It is even better to have 100 functions work on 100 data structures. Powerful programming languages like Lisp and Haskell give you that. Generics gives you most of that.
If every ounce of performance matters, e.g. in a database, you want 10000 functions, 100 for each data structure.
Lisp gives you an infinite amount of functions that operate on one data structure: the cons cell. :P
I guess all you really need are dynamically allocated arrays. A cons cell is an array of two. A struct with N fields is an array of N. Everything else is built on that.
If every ounce of performance matters, e.g. in a database, you want 10000 functions, 100 for each data structure.