I'm trying to figure out the best way to design a multi-field table in which there are no inter-dependencies, but where there is bound to be much tuple-duplication.
I would like to eliminate this duplication, but reel at the idea of making a unique index on multiple fields, nearly the whole table. On the other hand, the prospect of coding a find-and-search by hand is equally horrifying.
So which way to go:
1) create a unique index on multiple fields; 2) avoid duplication by searching at create time; or 3) shut up and accept the wasted space?