Question

I need an algorithm that introduces data like a stack, so when I scan the structure I can read them in the same sequence as they were introduced, for sequential access. Also these values are stored in buckets, like a hashtable, so I can fragment the whole structure for disk storage and have fast random access.

Is there an algorithm like this, or should I have two separate structures? What’s the best strategy?

Best regards!

Answer 1

What I would probably do would be this:

1. Create a hash table where you actually store the entries
2. Create a stack that stores the true memory locations of the objects (not the objects themselves)
3. Abstract these two structures behind a class (or something similar), so that its true implementation is hidden from the user.