This solution does it in O(n) worst case where n is your list of exclusions, and constant memory. The code is a little longer but might be relevant if you:
- Possibly have a huge list of exclusions
- Need to run this many times
- Have a large range
It preserves the random distribution in the sense that it actually skips the exclusion list and generates a random number within the range excluding the set.
This is the implementation:
private static int RandomInRangeExcludingNumbers(Random random, int min, int max, int[] excluded)
{
if (excluded.Length == 0) return random.Next(min, max);
//array should be sorted, remove this check if you
//can make sure, or sort the array before using it
//to improve performance. Also no duplicates allowed
//by this implementation
int previous = excluded[0];
for (int i = 1; i < excluded.Length; i++)
{
if (previous >= excluded[i])
{
throw new ArgumentException("excluded array should be sorted");
}
}
//basic error checking, check that (min - max) > excluded.Length
if (max - min <= excluded.Length)
throw new ArgumentException("the range should be larger than the list of exclusions");
int output = random.Next(min, max - excluded.Length);
int j = 0;
//set the start index to be the first element that can fall into the range
while (j < excluded.Length && excluded[j] < min) j++;
//skip each number occurring between min and the randomly generated number
while (j < excluded.Length && excluded[j] <= output)
{
j++;
output++;
while (excluded.Contains(output))
output++;
}
return output;
}
And a test function to make sure it works (over 100k elements)
private static void Test()
{
Random random = new Random();
int[] excluded = new[] { 3, 7, 80 };
int min = 1, max = 90;
for (int i = 0; i < 100000; i++)
{
int randomValue = RandomInRangeExcludingNumbers(random, min, max, excluded);
if (randomValue < min || randomValue >= max || excluded.Contains(randomValue))
{
Console.WriteLine("Error! {0}", randomValue);
}
}
Console.WriteLine("Done");
}