sort arrays of double in C

Question

if I have an array

double i[5] = {1.023, 1.22, 1.56, 2, 5, 3.331};

how do i sort the values so that they look like this:

double i[5] = {1.023, 1.22, 1.56, 2, 3.331, 5};

i've tried qsort() with no luck, after trying some examples, i came up with:

qsort(i, 5, sizeof(double), sort);

int sort(const void *x, const void *y)
{
return (*(double*)x - *(double*)y);
}

with => error: incompatible type for argument 1 not sorting the array.....

Solution

The first argument to qsort is the pointer to the start of the array to be sorted. Instead of

qsort(i[5], 5, sizeof(double), sort);

it should read

qsort(i, 5, sizeof(double), sort);

Some further observations:

1. The length of i's initializer is incorrect (i has five elements, yet the initializer has six).
2. Hard-coding the 5 into the qsort call is asking for trouble later on.
3. The name "i" is most commonly used for loop counters and the like.
4. Calling the comparison function sort is confusing.
5. Your comparison function is wrong. Consider how it would compare the numbers 1.1 and 1.2. Also think about what would happen if the difference between the two values doesn't fit in an int.

I would rewrite your entire example like so:

double arr[] = {1.023, 1.22, 1.56, 2, 5, 3.331};

int cmp(const void *x, const void *y)
{
  double xx = *(double*)x, yy = *(double*)y;
  if (xx < yy) return -1;
  if (xx > yy) return  1;
  return 0;
}

int main() {
  qsort(arr, sizeof(arr)/sizeof(arr[0]), sizeof(arr[0]), cmp);
}

Note that the above comparison function still doesn't correctly handle NaNs; I leave it as an exercise for the reader to fix that.

Comments

[tom91136]

opps , typo, i updated the question, but now, it's not returning sorted array:(

[NPE]

@Tom91136: Your sort function is wrong. See the updated answer.

[tom91136]

thanks! i was so stupid, that sort function was originally used somewhere in my code to sort int arrays...

[Keith Thompson]

@Tom91136: Even for int arrays, that comparison function (please don't call it sort) won't always work. The difference between to int values won't necessarily fit in an int; for example, INT_MAX - INT_MIN will overflow. A trick I've seen is return (xx > yy) - (xx < yy);.