Before I proceed, I'm trying to understand all the issues around core data sync with icloud. Can anyone tell me or point me to some documentation as to what happens when the data model changes. For instance I send out a new release of my app which requires a migration to a new data model version. A user downloads the new release on their iPhone - but not their iPad. What happens at that point?
Syncing only happens between devices that have the same model version. If they update the app on one device but not another, syncing will stop until they update their other device. Once they update the other device, syncing is supposed to resume. I'm not completely certain about how reliable this is in practice, but that's what's supposed to happen.
Apple just released a new guide to using Core Data with iCloud that you might find useful.
Related
I'm currently learning how to use CloudKit Framework and lack of documentation or examples showing how to sync Core Data and CloudKit.
I have watched all WWDC videos (2014, 2015, 2016) Dedicated to CloudKit, but none of them telling us how to implement syncing with Core Data. I can't find any fresh examples, tutorial or books, showing how to implement this syncing.
I know that it is effective to use Operations API by CloudKit (not Convenience API) and to Subscribe to changes as it said in the new WWDC 2016 videos, dedicated to CloudKit, but mapping with CoreData is a real problem.
For example, let's say I would like to create an app similar to Notes app. while offline, user can create his notes and work with them saving them to his core data database. When the device going online the app checks what changed on the server and saves newly created records to server (CloudKit).
When the app starts, it also fetches for changes from the CloudKit and if there are changes , it updates local cache (Core Data) with the new changes.
I would appreciate to have a common pattern of syncing. Where to place syncing with Core Data methods and how they should look like?
Would appreciate any information or help about this.
I'm using Swift 3, Xcode 8 , iOS 10.
As of iOS 13, there are new API's that simplify this synchronization for developers. I would recommend you to watch the WWDC19 session about the new synchronization between CoreData and CloudKit. Please note that these new API's only work for iOS 13+.
Video: https://developer.apple.com/videos/play/wwdc2019/202/
In short, you need to start using NSPersistentCloudKitContainer instead of NSPersistentContainer. This will let the syncing work automatically using automatic conflict resolution with a last-writer-wins merge strategy. If you want to build a good working app, you'll also need to do some modifications to improve the syncing for your app.
Official documentation can be found at:
Setting Up Core Data with CloudKit
Syncing a Core Data Store with CloudKit
Data modeling for collaboration (conflict-free replicated data type)
At the end of the session they also demonstrated an example of better sync merging than the default 'last-writer-wins merge strategy'. The usage of Causal Trees allow multiple users to edit the same string (and to some extend other types of data) without losing any data. I would really recommend everyone to read this article from Archagon that describes how this works and how to implement it (also with CloudKit syncing, but not the new automatic one). As demonstrated in the session, you can also implement this with the new automatic syncing between CoreData and CloudKit.
Core Data already provides the user with the ability to sync to iCloud. There's no need to use CloudKit.
Design For Core Data In iCloud
But yes, Core Data with iCloud has been deprecated. Even so, it has not been discontinued, and there are no immediate plans at apple to discontinue it, they just want to discourage its use. But it also has problems with rationalising updates from multiple devices.
In any case, I have been looking into the question of how to do this with cloud kit myself. Two answers; the first is to use the following;
Seam in GitHub
The second is to do it manually;
Designing for CloudKit
The key here is that Cloud Kit needs the record metadata to be able to handle record updates reliably, so you have to save that metadata in your Core Data database. The CKRecord class includes a method encodeSystemFields(with:) which will encode those fields into a Data record that can be stored in your database, and then your can use the appropriate decoder when you need to restore the CKRecord.
Anyway, I am about to start doing this myself. I'll update this with more information when I have it.
Apple has recently published a guide that seems to answer this question. Check out Apple's Maintaining a Local Cache of CloudKit Records to see how to store CloudKit data on device.
While this guide doesn't provide sample code to write to the device, it does answer the rest of the question. This tells you how to fetch changes from CloudKit and create data which can be stored on the device.
I have a coredata app that I would love to be able to share the same data with multiple devices, possibly with iCloud/cloud kit. I am not sure where to start, or how to go about it? The only thing I can think of, but still not sure how to do, would be to sync the SQLite files with iCloud? Not sure if thats a good idea or not? I just recently converted my app over to swift 3 and iOS10 core data code. The only way I am able to share data between devices currently is thru iTunes files sharing.
For whatever reason this topic is hard to find modern info on.
Core Data doesn't have support for this. Except for the built-in iCloud sync, but that's deprecated as of iOS 10.
You could use CloudKit to sync data, but you'll have to write your own code to convert between Core Data's persistent store and CloudKit's online store. It's not impossible but it's certainly not automatic.
Syncing the SQLite file is not a good idea unless you really want to corrupt the data.
I have an app with multiple updates on the AppStore already, funny thing happened, I thought that the lightweight migration happens automatically, however, my recent discovery that I need to add the
NSDictionary *storeOptions = #{NSMigratePersistentStoresAutomaticallyOption:#YES, NSInferMappingModelAutomaticallyOption:#YES};
to my persistentStoreCoordinator shook my confidence when I realized I already have 5 core data models.
The question is: when I add the above line to the next version of the app, is it going to work for everyone when they update? Because right now everything that happens when they open the app .. is a fancy CRASH.
Thx
It will work if automatic lightweight migration is possible for the migration you're trying to perform. Whether this will work depends on the differences between the model used for the existing data and the current version of the model. Many common changes permit automatic lightweight migration but not all. You'll need to review the docs on this kind of migration and decide whether it will work in your case.
If it doesn't work, there are other ways to handle it, for example by creating a mapping file to tell Core Data how to make changes that it can't infer automatically.
What is the best way to check iCloud for existing data?
I need to check that data doesn't exist on the local device, or iCloud so I can then download it.
Since you included the core-data tag I'm assuming you mean that you're using Core Data rather than iCloud file APIs or the ubiquitous key-value store.
With Core Data's built-in iCloud support, you check on existing data in exactly the same way as if you were not using iCloud. Once you create your Core Data stack, you check what data exists by doing normal Core Data fetches. There's no (exposed) concept of local vs. cloud data, there's just one data store that happens to know how to communicate with iCloud. You don't explicitly initiate downloads-- they happen automatically.
At app launch time when you call addPersistentStoreWithType:configuration:URL:options:error:, Core Data internally initiates the download of any data that's not available locally yet. As a result this method may block for a while. If it returns successfully, then all current downloads can be assumed to be complete.
If new changes appear while your app is running, Core Data will download and import them and, when it's finished, will post NSPersistentStoreDidImportUbiquitousContentChangesNotification to tell you what just happened.
This all describes how Core Data's iCloud is supposed to work. In practice you'll probably find that it doesn't always work as intended.
Thanks to #Tom Harrington for pointing out, this error is nothing to do with the developer/coding - it's purely to do with iCloud/Apple/connection issues.
More on this SO answer I found.
I recently submitted an upgrade of my app which included a lightweight coredata migration (including new fields in existing tables and a couple of new tables). I followed every tip regarding this migration, including some I found on this site.
I thoroughly tested the update on three different devices and it all went ok!!!
However, this update is crashing an all my devices and probably on all my customers. I can't explain why this is happening.
Could you please help me understand this debacle?
To truly test your app and migration, you need to run your original app to create data store according to the original data model. Then you need to run your new app, opening data store that was generated with original app. This can be a real pain and is easier (at least initially) to do in Simulator because you have more control over the file system and can swap in a saved original data store. On iDevice you need to regenerate original data store for each test.
If you are testing on your own development devices then you have already migrated your data store. Is it possible that your test devices created their data stores with new data model - and never actually performed a migration?
I only generally use automatic migration during beta testing, for quick revisions, other than that I always use a mapping model, so that you have control.
the other issue is that if your model shifts far enough between releases, auto migration from v1-v2 could be fine, and v2-v3 could be ok, but v1-v3 could be too drastic to be inferred. by making maps for them, you retain control of the migration.