Latest release: September 9, 2022 • version 3.0.0 • Release Notes

Requirements: iOS 11.0+ / macOS 10.13+ / tvOS 11.0+ / watchOS 4.0+ • Swift 5.7+ / Xcode 14+

To connect to the database, please refer to GRDB, the database library that supports RxGRDB.

// Single<[Player]>
let players = dbQueue.rx.read { db in
    try Player.fetchAll(db)
}

players.subscribe(
    onSuccess: { (players: [Player]) in
        print("Players: \(players)")
    },
    onError: { error in ... })

// Single<Void>
let write = dbQueue.rx.write { db in 
    try Player(...).insert(db)
}

write.subscribe(
    onSuccess: { _ in
        print("Updates completed")
    },
    onError: { error in ... })

// Single<Int>
let newPlayerCount = dbQueue.rx.write { db -> Int in
    try Player(...).insert(db)
    return try Player.fetchCount(db)
}

newPlayerCount.subscribe(
    onSuccess: { (playerCount: Int) in
        print("New players count: \(playerCount)")
    },
    onError: { error in ... })

// Observable<[Player]>
let observable = ValueObservation
    .tracking { db in try Player.fetchAll(db) }
    .rx.observe(in: dbQueue)

observable.subscribe(
    onNext: { (players: [Player]) in
        print("Fresh players: \(players)")
    },
    onError: { error in ... })

// Observable<Int?>
let observable = ValueObservation
    .tracking { db in try Int.fetchOne(db, sql: "SELECT MAX(score) FROM player") }
    .rx.observe(in: dbQueue)

observable.subscribe(
    onNext: { (maxScore: Int?) in
        print("Fresh maximum score: \(maxScore)")
    },
    onError: { error in ... })

// Observable<Database>
let observable = DatabaseRegionObservation
    .tracking(Player.all())
    .rx.changes(in: dbQueue)

observable.subscribe(
    onNext: { (db: Database) in
        print("Exclusive write access to the database after players have been impacted")
    },
    onError: { error in ... })

// Observable<Database>
let observable = DatabaseRegionObservation
    .tracking(SQLRequest<Int>(sql: "SELECT MAX(score) FROM player"))
    .rx.changes(in: dbQueue)

observable.subscribe(
    onNext: { (db: Database) in
        print("Exclusive write access to the database after maximum score has been impacted")
    },
    onError: { error in ... })

• Installation
• Demo Application
• Asynchronous Database Access
• Database Observation

To use RxGRDB with the Swift Package Manager, add a dependency to your Package.swift file:

let package = Package(
    dependencies: [
        .package(url: "https://github.com/RxSwiftCommunity/RxGRDB.git", ...)
    ]
)

To use RxGRDB with CocoaPods, specify in your Podfile:

# Pick only one
pod 'RxGRDB'
pod 'RxGRDB/SQLCipher'

RxGRDB provide observables that perform asynchronous database accesses.

• rx.read(observeOn:value:)
• rx.write(observeOn:updates:)
• rx.write(observeOn:updates:thenRead:)

This methods returns a Single that completes after database values have been asynchronously fetched.

// Single<[Player]>
let players = dbQueue.rx.read { db in
    try Player.fetchAll(db)
}

Any attempt at modifying the database completes subscriptions with an error.

When you use a database queue or a database snapshot, the read has to wait for any eventual concurrent database access performed by this queue or snapshot to complete.

When you use a database pool, reads are generally non-blocking, unless the maximum number of concurrent reads has been reached. In this case, a read has to wait for another read to complete. That maximum number can be configured.

This observable can be subscribed from any thread. A new database access starts on every subscription.

The fetched value is published on the main queue, unless you provide a specific scheduler to the observeOn argument.

This method returns a Single that completes after database updates have been successfully executed inside a database transaction.

// Single<Void>
let write = dbQueue.rx.write { db in
    try Player(...).insert(db)
}

// Single<Int>
let newPlayerCount = dbQueue.rx.write { db -> Int in
    try Player(...).insert(db)
    return try Player.fetchCount(db)
}

This observable can be subscribed from any thread. A new database access starts on every subscription.

It completes on the main queue, unless you provide a specific scheduler to the observeOn argument.

You can ignore its value and turn it into a Completable with the asCompletable operator:

// Completable
let write = dbQueue.rx
    .write { db in try Player(...).insert(db) }
    .asCompletable()

When you use a database pool, and your app executes some database updates followed by some slow fetches, you may profit from optimized scheduling with rx.write(observeOn:updates:thenRead:). See below.

This method returns a Single that completes after database updates have been successfully executed inside a database transaction, and values have been subsequently fetched:

// Single<Int>
let newPlayerCount = dbQueue.rx.write(
    updates: { db in try Player(...).insert(db) }
    thenRead: { db, _ in try Player.fetchCount(db) })
}

It publishes exactly the same values as rx.write(observeOn:updates:):

// Single<Int>
let newPlayerCount = dbQueue.rx.write { db -> Int in
    try Player(...).insert(db)
    return try Player.fetchCount(db)
}

The difference is that the last fetches are performed in the thenRead function. This function accepts two arguments: a readonly database connection, and the result of the updates function. This allows you to pass information from a function to the other (it is ignored in the sample code above).

When you use a database pool, this method applies a scheduling optimization: the thenRead function sees the database in the state left by the updates function, and yet does not block any concurrent writes. This can reduce database write contention. See Advanced DatabasePool for more information.

When you use a database queue, the results are guaranteed to be identical, but no scheduling optimization is applied.

This observable can be subscribed from any thread. A new database access starts on every subscription.

It completes on the main queue, unless you provide a specific scheduler to the observeOn argument.

Database Observation observables are based on GRDB's ValueObservation and DatabaseRegionObservation. Please refer to their documentation for more information. If your application needs change notifications that are not built in RxGRDB, check the general Database Changes Observation chapter.

• ValueObservation.rx.observe(in:scheduling:)
• DatabaseRegionObservation.rx.changes(in:)

GRDB's ValueObservation tracks changes in database values. You can turn it into an RxSwift observable:

let observation = ValueObservation.tracking { db in
    try Player.fetchAll(db)
}

// Observable<[Player]>
let observable = observation.rx.observe(in: dbQueue)

This observable has the same behavior as ValueObservation:

• It notifies an initial value before the eventual changes.

• It may coalesce subsequent changes into a single notification.

• It may notify consecutive identical values. You can filter out the undesired duplicates with the distinctUntilChanged() RxSwift operator, but we suggest you have a look at the removeDuplicates() GRDB operator also.

• It stops emitting any value after the database connection is closed. But it never completes.

• By default, it notifies the initial value, as well as eventual changes and errors, on the main thread, asynchronously.

  This can be configured with the scheduling argument. It does not accept an RxSwift scheduler, but a GRDB scheduler.

  For example, the .immediate scheduler makes sure the initial value is notified immediately when the observable is subscribed. It can help your application update the user interface without having to wait for any asynchronous notifications:

  // Immediate notification of the initial value
  let disposable = observation.rx
      .observe(
          in: dbQueue,
          scheduling: .immediate) // <-
      .subscribe(
          onNext: { players: [Player] in print("fresh players: \(players)") },
          onError: { error in ... })
  // <- here "fresh players" is already printed.

  Note that the .immediate scheduler requires that the observable is subscribed from the main thread. It raises a fatal error otherwise.

See ValueObservation Scheduling for more information.

⚠️ ValueObservation and Data Consistency

When you compose ValueObservation observables together with the combineLatest operator, you lose all guarantees of data consistency.

Instead, compose requests together into one single ValueObservation, as below:

// DATA CONSISTENCY GUARANTEED
let hallOfFameObservable = ValueObservation
    .tracking { db -> HallOfFame in
        let playerCount = try Player.fetchCount(db)
        let bestPlayers = try Player.limit(10).orderedByScore().fetchAll(db)
        return HallOfFame(playerCount:playerCount, bestPlayers:bestPlayers)
    }
    .rx.observe(in: dbQueue)

See ValueObservation for more information.

GRDB's DatabaseRegionObservation notifies all transactions that impact a tracked database region. You can turn it into an RxSwift observable:

let request = Player.all()
let observation = DatabaseRegionObservation.tracking(request)

// Observable<Database>
let observable = observation.rx.changes(in: dbQueue)

This observable can be created and subscribed from any thread. It delivers database connections in a "protected dispatch queue", serialized with all database updates. It only completes when a database error happens.

let request = Player.all()
let disposable = DatabaseRegionObservation
    .tracking(request)
    .rx.changes(in: dbQueue)
    .subscribe(
        onNext: { (db: Database) in
            print("Players have changed.")
        },
        onError: { error in ... })

try dbQueue.write { db in
    try Player(name: "Arthur").insert(db)
    try Player(name: "Barbara").insert(db)
} 
// Prints "Players have changed."

try dbQueue.write { db in
    try Player.deleteAll(db)
}
// Prints "Players have changed."

See DatabaseRegionObservation for more information.