I'm planning to use Core Data to save both a string that the user will input, as well as a canonicalized string version of that to make searching more efficient (so if I can do searches with CONTAINS, I won't need to add [cd] properties in my predicates). I'm just wondering if it's also recommended to add a fetched index for the canonicalized string or not? Does that help in making string searches more efficient in some way, or will it just use up space for no reason?
If it's a good idea, should I make it a Binary or R-Tree index?
Thanks.
