

Event Consumers

Now before we get to actually building projections, you need to deliver/publish events to consumers that will process those events to update their projection of the current state.

There are a couple ways to accomplish this.

The first is to simply use a message broker that publishes events after they are saved to the event stream.

If you’re crossing a boundary, you likely don’t want to expose the event your persisting to your event stream. That would be leaking data internal to your domain. You’d likely want to transform that event into an integration event that you have versioning and contracts defined for.

The second option is if your consumers are within your boundary, and your database/event store supports it, is to use the event stream directly.

Products like EventStoreDB support two types of subscriptions: Persistent and Catch-up.

Subscriptions

Persistent subscriptions mean that you have competing consumers to a subscription group. As events occur, the event will be published to one consumer in the group that will process the message. This is similar to how you would use a message broker, except the event store is the broker. The event store keeps track of which subscription group has processed which message in the stream.

Catch-up subscriptions work a bit differently as the consumer must ask the event store to send events from a particular version onwards. The consumer must keep track of which message (version) of the stream it has processed. Once the consumer is caught up and processed all the messages that have occurred since the one it requested, the event store will send new messages to the consumer. Again, the consumer must keep track of which index/version it has processed because once it re-connects (for whatever reason) it needs to tell the event store where to start in the event stream.

Source Code

Developer-level members of my CodeOpinion YouTube channel get access to the full source in this post available in a git repo. Check out the membership for more info.

Building Projections

For my example, I’m using the first example I showed with a project that is keeping track of the current quantity for a product (by SKU).

I’m using Entity Framework and here’s my simple Entity and DbContext.

Every time a new event is appended to our event stream, it will publish this event which will be consumed by our ProjectionBuilder.ReceiveEvent()

ReceiveEvent() will determine which event type it is, then call the appropriate Apply() method. Each Apply() method fetches the appropriate record from our database, then updates the appropriate property/column. Then of course save the changes.

Now if we wanted to display the quantity on hand for a product, we would simply query DbContext by SKU and have to Subtract the Shipped from the Received amount. We do not need to go to the event store, reply to all the events to get to the current state. We already have it.

Demo App

I’ve created a simple console application that has all the code above in an interactive way. Developer-level members of my CodeOpinion YouTube channel get access to the full source and demo available in a git repo. Check out the membership for more info.

Follow @CodeOpinion on Twitter

Enjoy this post? Subscribe!

Subscribe to our weekly Newsletter and stay tuned.

Event Sourcing Example & Explained in plain English

What is Event Sourcing? It’s a way of storing data that is probably very different than what you’re used to. I’ll explain the differences and show ab event sourcing example that should clear up all the mystery around it.

YouTube

Check out my YouTube channel where I post all kinds of content that accompanies my posts including this video showing everything that is in this post.

Source Code

Developer-level members of my CodeOpinion YouTube channel get access to the full source in this post available in a git repo. Check out the membership for more info.

Current State

The vast majority of apps persist current state in a database. Regardless if you’re using a relational database (RDBMS) or a document store (NoSQL), you’re likely used to storing current state.

To illustrate this, here’s a table that represents products.

Event Sourcing

If we were to have some behavior in our system that is to receive product into the warehouse, we would increment the quantity value. So for example, if we received more quantity for SKU ABC123, we would update quantity value. If we shipped product out of the warehouse, we would decrease the quantity value.

One question to ask from the table above, how did we get to the current state of Quantity of 59 for product ABC123?

Because we only record current state, we have no way to know with absolutely certainty how we got to that number.

Yes, if added logging you could infer with some degree of certainty how you got to a particular state. However, it would not be guaranteed because it actually requires you to write logs in every place that you’re changing state, include outside of your application. This could be incredibly difficult. Ultimately your current state is the point of truth, no matter how you got to that state.

Event Sourcing

Event Sourcing is a different approach to storing data. Instead of storing the current state, you’re instead going to be storing events. Events represent the state transitions of things that have occurred in your system.

They are facts.

To illustrate the exact same product of SKU ABC123 that had a current state quantity of 59, this is how we would be storing this data using event sourcing.

It’s important to note that events are persisted in what is called an event stream. Event streams are generally per unique aggregate. So in my example, a single product SKU. The above is the stream of events for SKU ABC123.

With the above events, we can see that we Received a quantity of 10. Then we Received 5 more. Followed up by having Shipped out 6. Finally there some extra quantity that was magically found in the warehouse so it was adjusted by another 50. This got us to our current state of 59.

Event Sourcing Implementation

First is to define the events that occur that we want to record. Events are facts that something has occurred. They are generally the result of a state changes from commands. Here’ are the 3 events I’ve defined in our the event stream.

Next comes our aggregate. It is responsible for creating events that will get persisted to the event stream. The aggregate exposes methods to perform commands/actions to our domain. If our business logic passes, then we’ve confirmed that an event has occured.

When an event is added, we call the appropriate Apply() method. These methods are to keep track of the current state within our aggregate so we can perform the relevant business logic (which throws an InvalidDomainException).

Repository

For demo purposes, I’m not using an actual database to store our event stream, but rather just in-memory dictionary and list to illustrate. The important part is your repository is responsible two things, building your aggregate and saving the events from your aggregate.

When you want to build/get a WarehouseProduct from the Repository, it will get the events from the event stream, then call ApplyEvent() for each existing event. This is replaying all the events in the aggregate to get back to current state.

Then after you have called commands like ShipProduct/ReceiveProduct/AdjustInventory, the new events will get appended to the event stream from the Repositories Save() method.

Projections

The current state used in the aggregate is called a projection. It represents the current state of the event stream. I’ve covered more about projections and how they are used in UI and reporting to build many different models from your event stream.

Demo App

I’ve created a simple console application that has all the code above in an interactive way. Developer-level members of my CodeOpinion YouTube channel get access to the full source and demo available in a git repo. Check out the membership for more info.

Follow @CodeOpinion on Twitter

Enjoy this post? Subscribe!

Subscribe to our weekly Newsletter and stay tuned.

Aggregate (Root) Design: Behavior & Data

How do you persist your Aggregate using Entity Framework? Using a Repository to get/set your Aggregate Root? I see a lot of examples that bend over backward to make their Entities expose behaviors while try to encapsulate data. It doesn’t need to be difficult. And if you’re not using an ORM, what’s a way that you can capture the changes made by the Aggregate Root so that you can persist them writing SQL?

YouTube

Check out my YouTube channel where I post all kinds of content that accompanies my posts including this video showing everything that is in this post.

Entity Framework

The most common examples I see are creating an Aggregate using Entity Framework entities. In the very simplest of use-cases, I think it can work. However, I’ve watched/read enough content that makes me shake my head about the hoops people will jump through to make their data persistence model also be their domain model. This is often because they are trying to hide the actual data or conform to how EF needs the entity to behave.

They don’t need to be the same object.

Your domain model is about exposing behaviors and encapsulating data.

If you’re not exposing data from an aggregate, meaning there’s no way for callers/consumers to get data from it (think CQRS), then you can encapsulate your data model inside your aggregate root. Your data model at that point is your Entity Framework entities. Your aggregate root is your aggregate.

Separate Aggregate & Data

This is very simple but allows you to use EF exactly how it’s intended and create an Aggregate Root that simply exposes behaviors and encapsulate your data model (EF).

First, let’s start with our Entity Framework Entities, which are basically now just our data model. They contain no behavior. Simply a data model used for persistence.

Next, all of our behavior goes into a separate class that takes the ShoppingCart as a parameter in the ctor. We’ll manipulate this data model in all of our behavior methods. We don’t ever expose the data model to any outside caller/consumer.

Finally, we use a repository to Get and Save a ShopingCartDomain.

There are a lot of opinions about Repositories. My opinion is they should be for constructing and saving your behavior only aggregate. This generally means you’ll only ever have 2 methods: Get() and Save()

The repository will get the data using Entity Framework, then construct a new instance of our Aggregate and pass in the data model.

Aggregates (Root) without an ORM

If you’re not using an ORM like Entity Framework, but still want to create a domain with behavior then the ultimate problem is change tracking. With an ORM, it’s doing the change tracking of knowing which properties on your entities changed that it needs to persist to your database.

This is what needs to be implemented (change tracking) if you’re not using an ORM.

To do so, I like to implement this using events to represent state changes.

In our behavior methods, we’re recording and keeping track of them in the _events member. We’re also taking a parameter that represents the data model of the current state that our repository will build us.

When we want to save our state changes, our repository will iterate through the events, then have the appropriate SQL statements for them.

Expose Behavior and Encapsulate Data

If you’re doing all sorts of odd things to make your ORM be a data model that also contains behaviors, just separate the two. If you’re not using an ORM, it doesn’t mean you can’t create aggregates. Use events as a method of change tracking and implement the SQL statements for each event that represents a state change.

Links

Clean up your Domain Model with Event Sourcing

Follow @CodeOpinion on Twitter

Enjoy this post? Subscribe!

Subscribe to our weekly Newsletter and stay tuned.