Generating Unique Random Numbers in Java

With Java 8+ you can use the ints method of Random to get an IntStream of random values then distinct and limit to reduce the stream to a number of unique random values.

ThreadLocalRandom.current().ints(0, 100).distinct().limit(5).forEach(System.out::println);

Random also has methods which create LongStreams and DoubleStreams if you need those instead.

If you want all (or a large amount) of the numbers in a range in a random order it might be more efficient to add all of the numbers to a list, shuffle it, and take the first n because the above example is currently implemented by generating random numbers in the range requested and passing them through a set (similarly to Rob Kielty's answer), which may require generating many more than the amount passed to limit because the probability of a generating a new unique number decreases with each one found. Here's an example of the other way:

List<Integer> range = IntStream.range(0, 100).boxed()
        .collect(Collectors.toCollection(ArrayList::new));
Collections.shuffle(range);
range.subList(0, 99).forEach(System.out::println);

Use Collections.shuffle() on all 100 numbers and select the first five, as shown here.

• Add each number in the range sequentially in a list structure.
• Shuffle it.
• Take the first 'n'.

Here is a simple implementation. This will print 3 unique random numbers from the range 1-10.

import java.util.ArrayList;
import java.util.Collections;

public class UniqueRandomNumbers {
    
    public static void main(String[] args) {
        ArrayList<Integer> list = new ArrayList<Integer>();
        for (int i=1; i<11; i++) list.add(i);
        Collections.shuffle(list);
        for (int i=0; i<3; i++) System.out.println(list.get(i));
    }
}

---

The first part of the fix with the original approach, as Mark Byers pointed out in an answer now deleted, is to use only a single Random instance.

That is what is causing the numbers to be identical. A Random instance is seeded by the current time in milliseconds. For a particular seed value, the 'random' instance will return the exact same sequence of pseudo random numbers.

1. Create an array of 100 numbers, then randomize their order.
2. Devise a pseudo-random number generator that has a range of 100.
3. Create a boolean array of 100 elements, then set an element true when you pick that number. When you pick the next number check against the array and try again if the array element is set. (You can make an easy-to-clear boolean array with an array of long where you shift and mask to access individual bits.)