Parallel sort two arrays - c

Question

Suppose I have two arrays:
char **words = (char**) malloc(MAX * sizeof(char*));
int *count = malloc(MAX * sizeof(int));
The first one stores the words in a given list:
words[0] = "myword0" / words[1] = "myword1" / words[2] = "myword3" ...
and the second one stores the number of occurences of each word in the list:
count[0] = 4 //means that "myword0" appears 4 times in the list
I want to print the most frequent word and its number of occurrences. If there are words that appear the same number of times, print the first one in alphabetical order.
To do that, I thought about alphabetically sorting words:
int sortWordsAlph(const void* a, const void* b)
{
char **x = (char**)a;
char **y = (char**)b;
return (strcmp(*x,*y));
}
and in int main():
qsort(words, MAX, sizeof(char*), sortWordsAlph);
Now, the problem is with count. It should still point to the occurrences of each word, thus meaning I have to sort it too. How can I do this?
Maybe I should use a swapping algorithm instead of qsort?

Answer 1

If you want to use qsort(), you should use an array of structs that contain both the word and the count. You can then pass your array to qsort() with a custom comparison function:
struct wc_s {
char *word;
int count;
};
int cmp_words (const void *p1, const void *p2)
{
const struct wc_s *s1 = p1;
const struct wc_s *s2 = p2;
return strcmp (s1->word, s2->word);
}
int cmp_counts (const void *p1, const void *p2)
{
const struct wc_s *s1 = p1;
const struct wc_s *s2 = p2;
return s2->count - s1->count;
}
...
struct wc_s *wc_list = malloc (MAX * sizeof *wc_list);
...
qsort (wc_list, MAX, sizeof *wc_list, cmp_counts);
If qsort() was guaranteed to perform a stable sort (i.e. two elements that compare the same retain their original order), you could solve your problem by first sorting by word, and then sorting by count. Unfortunately, the resulting order of two equal elements is unspecified.
What you could do is sort the array by count, and then go through the array to find sub-arrays with the same counts, and sort those individually:
int start = 0;
int length;
while (start < MAX) {
for (length = 1; start+length < MAX; length++) {
if (wc_list[start].count != wc_list[start+length].count)
break;
}
qsort (wc_list+start, length, sizeof *wc_list, cmp_words);
start += length;
}