Desk joins in Fluent 4

Database fashions

Fluent is a Swift ORM framework written for Vapor. You should use fashions to signify rows in a desk, migrations to create the construction for the tables and you may outline relations between the fashions utilizing Swift property wrappers. That is fairly a easy means of representing guardian, youngster or sibling connections. You possibly can “keen load” fashions by way of these predefined relation properties, which is nice, however typically you do not need to have static sorts for the relationships.

I am engaged on a modular CMS and I am unable to have hardcoded relationship properties contained in the fashions. Why? Nicely, I would like to have the ability to load modules at runtime, so if module A relies upon from module B by way of a relation property then I am unable to compile module A independently. That is why I dropped many of the cross-module relations, however I’ve to put in writing joined queries. 😅

Buyer mannequin

On this instance we’re going to mannequin a easy Buyer-Order-Product relation. Our buyer mannequin could have a primary identifier and a reputation. Take into account the next:

last class CustomerModel: Mannequin, Content material {
    static let schema = "prospects"
    
    @ID(key: .id) var id: UUID?
    @Discipline(key: "title") var title: String

    init() { }

    init(id: UUID? = nil, title: String) {
        self.id = id
        self.title = title
    }
}

Nothing particular, only a primary Fluent mannequin.

Order mannequin

Prospects could have a one-to-many relationship to the orders. Which means a buyer can have a number of orders, however an order will at all times have precisely one related buyer.

last class OrderModel: Mannequin, Content material {
    static let schema = "orders"
    
    @ID(key: .id) var id: UUID?
    @Discipline(key: "date") var date: Date
    @Discipline(key: "customer_id") var customerId: UUID

    init() { }

    init(id: UUID? = nil, date: Date, customerId: UUID) {
        self.id = id
        self.date = date
        self.customerId = customerId
    }
}

We might make the most of the @Guardian and @Baby property wrappers, however this time we’re going to retailer a customerId reference as a UUID kind. In a while we’re going to put a international key constraint on this relation to make sure that referenced objects are legitimate identifiers.

Product mannequin

The product mannequin, similar to the client mannequin, is completely impartial from anything. 📦

last class ProductModel: Mannequin, Content material {
    static let schema = "merchandise"
    
    @ID(key: .id) var id: UUID?
    @Discipline(key: "title") var title: String

    init() { }

    init(id: UUID? = nil, title: String) {
        self.id = id
        self.title = title
    }
}

We are able to create a property with a @Sibling wrapper to specific the connection between the orders and the merchandise, or use joins to question the required information. It actually would not matter which means we go, we nonetheless want a cross desk to retailer the associated product and order identifiers.

OrderProductModel

We are able to describe a many-to-many relation between two tables utilizing a 3rd desk.

last class OrderProductModel: Mannequin, Content material {
    static let schema = "order_products"
    
    @ID(key: .id) var id: UUID?
    @Discipline(key: "order_id") var orderId: UUID
    @Discipline(key: "product_id") var productId: UUID
    @Discipline(key: "amount") var amount: Int

    init() { }

    init(id: UUID? = nil, orderId: UUID, productId: UUID, amount: Int) {
        self.id = id
        self.orderId = orderId
        self.productId = productId
        self.amount = amount
    }
}

As you possibly can see we will retailer additional information on the cross desk, in our case we’re going to affiliate portions to the merchandise on this relation proper subsequent to the product identifier.

Migrations

Happily, Fluent offers us a easy method to create the schema for the database tables.

struct InitialMigration: Migration {

    func put together(on db: Database) -> EventLoopFuture<Void> {
        db.eventLoop.flatten([
            db.schema(CustomerModel.schema)
                .id()
                .field("name", .string, .required)
                .create(),
            db.schema(OrderModel.schema)
                .id()
                .field("date", .date, .required)
                .field("customer_id", .uuid, .required)
                .foreignKey("customer_id", references: CustomerModel.schema, .id, onDelete: .cascade)
                .create(),
            db.schema(ProductModel.schema)
                .id()
                .field("name", .string, .required)
                .create(),
            db.schema(OrderProductModel.schema)
                .id()
                .field("order_id", .uuid, .required)
                .foreignKey("order_id", references: OrderModel.schema, .id, onDelete: .cascade)
                .field("product_id", .uuid, .required)
                .foreignKey("product_id", references: ProductModel.schema, .id, onDelete: .cascade)
                .field("quantity", .int, .required)
                .unique(on: "order_id", "product_id")
                .create(),
        ])
    }

    func revert(on db: Database) -> EventLoopFuture<Void> {
        db.eventLoop.flatten([
            db.schema(OrderProductModel.schema).delete(),
            db.schema(CustomerModel.schema).delete(),
            db.schema(OrderModel.schema).delete(),
            db.schema(ProductModel.schema).delete(),
        ])
    }
}

If you wish to keep away from invalid information within the tables, you need to at all times use the international key and distinctive constraints. A international key can be utilized to examine if the referenced identifier exists within the associated desk and the distinctive constraint will be sure that just one row can exists from a given area.

Becoming a member of database tables utilizing Fluent 4

We’ve got to run the InitialMigration script earlier than we begin utilizing the database. This may be carried out by passing a command argument to the backend software or we will obtain the identical factor by calling the autoMigrate() technique on the applying occasion.

For the sake of simplicity I will use the wait technique as a substitute of async Futures & Guarantees, that is high quality for demo functions, however in a real-world server software you need to by no means block the present occasion loop with the wait technique.

That is one doable setup of our dummy database utilizing an SQLite storage, however in fact you need to use PostgreSQL, MySQL and even MariaDB by way of the accessible Fluent SQL drivers. 🚙

public func configure(_ app: Utility) throws {

    app.databases.use(.sqlite(.file("db.sqlite")), as: .sqlite)

    app.migrations.add(InitialMigration())

    attempt app.autoMigrate().wait()

    let prospects = [
        CustomerModel(name: "Bender"),
        CustomerModel(name: "Fry"),
        CustomerModel(name: "Leela"),
        CustomerModel(name: "Hermes"),
        CustomerModel(name: "Zoidberg"),
    ]
    attempt prospects.create(on: app.db).wait()
    
    let merchandise = [
        ProductModel(name: "Hamburger"),
        ProductModel(name: "Fish"),
        ProductModel(name: "Pizza"),
        ProductModel(name: "Beer"),
    ]
    attempt merchandise.create(on: app.db).wait()

    
    let order = OrderModel(date: Date(), customerId: prospects[0].id!)
    attempt order.create(on: app.db).wait()

    let beerProduct = OrderProductModel(orderId: order.id!, productId: merchandise[3].id!, amount: 6)
    attempt beerProduct.create(on: app.db).wait()
    let pizzaProduct = OrderProductModel(orderId: order.id!, productId: merchandise[2].id!, amount: 1)
    attempt pizzaProduct.create(on: app.db).wait()
}

We’ve got created 5 prospects (Bender, Fry, Leela, Hermes, Zoidberg), 4 merchandise (Hamburger, Fish, Pizza, Beer) and one new order for Bender containing 2 merchandise (6 beers and 1 pizza). 🤖

Interior be a part of utilizing one-to-many relations

Now the query is: how can we get the client information based mostly on the order?

let orders = attempt OrderModel
    .question(on: app.db)
    .be a part of(CustomerModel.self, on: OrderModel.$customerId == CustomerModel.$id, technique: .inside)
    .all()
    .wait()

for order in orders {
    let buyer = attempt order.joined(CustomerModel.self)
    print(buyer.title)
    print(order.date)
}

The reply is fairly easy. We are able to use an inside be a part of to fetch the client mannequin by way of the order.customerId and buyer.id relation. After we iterate by way of the fashions we will ask for the associated mannequin utilizing the joined technique.

Joins and plenty of to many relations

Having a buyer is nice, however how can I fetch the related merchandise for the order? We are able to begin the question with the OrderProductModel and use a be a part of utilizing the ProductModel plus we will filter by the order id utilizing the present order.

for order in orders {
    

    let orderProducts = attempt OrderProductModel
        .question(on: app.db)
        .be a part of(ProductModel.self, on: OrderProductModel.$productId == ProductModel.$id, technique: .inside)
        .filter(.$orderId == order.id!)
        .all()
        .wait()

    for orderProduct in orderProducts {
        let product = attempt orderProduct.joined(ProductModel.self)
        print(product.title)
        print(orderProduct.amount)
    }
}

We are able to request the joined mannequin the identical means as we did it for the client. Once more, the very first parameter is the mannequin illustration of the joined desk, subsequent you outline the relation between the tables utilizing the referenced identifiers. As a final parameter you possibly can specify the kind of the be a part of.

Interior be a part of vs left be a part of

There’s a nice SQL tutorial about joins on w3schools.com, I extremely suggest studying it. The primary distinction between an inside be a part of and a left be a part of is that an inside be a part of solely returns these information which have matching identifiers in each tables, however a left be a part of will return all of the information from the bottom (left) desk even when there aren’t any matches within the joined (proper) desk.

There are a lot of several types of SQL joins, however inside and left be a part of are the most typical ones. If you wish to know extra in regards to the different sorts you need to learn the linked article. 👍

Abstract

Desk joins are actually helpful, however you must watch out with them. You need to at all times use correct international key and distinctive constraints. Additionally think about using indexes on some rows whenever you work with joins, as a result of it may possibly enhance the efficiency of your queries. Pace will be an essential issue, so by no means load extra information from the database than you really want.

There is a matter on GitHub in regards to the Fluent 4 API, and one other one about querying particular fields utilizing the .area technique. Lengthy story brief, joins will be nice and we want higher docs. 🙉