0

Suppose there is a web service returning a JSON response like this:

{
   "id": "123",
   "name": "John Doe",
   "phone": {
      "country_code": "1",
      "area_code": "11",
      "number": "55544433"
   }
}

This is describing a Person entity, but inside this entity, there is the phone property containing an embedded object.

Phone is not an entity and it is not referenced by an id, it is fully contained inside the scope of this Person entity. However, it would be desirable to map it to a Phone @interface in my code (for example, it could have a method "formatAsString" that returns a formatted string like this: "+1 (11) 55544433")

How should I handle this case when mapping the model to iOS Core Data?

Rafael Jannone
  • 90
  • 1
  • 7

1 Answers1

2

Create a Person entity with a to-many (or to-one if there is only one phone #) relationship to a Phone entity.
The Phone entity will have a to-one relationship to a Person entity.
In addition it will have the inner structure you described (country_code,area_code,number).

You don't need an id property in order to create a CoreData entity.

You could set the relationship to 'cascade', so that when a person is deleted all related phone numbers are deleted.

When you traverse you JSON response, read the "phone" key into a Phone entity and the its person property to the person you are currently traversing.

Dan Shelly
  • 5,991
  • 2
  • 22
  • 26
  • Sounds correct. Is the id completely optional, even when saving the entity to a persistent store? (say, for instance, that I'm saving the Person together with its Phone). Thanks! – Rafael Jannone Jul 04 '13 at 04:39
  • CoreData does not provide an automatic way of uniquing a given property. the only thing guarantied to be unique is the `objectID` property of the managed object. the `objectID` property is managed by CoreData, and you cannot change it yourself. When you save, the entities will be given a permanent `objectID`, regardless of the properties you set on those objects. – Dan Shelly Jul 04 '13 at 05:40