Is there an efficient implementation of merging in space?

I just coded this working version of mergesort:

static int[] merge(int[] first, int[] second){
    int totalsize = first.length + second.length;
    int[] merged_array = new int[totalsize];
    int i = 0, firstpointer = 0, secondpointer = 0;
    while(i < totalsize){
        if(firstpointer == first.length){
            merged_array[i] = second[secondpointer];
            ++secondpointer;
        }
        else if(secondpointer == second.length){
            merged_array[i] = first[firstpointer];
            ++firstpointer;
        }
        else if(first[firstpointer] < second[secondpointer]){
            merged_array[i] = first[firstpointer];
            ++firstpointer;
        }
        else{
            merged_array[i] = second[secondpointer];
            ++secondpointer;
        }
        ++i;
    }
    return merged_array;
}

static int[] mergesort(int[] array){

    if(array.length == 1){
        return array;
    }
    else{
        int length = array.length;
        int[] first = Arrays.copyOfRange(array, 0, (int) length / 2);
        int[] second = Arrays.copyOfRange(array, (int) length / 2, length);
        return merge(mergesort(first), mergesort(second));
    }

}

      

However, if you noticed, I use the copyOfRange function, which creates a new array that is a copy of a specific portion of the parent array. Is there a merge implementation in java that is more space efficient than this?

+3


source to share


1 answer


Duplicate: How to sort using merge sort algorithm?

Synopsis: Yes, there are memory-related merge types, but they are either a) very complex or b) inefficient in time: O (n ^ 2 log n)



Basically, don't worry. It's not really that much memory you save, and if you really want to, just use quicksort instead.

+2


source







All Articles