Can I create a Core Data object graph in one application, then use it in another from a preset state?

Question

My app as it stands now has a .plist of around 5,000 dictionaries which, at runtime, I use to create 5,000 objects of my custom class. Each one contains a few strings, some boolean values, etc. This is a fairly quick process, takes a few seconds, tops, but it's a bit unwieldy using a .plist with thousands of objects.

I'm going to migrate to Core Data (which I've never used before), so the initial step is going to be getting the data from that .plist into those custom objects, then saving them in the context. I've watched the videos and such, and think I'm going to be okay doing that.

My first question is: can I create the object graph and data model in a small, separate app (say, called ObjectMaker), then save it there, and import the .sqlite or whatever format I export my data as into the actual app I'm making? From there, I'd be performing all of the fetch requests, further editing and saving of the data, etc.

My second question is: is it possible to provide a "first launch state" of an object graph that a user would get on first opening my app (with those 5,000 objects ready-made and ready to be accessed / queried, etc.? I'd like the app to ship with the .sqlite or whatever format I export my data to in the ObjectMaker app, and use that exclusively, without ever knowing that the .plist existed.

Thanks very much for any and all answers ^_^

Answer 1

We do this quite often with many of our apps. During development we build out the data model but often times need to import a bunch of master data and perform some default app configuration to get to an "intial state" of the sqlite database that we would ship the product with.

In the app delegate we have a #define (or variable) you can use to toggle " DB create & import" mode. When running the create mode we delete any existing .sqlite file, reinitialize the Core Data stack, then run the import routines to convert our .plist and .csv data into the required Core Data objects.

We use the simulator and build and run. Once we have the data imported and the default options of the app configured we navigate to the simulator folder that contains the app you are executing and copy the database and drop it back into the project file.

When the "DB create & import" mode is not enabled it first checks to see if there is a database at the proper location. If not, it copies the default database from the main bundle into the proper location on the device.

Answer 2

My first question is: can I create the object graph and data model in a small, separate app (say, called ObjectMaker), then save it there, and import the .sqlite or whatever format I export my data as into the actual app I'm making? From there, I'd be performing all of the fetch requests, further editing and saving of the data, etc.

Sure. You can create your sql store in your ObjectMaker app. Once done, you can take that store and copy (you have to do it yourself) in the main bundle of your app.

At startup, the app will copy the store into the main folder like the following. Persistent store has to point to that folder store.

NSString *storePath = [[self applicationDocumentsDirectory] 
    stringByAppendingPathComponent: @"yourStore.sqlite"];
NSURL *storeUrl = [NSURL fileURLWithPath:storePath];

NSFileManager *fileManager = [NSFileManager defaultManager];
if (![fileManager fileExistsAtPath:storePath]) {
    NSString *defaultStorePath = [[NSBundle mainBundle] 
        pathForResource:@"yourStore" ofType:@"sqlite"];
    if (defaultStorePath) {
        [fileManager copyItemAtPath:defaultStorePath toPath:storePath error:NULL];
    }
}

My second question is: is it possible to provide a "first launch state" of an object graph that a user would get on first opening my app (with those 5,000 objects ready-made and ready to be accessed / queried, etc.? I'd like the app to ship with the .sqlite or whatever format I export my data to in the ObjectMaker app, and use that exclusively, without ever knowing that the .plist existed.

About this question, I don't understand your goal but I'll try to provide some hints. First, I'm not very sure that a user will use 5,000 objects all in once. So you can display a bunch of them performing ad hoc queries (NSFetchRequest) to display them only when requested. In addition, if you need to display objects through tables I really suggest to play with NSFetchedResultsController. This class is there for this. It allows to have more control on data and let Core Data do memory management for you. For example, it allows to specify how many objects to retrieve (say 10). Scrolling the table, other 10 objects will fetched and so on...

Hope that helps. If you need something else, let me know.