Interview Question for Software Development Engineer In Test (SDET) at Microsoft:

Assuming the array items are of type unsigned short int, here is possible solution:
void find_duplicates(unsigned short *items, int length)
{
    int counts[USHRT_MAX + 1];

    memset(counts, 0, sizeof(counts));

    for (int i = 0; i < length; i++)
    {
        counts[items[i]]++;
    }

    for (unsigned i = 0; i <= USHRT_MAX; i++)
    {
        if (counts[i] > 1)
        {
            printf ("%d: %d\n", i, counts[i]);
        }
    }
}

Another solution with using sorting:

void find_duplicates(char *string)
{
    char *duplicate = strdup(string);
    int length = strlen(duplicate);

    quicksort(duplicate, 0, length - 1);

    for (int i = 0; i < length;)
    {
        int count = 1;

        for (i = i + 1; i < length && duplicate[i] == duplicate[i - 1]; i++)
        {
            count++;
        }

        if (count > 1)
        {
            printf("%c: %d\n", duplicate[i - 1], count);
        }
    }
}

Two solutions that are based on constraints of the question. One solution depends on the constraint of values (i.e. all values range from 1 to 100), in this solution you can use the indicies to represent the values and increment the count with the array. Some of the issues with this is that it doesn't scale well, if the range of values increases then the amount of memory associated with an array that huge is pointless. Also if negatives are used then the array doesn't work. The other solution is to use a hash table where the number is the key and the value is the count of the items. Depending on the constraints of the sets then both solutions can be good and both run in O(n) time.