I would wrap the UserSettings type inside an observable class and then use that class as an environment value @Environment(SettingsManager.self) var settingsManager Then this manager object could be made to handle all loading, updating and saving of the UserSettings object internally to have only one single source of truth.

One option is to create and work with a local instance of ModelContext in your view and then call save() or rollback (or simply discard it) on that local instance. Once you call save the changes will be propagated to the main context. One article about this can be found on https://www.hackingwithswift.com/quick-start/swiftdata/how-to-discard-changes-to-a-swiftdata-object or as part of this longer article, https://medium.com/@matgnt/the-art-of-swiftdata-in-2025-from-scattered-pieces-to-a-masterpiece-1fd0cefd8d87

When you are using @ModelActor you have access to a specific model context for that actor via the property modelContext so you should have try modelContext.delete(… and you should definitely not use the context from the main actor in any other actor. This alone will probably not fix the crash since you have a reference to the deleted objects in self._rooms that you need to clear before doing the delete.

Do you get any error message in the console when you save? Could you replace try? with try when saving to see if any error is thrown do { try modelContext.save() } catch { print(error) } How is your model container declared and used?

I find the question a bit confusing. Firstly, are you talking about saving and fetching objects in general or is it the specific image property of type Data you mean when you write “data” in the question? Secondly it’s very hard to understand the last segment of code.

And we are just supposed to know what ETL means?

What do you mean with "more verbose way"? It seems to be the same command but one is wrapped in a generic function.

Here is a good Apple tutorial on handling errors in a view.

You should read the answers to your previous question carefully and also see this question on stackoverflow.com on using optionals with a picker

The selection parameter for the picker should be a single item and not an array @State private var contractType: TypeOfContract = // some default type also the type of the value passed to .tag() must be the same type as the property above. So this issue is not so much about SwiftData but instead understanding how a picker works.

My guess is that a view is trying to access an instance of the object after it has been removed.

That’s a completely different issue and it’s not something you checked either in the code you have provided

You don’t need to unwrap anything so just skip the if let self.project.subcategories?.removeAll { $0 == subcategory} calling a method on something that is nil in swift is perfectly fine

MKMapPoint doesn’t conform to Codable so you can’t use it directly in your model. One solution is to store the coordinate properties in the model instead.

An array is a value type in itself even if the elements are reference types. So if you do var myArray = project.subcategories ?? [] then any change to myArray will not affect project.subcategories but if you make a change to an object in myArray (a Subproject object?) then that object in the project.subcategories array will also be updated. So I guess the short answer is that if you want to add or remove objects to the project you need to work with project.subcategories I can't really help you with the rest because I don't really understand what you are saying in the paragraph after the last code snippet.