swiftdata-pro

SwiftData Pro

Model, query, and persist data with SwiftData correctly and efficiently.

When to use

• Defining @Model types and relationships.
• Writing @Query / fetching in views.
• Setting up ModelContainer, migrations, or CloudKit sync.

Trigger: /swiftdata-pro.

Core principles

• @Model classes are reference types managed by a ModelContext.
• Inject the container at the app root with .modelContainer(for:).
• Read in views with @Query; write through the modelContext.
• Keep the main-actor context for UI; use a background ModelContext for bulk work.

Modeling

@Model
final class Trip {
    var name: String
    var startDate: Date
    @Relationship(deleteRule: .cascade) var stops: [Stop] = []

    init(name: String, startDate: Date) {
        self.name = name
        self.startDate = startDate
    }
}

• Set an explicit deleteRule on relationships — don't rely on defaults for ownership.
• Use @Attribute(.unique) for natural keys.
• Mark large blobs @Attribute(.externalStorage).

❌ No delete rule on an owning relationship

var stops: [Stop] = []   // orphans Stop rows when a Trip is deleted

✅

@Relationship(deleteRule: .cascade) var stops: [Stop] = []

Container setup

@main
struct TripsApp: App {
    var body: some Scene {
        WindowGroup { ContentView() }
            .modelContainer(for: Trip.self)
    }
}

Querying

Use @Query with sort/filter in the view; don't fetch-all then filter in Swift.

❌

@Query private var trips: [Trip]
var upcoming: [Trip] { trips.filter { $0.startDate > .now } }   // loads everything

✅

@Query(filter: #Predicate<Trip> { $0.startDate > Date.now },
       sort: \Trip.startDate)
private var upcoming: [Trip]

Writing

@Environment(\.modelContext) private var context

func add(_ trip: Trip) {
    context.insert(trip)
    // SwiftData autosaves; call try? context.save() only when you need it now.
}

Delete with context.delete(trip).

Background work

For imports/bulk writes, use a separate context off the main actor and save in batches:

let context = ModelContext(container)
for row in rows { context.insert(Item(row)) }
try context.save()

Don't do thousands of inserts on the main-actor context — it blocks the UI.

Migrations

• Lightweight changes (adding optional properties) migrate automatically.
• Breaking changes need a SchemaMigrationPlan with versioned schemas and migration stages. Define VersionedSchema types; never silently mutate a shipped model.

CloudKit sync

• Every property must be optional or have a default; relationships must be optional.
• No @Attribute(.unique) (CloudKit can't enforce it).
• Configure with a ModelConfiguration using a CloudKit container identifier.

❌ (breaks CloudKit)

@Attribute(.unique) var code: String

✅

var code: String = ""

Common mistakes checklist

• Missing deleteRule on owning relationships.
• Filtering fetched arrays in Swift instead of #Predicate in @Query.
• Bulk inserts on the main-actor context.
• .unique or non-optional properties on a CloudKit-synced model.
• Breaking schema change with no migration plan.

Output format (when reviewing)

Per issue: file:line, rule, before/after. Lead with data-loss / migration risks.