in reply to Array vs. Hash for sparsely integer-indexed data
It's hard to know how sparse your data set is, and how large the range. If you have 100 items with indices between 0 and 3_000_000, the array isn't going to work out so well. On the other hand, if you have 10_000 items with indices between 0 and 32767, the array wouldn't be so bad after all.
While hash insertions and lookups are both O(1) operations in order of growth, the constant factors are more expensive with hashes than with arrays. But do you know speed efficiency to be an issue?
Without knowing enough about the specifics of the problem, I would have to recommend a hash. Once specifics are known (how sparse, how wide the range of indices, and how time-critical the code is), that recommendation could change. But the hash is the "general" solution, where the array would be a solution tailored to a specific set of criteria.
Dave
|
|---|