Technical reasons behind using Arrays.deepHashCode() when calculating hashcode for multidimensional array

Question

I have read that to calculate the hashcode for multidimensional array, one has to use Arrays.deepHashCode() method instead of Arrays.hashCode() method, but I don't really understand the technical reason behind it. Could someone please explain it to me?

    Object[][] one = {
        {1, 2, 3},
        {11, 22, 33}
    };
    int two = Arrays.deepHashCode(one);
    int three = Arrays.hashCode(one);
    System.out.println("two " + two);
    System.out.println("three " + three);

Result:

two 997365
three 312355675

Answer 1

The technical reason is that if you simply you call hashCode() on the array, you will get an "identity" hashcode; i.e. a hashcode that ignores the values stored in the array. Thus

Object[][] one = {
    {1, 2, 3},
    {11, 22, 33}
};

Object[][] won = {
    {1, 2, 3},
    {11, 22, 33}
};

println(won.hashCode() == one.hashCode());

will print false. In typical cases, you would want hash codes for one and won to be equal. To get that, you need the hash code calculation to include the values of all of the array's elements.