I'm a little fuzzy on how the delete rules for relationships in Core Data work, at least beyond the simple cases described in the documentation.
Most of those cases, and most of the answers I've seen for questions here, use a model where the object on left side of a one-to-many relationship "owns" the objects on the right side: e.g. a Person
has PhoneNumber
s, and if you delete the person you delete all their associated numbers. In that kind of case, the solution is clear: Core Data will handle everything for you if you set the relationships like so:
Person --(cascade)-->> PhoneNumber
PhoneNumber --(nullify)--> Person
What I'm interested in is the opposite: A to-many relationship where the "ownership" is reversed. For example, I might extend the CoreDataBooks sample code to add an Author
entity for collecting all info about a unique author in one place. A Book
has one author, but an author has many books... but we don't care about authors for whom we don't list books. Thus, deleting an Author
whose books
relationship is non-empty should not be allowed, and deleting the last Book
referencing a particular Author
should delete that Author
.
I can imagine a couple of ways to do this manually... what I'm not sure of is:
One approach to delete everything and reset Core Data is to destroy the persistent store. Deleting and re-creating the persistent store will delete all objects in Core Data.
Setting the Delete Rule to Protect prevents deleting records of the main Entity while there are associated records in the related Entity. This behavior is ensured by a database constraint created on the reference attribute.
The following worked for me:
Set the deletion rule on the 'book' relationship of your author entity to 'Deny' meaning that as long as there is a book linked to your author it cannot be deleted.
Subclass your book entity and override the prepareForDeletion() function as follows:
public override func prepareForDeletion() {
super.prepareForDeletion()
do {
try author.validateForDelete()
managedObjectContext?.delete(author)
} catch {}
}
Validate for delete will throw an error unless the book relationship is empty. You can optionally handle the error.
You could override prepareForDeletion
in your Book
class and check if the author has any other books. If not you could delete the author.
- (void)prepareForDeletion {
Author *author = self.author;
if (author.books.count == 1) { // only the book itself
[self.managedObjectContext deleteObject:author];
}
}
Edit: To prevent deletion of an author with books you could override validateForDelete
or even better: don't call deleteObject with an author with books in the first place
Rickstr,
Check below for the relationships to get your two criteria done.
Author -- (Deny) -->> Books
deleting an Author whose books relationship is non-empty should not be allowed
DENY: If there is at least one object at the relationship destination, then the source object cannot be deleted.
Book -- (Cascade)-- > Author
deleting the last Book referencing a particular Author should delete that Author
You cannot delete the Author, as our first rule is saying, if there are any Books which are non-empty should not be deleted. If they are not present the Author gets deleted.
I think theoretically it should work. Let me know, if this works or not.
If you love us? You can donate to us via Paypal or buy me a coffee so we can maintain and grow! Thank you!
Donate Us With