Thread: Array that tests if any elements repeat themselves

Hey guys, I am trying to write a program that takes an array of float values and determines if all the numbers are different for each other. I am trying to set a temp variable that stores the value of the current element in the array and compares if to the next one. If they match, then break the loop and return false.

However, I get an exception. What did I do wrong?

package Creativity3;
 
public class FloatingArray
{
	public static void main(String[] args)
	{
		float[] array = new float[3];
		array[0] = 1; 
		array[1] = 3;
 
		System.out.println(FloatingArray.isDifferent(array));
	} 
 
	public static boolean isDifferent(float[] array)
	{
		float temp;
		boolean count = true;
		if(array.length == 0)
		{
			throw new IllegalArgumentException("Array can't be empty!");
		}
 
		else
		{
			for(int i = 0; i <= array.length - 1; i++)
			{
				temp = array[i];
				if(temp == array[i + 1])
				{
					count = false;
					break;
				}
			}
		}
		return count;
	}
}

Exception:

Exception in thread "main" java.lang.ArrayIndexOutOfBoundsException: 3
at Creativity3.FloatingArray.isDifferent(FloatingArra y.java:28)
at Creativity3.FloatingArray.main(FloatingArray.java: 11)

Thank you.

What's going to happen on the indicated line

			for(int i = 0; i <= array.length - 1; i++)
			{
				temp = array[i];
				if(temp == array[i + 1]) // ***** here
				{
					count = false;
					break;
				}
			}

when i = array.length - 1?

Another problem is that of your logic. Remember that floating point numbers are not represented on digital computers with 100% precision (or is it accuracy? I get the two confused!), and so testing for equality can be problematic. You'd be much better off using ints or if you can't use ints, then use doubles which allow for more significant digits than floats, and test not for equality, but for being "close enough".

One way would be to test for equality like this:

public static boolean closeEnough(double expectedValue, double observedValue) {
   if (expectedValue == 0.0) {
      return EPSILON < Math.abs(expectedValue - observedValue);
   } else {
      return EPSILON < Math.abs((expectedValue - observedValue) / expectedValue);
   }
}

Where EPSILON is a constant double of small value.

Originally Posted by Norm

At line 28 the index to the array goes past the end of the array. Check the code to see how the index got past the end of the array.
Remember that array indexes range from 0 to the array's length-1

I put array.length - 2 and it worked thanks.

It may not work as well as you think. The logic I believe will require *nested* for loops. Test your code out on some arrays and see for yourself. Your current code only appears to work for same numbers next to each other in the array.

Originally Posted by curmudgeon

It may not work as well as you think. The logic I believe will require *nested* for loops. Test your code out on some arrays and see for yourself. Your current code only appears to work for same numbers next to each other in the array.

I tried to do another for loop but this one always gives me false (reasonable, considering that if I loops through it I will always find the value that I stored in temp. How do I fix this?

Think of the algorithm first, how you would solve this on paper if you had a large collection of numbers. You'd walk down the list, first seeing if there are duplicates of the first number in all the numbers *above* it, then seeing if there are any duplicates of the second number in any numbers *above* it, then checking the third number against all the numbers *above* it,... do you see the logic inherent in this process?

Originally Posted by curmudgeon

Think of the algorithm first, how you would solve this on paper if you had a large collection of numbers. You'd walk down the list, first seeing if there are duplicates of the first number in all the numbers *above* it, then seeing if there are any duplicates of the second number in any numbers *above* it, then checking the third number against all the numbers *above* it,... do you see the logic inherent in this process?

I think that this one should work, but for some reason it doesn't. Does anything clear seem wrong now again?

	public static boolean isDifferent(float[] array)
	{
		float temp;
		boolean count = true;
 
		if(array.length == 0)
		{
			throw new IllegalArgumentException("Array can't be empty!");
		}
 
		else
		{
			for(int i = 0; i < array.length; i++)
			{
				temp = array[i];
				for(int c = i + 1; c < array.length; c++)
				{
					if(i != c && array[i] == array[c])
					{
						count = false;
					}
				}
			}
		}
		return count;
	}

Originally Posted by jean28

... but for some reason...

So, what if the array contains [1.0, 3.0, 0.0]? What would you expect? What does it give?

What if the array contains [1.0, 2.0, 3.0, 2.0, 5.0, 6.0]? What would you expect? What does it give?

Well, I'll create a couple of arrays and print out each array and result from your method...

import java.util.*;
 
public class FloatingArray {
    public static void main(String [] args) {
        float [] x = {1, 3, 0};
        System.out.println("x = " + Arrays.toString(x));
        System.out.println("FloatingArray.isDifferent(x) = " + FloatingArray.isDifferent(x));
        System.out.println();
 
        float [] y = {1, 2, 3, 2, 5, 6};
        System.out.println("y = " + Arrays.toString(y));
        System.out.println("FloatingArray.isDifferent(y) = " + FloatingArray.isDifferent(y));
 
    } // End of main
 
    public static boolean isDifferent(float [] array) {
        .
        .
        .

Output

x = [1.0, 3.0, 0.0]
FloatingArray.isDifferent(x) = true

y = [1.0, 2.0, 3.0, 2.0, 5.0, 6.0]
FloatingArray.isDifferent(y) = false


Are the outputs for these examples correct? (I think they are.)


Bottom line:
How about showing us an example for which you think it gives an incorrect result? Maybe someone can walk you through it...



Cheers!

Z