The best approach to search for closest object in hash map holding arraylist without changing data structure

Question

Imagine the game handled server sided, don't think of a visual problem.

Given a game 2D map, with size of n * n. The game contains objects (such as door, wall, a table).

Each tile position (x,y) can contain multiple objects, but this is not the catch of the question, it can even be limited to one.

The data structure for the game objects linked to the map is designed like this:

HashMap<Integer, ArrayList<GameObject>>

Where the Key integer represents X position of the object, and the value represents a list of all objects with the same X position, different Y position.

Now, I have a player location, I want to fetch a list of all 10 closest object to the player's position, within 100 tiles distance of the player.

My way of doing so:

Starting from player's position, and simply going x - 1 and x + 1 grabbing the objects lists limited to Z or until you have 10 closest objects collected, and then performing a position distance comparing search on each object. (Note, the objects lists are not sorted)

But this way is not stable, in my opinion, but I am not sure how can you solve this without changing the data stucture?

Answer 1

If it is for some reason important to use hashing, there is a concept called locality sensitive hashing, however it is somewhat complicated. Typically, nearest neighbor search is done with tree indexes, such as quadtrees, R-Trees or kd-trees.

Answer 2

You probably can't. The principled solution is to use a data structure that supports nearest neighbor search.