Implementing Find-Or-Create During Core Data Flat File Import

Question

I am importing a large plist/XML into Core Data. The structure is simple: let's say there is Person and Company, and a Company can have many Persons. The import goes successfully, but the plist has no established relations, so duplicates of Company are inserted every time multiple people have the same Company.

A potential solution lies in Apple's Core Data docs under 'Implementing Find-or-Create Efficiently':

Or if you import "flat" data with no relationships, you can create managed objects for the entire set and weed out (delete) any duplicates before save using a single large IN predicate.

I've stared at this sentence for ages and can't parse it. Wasn't I already using managed objects to import the entire set? What fetch request are they alluding to?

An algorithm or clarification would be much appreciated.

Answer 1

In your example your best bet is to do a lookup/search on the company during the import so that you can set up the relationships correctly. Depending on the size of your data you may even want to keep the company objects in memory in a NSDictionary so that you can easily join them to the person objects as they are imported.