使用for循环合并两个排序的数组

Question

I have a function that merges two sorted arrays into one and returns a pointer to it. I want to use a for loop rather than a while. However in some test cases the last 1 or 2 elements of the merge array are not in their place. I would appreciate if someone can help solve this problem keeping the for loop.

int * mergeSort(int arr1[], int arr2[],int len)
{

  /* len is the combined length of the two arrays */

    static int sorted[100];

    int pos1=0, pos2=0;

    for (int i=0; i<len; i++)
    {
        if (arr1[pos1]<=arr2[pos2])
        {
            sorted[i]=arr1[pos1];
            pos1++;
        }
        else
        {
            sorted[i]=arr2[pos2];
            pos2++;
        }
    }

    return sorted;
}

Answer 1

Your problem is that you don't seem to handle going past the end of the input arrays. If there is uninitialized memory - you get undefined behaviour.

You can avoid this by terminating your arrays with a sentinel value, for example INT_MAX , which should always be bigger than all possible values in the arrays:

int a[] = { 1, 2, 104, INT_MAX};
int b[] = { 101, 102, 105, INT_MAX};

int* ptr = mergeSort(a,b,6);

for(int i = 0; i < 6; i++){
    cout << i << " " << ptr[i] << endl;
}

live demo

Or you can pass the actual lengths of both arrays and handle them correctly:

int * mergeSort(int arr1[], int len1, int arr2[],int len2)
{

  /* len is the combined length of the two arrays */

    static int sorted[100];

    int pos1=0, pos2=0;

    for (int i=0; i< len1 + len2; i++)
    {
        if ((pos2 == len2) || (arr1[pos1] <= arr2[pos2] && (pos1 < len1)))
        {
            sorted[i]=arr1[pos1];
            pos1++;
        }
        else
        {
            sorted[i]=arr2[pos2];
            pos2++;
        }
    }

    return sorted;
}

live demo

Answer 2

This doesn't answer the question of what's wrong with your code, but to answer the question of how to merge two sorted ranges I would suggest std::merge :

    int * mergeSort(int arr1[], int arr2[], int len1, int len2)
    {
        //I am not condoning the use of a static buffer here,
        //I would probably use a std::vector or std::array,
        //possibly a boost::static_vector if really necessary
        static int sorted[100];
        std::merge(arr1, arr1 + len1, arr2, arr2 + len2, sorted);
        return sorted;
    }

I also changed int len to int len1, int len2 because you need to know the lengths of the individual arrays, not just their combined length, to avoid reading past the end of the input arrays.