I have an NSManagedObject that has a to-many relationship to another NSManagedObject.
During creation of the NSManagedObject I can use the generated accessors 'removeNotesObject' and the deletion works fine. I can create an object to add to the parent object, save the object, delete the object and then save again. When I fetch this parent object the object I created and deleted is still deleted.
However, after I add the object and then save it (but don't delete and save after) and then fetch it, I can't seem to delete the object that was previously created. I am using the generated accessors to try and remove the object, which appears to work but when I fetch it again the object hasn't been deleted.
(Note: Adding objects does work so it is not a problem with the saving)
To delete the object I retrieve the set of object and select the objects I want to delete. Then I remove the objects
NSSet *notes = summary.notes;
NSSet *oldNotes = [notes objectsPassingTest:^(id obj,BOOL *stop){
Note *oldNote = (Note *)obj;
BOOL sameRow = (oldNote.row == newNote.row);
BOOL sameColumn = (oldNote.column == newNote.column);
BOOL success = (sameRow && sameColumn);
return success;}];
[summary removeNotes:oldNotes];
I have tried making the relationship inverse to delete the objects which didn't delete them. I have also tried different delete rules (cascade and nullify) which again didn't work. Finally, I tried to remove each object separately and deleting each object from the context after I had removed it from the parent object which again unfortunately didn't work.
I assume the problem must be something to do with it being a fetched object. If anyone could help I would really appreciate it as I can't think of any other ways to test or solve this problem.
You need to do
NSManagedObjectContext * moc = .......;
[moc deleteObject:note]
edit: The core data generated accessors simply remove the object from the relationship, but do not delete the object permanently. This makes sense because you may have one NSManagedObject associated to multiple other NSManagedObjects via relationships.
edit: Deleting in the above mentioned fashion will invoke the deletion rules. I suggest you double check that they are setup correctly.
The reason the above code did not work is that == will not actually compare the NSNumber. Instead you need to call 'isEqualTo:'. I think before it was checking the address hence working before I saved it. What's more it was returning an object in the NSSet so appeared to be working. During debugging it wasn't clear what the object was but clearly wasn't the one I needed.
Related
I'm facing this problem where I have an NSMutableArray containing NSManagedObjects.
And I want to iterate each one NSManagedObject, set a value to a variable and then save them.
The problem is that I have a verification to see if the objects are in Database or not, and after I save my first NSManagedObject to the database all the others NSManagedObjects that are in that array are also inserted into the database...
Here's a piece of code that describes what is my issue:
for (Category* c in categories) {
Category* original = [Database getCategoryByID:c.categoryid];
// After the first save this will have value because it saves all the objects
// that are in categories Array and the version will be the same...
if (original != nil && original.version >= c.version) {
// This object is already up to date so do not make any changes
} else {
// This object is not up to date, or it does not exist
// update it's contents
c.name = name;
[[c managedObjectContext] MR_saveToPersistentStoreAndWait];
// Here it saves all the objects instead of only 1 object
}
}
Is there a way to only save 1 object at a time while having other NSManagedObjects inside an NSMutableArray ?
With Core Data, you tell a context to save, and it saves everything it has. There's no way to save just one object unless that object is the only one in a context with changes.
In your case, it looks like your array of Category objects are managed objects and that they belong to a managed object context. The easiest way to avoid unexpected saving is to move the save command until after the loop. When you reach that point,
Any Category that needed to be created or updated is ready to save
Any Category that didn't need to be updated has no changes, so it won't be affected by saving the context.
So it should be safe to save everything in the context then. If it has changes, it gets updated, and if it doesn't, it won't change.
Say I have a CoreData entity type called Player and it has a to-one relationship (purpose) with an entity type called PlayerPurpose. For completeness, say we have an inverse relationship in PlayerPurpose called parentPlayer. Consider the following swift code:
// Assume we already have a player object in a NSManagedObjectContext called context:
player.purpose = NSEntityDescription.insertNewObjectForEntityForName("PlayerPurpose",
inManagedObjectContext: context) as PlayerPurpose;
// Later in the code, we set the value to nil (or we could have replaced
// it with another call to insertNewObjectForEntityForName)
player.purpose = nil;
// What happens to the previous playerPurpose object within the Managed Object Context?
My question: what happens to the original playerPurpose object within the Managed Object Context when the only reference it has in the data is set to nil (or replaced with another object)?
This is not really related to relationship deletion rules because I'm not explicitly deleting any object -- I'm removing it from any meaningful relationships, making it an orphan.
From an ARC perspective (if the PlayerPurpose was just a normal, non-managed object), the original PlayerPurpose instance now has no references, so it can be cleared from memory -- but what happens in the Managed Object Context? Does CoreData recognize this as an orphaned object and delete it via the context?
If not, then I assume I have to be careful to delete any managed object created via a context if I'm going to get rid of all references to it. Assuming that's the case, is there a good pattern go use for making sure that orphaned objects get cleared from the NSManagedObjectContext and that they are no longer stored in the persistent store?
Thanks!
Core Data does not automatically delete objects in this scenario, because "orphaned" is a concept that your code has but not one that Core Data recognizes. There's no reason for it to delete a PlayerPurpose object just because one of its relationships is nil.
The most reliable way to ensure that PlayerPurpose instances are deleted would be to
Create custom NSManagedObject subclasses for your entities (if you don't have them already).
Override the setter method for purpose on the Player subclass. If the new value is nil, delete the old one.
You can also handle this by just making sure to call deleteObject: at the appropriate times. Or you could run a clean-up step where you fetch every PlayerPurpose with a nil value for parentPlayer and delete them.
Quick question, if I want to replace a to-many relationship completely. Do I need to delete all the previous objects, or is that done automatically
newVariable.hasRel = [NSSet setWithSet:newSet];
Where the variable has had some objects in that set, will just replacing it with the new set get rid of the old entities ???
Quick answer: Assigning a new set of related objects does not delete the previously related objects. You have to do that manually.
i want to insert a previously created NSManagedObject which has some string attributes into my NSManagedObjectContext via the insertObject method. This seems to work without error but when i reload the saved object again all my string attributes are null.
I have created my entity with:
[[NSManagedObject alloc] initWithEntity:description
insertIntoManagedObjectContext:nil];
Thanks for any constructive feedback.
It is not advisable to insert a "previously created" managed object because it is much more robust to insert it during creation.
If you want to copy a managed object (i.e. "insert it again") effectively creating two instances of it in the persistent store, you will have to create a new one and then copy the all the attributes.
It is certainly more reasonable to insert the object into the context and then delete it if desired. It is just the much more intuitive and documented way to achieve what you want.
I strongly advise against creating a context-less managed object in your usage case!
In the documentation and in the broad literature the generated factory method to delete/remove a subclassed managed object in CoreData for IOS is
(void)removeXXXObject:(NSManagedObject *)value
where XXX is the corresponding relationship or we can use simply removeObject.
In my code I used this:
Data *lastData = [[self sortedPersonDatas] objectAtIndex:0];
[selectedPerson removePersonDatasObject:lastData];
where PersonDatas is a one-to-many relationship to Data managed object from I took the last data (lastData resulted from a sorted array of all data)
But using the first two remove methods and checking the SQL database behind we can find that the actual data is existing just the inverse relationship is null.
To completely delete the data (all attributes and the object) I had to use:
[selectedPerson.managedObjectContext deleteObject:lastData];
The question: which is the better method and is it correct that CoreData leaves the data intact?
removeXXXObject only removes an object from a to-many relationship, but does not delete the object from the store. To do so, you have to indeed use deleteObject - this is the desired behavior. Calling deleteObject will by default also set the corresponding relationships to nil (see https://developer.apple.com/library/mac/#documentation/Cocoa/Conceptual/CoreData/Articles/cdRelationships.html#//apple_ref/doc/uid/TP40001857-SW1).