core-data-pro

Core Data Pro

Use Core Data correctly: safe contexts, efficient fetches, clean migrations. (For new apps, also consider SwiftData — see swiftdata-pro.)

When to use

• Existing Core Data stacks, or apps needing fine-grained control.
• Fixing threading crashes, slow fetches, or migration failures.

Trigger: /core-data-pro.

Core principles

• The view context is main-queue only; never touch it off the main thread.
• Do writes/imports on a background context, then merge.
• NSManagedObjects belong to their context — don't pass them across threads, pass IDs.
• Fetch only what you need (predicates, limits, batching).

Stack setup

let container = NSPersistentContainer(name: "Model")
container.loadPersistentStores { _, error in
    if let error { fatalError("Store load failed: \(error)") }
}
container.viewContext.automaticallyMergesChangesFromParent = true

Threading

❌ Background work on the view context (crashes / corruption)

DispatchQueue.global().async {
    let obj = Item(context: container.viewContext)   // wrong queue
}

✅ Background context with perform

container.performBackgroundTask { context in
    let obj = Item(context: context)
    obj.title = "New"
    try? context.save()   // merges into viewContext automatically
}

Pass object IDs across contexts, not objects:

let id = obj.objectID
container.performBackgroundTask { ctx in
    let bgObj = ctx.object(with: id)
}

Fetching efficiently

❌ Fetch everything, filter in Swift

let all = try context.fetch(Item.fetchRequest())
let recent = all.filter { $0.date > cutoff }     // loads the whole table

✅ Predicate + sort + limit in the request

let req = Item.fetchRequest()
req.predicate = NSPredicate(format: "date > %@", cutoff as NSDate)
req.sortDescriptors = [NSSortDescriptor(key: "date", ascending: false)]
req.fetchLimit = 50
let recent = try context.fetch(req)

Use fetchBatchSize for large lists; NSFetchedResultsController (UIKit) or @FetchRequest (SwiftUI) for table/list binding.