Question

I’ve been using RandomStringUtils to generate random IDs to be used as database keys:

import org.apache.commons.lang.RandomStringUtils;
public class RandomStringTest {
    public static void main(final String[] args) {
        for (int i = 0; i <= 10; i++) {
            final String id = RandomStringUtils.random(8,
                    "0123456789abcdefghijklmnopqrstuvwxyz");
            System.out.println(id);
        }
    }
}

The key space is large enough,

len("0123456789abcdefghijklmnopqrstuvwxyz")^8 = 2821109907456 ≃ 10^12

Is the random mechanism properly seeded? I need to know that the keys are properly distributed before applying this to production.

By the way, the test code showed no repetitions when executed a few times, but that’s far from solid proof.

Answer 1

There is no seed provided in the implementation I found, there is just a new Random(). But there is a method where you can supply your own source of randomness

random(int count, int start, int end, 
        boolean letters, boolean numbers, 
        char[] chars, java.util.Random random)