How to represent 6 integer key better than as indices to a 6-dimensional array?

Question

In my program, a state can be uniquely identified by six integers. Each integer, i satisfies 0 <= i <= 10 and each state has an associated value. I am currently using a 6-dimensional array to keep track of each states value. I am storing the state values in an array like so state[11][11][11][11][11][11][11], where each value has the 6 integers as a key. However, this array would be very sparse, as I only visit a small number of possible states. Is there a better way to represent the key for the value of a state?

Answer 1

If you don't need to have a value in every state you can get away with using a map

Concept code, totally untested.

constexpr int dimension = 6;
using KeyType = std::array<char, dimension>;
int32_t Key(const & KeyType keys) {
  int32_t res = 0;
  for (auto key : keys) {
    res <<= 4;
    res += key;
  }
  return res;
}

void Key2Array(int32_t keyValue, KeyType& keys) {
  int idx = dimension-1;
  for (auto& key : keys) {
    keys[idx--] = keyValue&0x16;
    keyValue >>= 4;
  }
}

std::map<int32_t, value> states;

states[Key({1,2,3,4,5,6}] = 42;

KeyType key;
Key2Array(0x123456, key);