Question

I am making a java program that uses the interpolation search below I have gotten from wikipedia. In my main program I have created an int array that will have 100,000 spots. Then I fill all those spots with random numbers and sort it. I then generate a random search key and call the function. I am also looping the function call 100 times each time with a different search key. When I do this I get an array out of bounds error on this statement if (sortedArray[mid] < toFind). The program works fine with an array with 10 spots, 100 spots, 1000 spots, but when I get to 100,000 I get the error. Do you know what I can do to fix this problem?

 public int interpolationSearch(int[] sortedArray, int toFind){
  // Returns index of toFind in sortedArray, or -1 if not found
  int low = 0;
  int high = sortedArray.length - 1;
  int mid;

  while (sortedArray[low] <= toFind && sortedArray[high] >= toFind) {
   mid = low +
         ((toFind - sortedArray[low]) * (high - low)) /
         (sortedArray[high] - sortedArray[low]);  

   if (sortedArray[mid] < toFind)
    low = mid + 1;
   else if (sortedArray[mid] > toFind)
    // Repetition of the comparison code is forced by syntax limitations.
    high = mid - 1;
   else
    return mid;
  }

  if (sortedArray[low] == toFind)
   return low;
  else
   return -1; // Not found
 }

Answer 1

You are overflowing the max value that can be stored in an int which is 2,147,483,647.

When calculating the numerator in your equation you are doing ((39258-0)*(99999-0)) which is 3,925,760,742.

If you change it to this you won’t have this issue:

long numerator = (long)(toFind - sortedArray[low]) * (long)(high - low);
mid = low + numerator / (sortedArray[high] - sortedArray[low]);

I think you also need to change your while loop to be:

while (sortedArray[low] < toFind && sortedArray[high] >= toFind) {

Otherwise you have a situation where sortedArray[low] == sortedArray[high] == toFind

Then your equation reduces to

mid = low + (0 * (high - low)) / 0 = 0/0

Java allows division by zero. I’m not exactly sure what happens though, it may be that in this case mid = infinity or NaN.