Java - Finding duplicate entries in multiple arrays

Question

I wish to find unique points in a bunch of Point arrays (i.e. removing all the duplicates and seeing what is left).

Here is the code I have tried:

int numCount = 0;
 for (int x = 0; x < ALLARRAYS.length; x++) {
 for (final Point test : ALLARRAYS[x]) {
 NUMBER: for (final Point[] numbers : ALLARRAYS) {
 if (numbers == ALLARRAYS[x]) {
 continue;
 }
 for (int i = 0; i < numbers.length; i++) {
 if (test.equals(numbers[i])) {
 break NUMBER;
 } else {
 if (i == numbers.length - 1) {
 numCount++;
 }
 }
 }
 }
 if (numCount == 10) {
 System.out.println(x + "\tPoint: " + test.x + ", " + test.y + " is unique.");
 }
 numCount = 0;
 }
 System.out.println();
 }

Basically, I have eleven arrays I want to check, therefore the numCount checks for 10 arrays to be confirmed. My approach was to cycle through each array, then cycle through all the points in that array, and then cycle through each point in every other array. If it sees a duplicate, it skips the whole number entirely.

My problem: I am getting false readings. I am not sure what is going on, but the program spits out points that already exist in other arrays.

My goal: Each array works as a whole. If one element is not satisfied, then the whole array is skipped. My end result should be 11 smaller-in-size arrays that are all unique so that a method can check through the elements of each set and determine that they are all unique.

NOTE: The arrays are all unique now, they are just ridiculously huge. I am looking for a way to truncate it by creating this mini-program to eliminate duplicates.

My question: Can someone please help me out?

• Greg Hewgill: What is your question? Is this code working as you expect, or is there something wrong with it?
• Greg Hewgill: Could you provide some sample input, actual output, and expected output so that we can help figure out the reason that's happening?
• Anthony Grist: test == numbers[i]. Point is a class, comparing it using the == operator is only going to return true if test and numbers[i] have references that are the same. Try using test.equals(numbers[i]) instead?
• Adam Liss: You'll also need to account for the case that compares a Point to itself.
• Adam Liss: If your arrays are "ridiculously huge," consider sorting them. Then you can skip to the next array as soon as you encounter a point that occurs after the one you're testing.

Answer 1

Put all arrays entries in a Set, it will skip the duplicates.

Just be sure to correctly override the equals method on your Point class because the implementations of Set interface will use it.

HTH

• Luiggi Mendoza: that can work if he wants the union of the arrays in a new array
• Dylan Wheeler: I am not looking for this. Please use my example above.
• yshavit: I think a better approach would be to put the entries into a map from entry to number of times entry appeared in the arrays. The key could be an Integer or, if you're worried about object allocation, a mutable Integer-type class. Once you have the Map created, you can iterate over the Map entries and get the ones whose count is 1.

Answer 2

Your code appears to check all arrays for the Point in question, including the array in which the Point you're looking for appears. So, your code will always find the Point it's looking for, and you will end up with no results.

You may need to do something like this:

 NUMBER: for (final Point[] numbers : ALLARRAYS) {
 if (numbers == array) {
 continue; // skip the array that contains 'test'
 }
 ...
 }

That way, you will skip the array that contains the test element.

• Anthony Grist: @JavaCoder-1337 Greg's edit should cover it, provided you don't have any duplicate Points in a single array that you want to ignore.
• Dylan Wheeler: Your suggestion helped, but still gave me some duplicates. I have edited my original post including my current code.
• Adam Liss: The bulletproof answer is to test (numbers == array && test == numbers[i]) in the inner loop. But why use two different styles of for loop? No need for the counter i. Re: being "too pushy," just be polite, and if you're asking entry-level questions you may want to reconsider "JavaCoder" and "1337" as a handle.
• Greg Hewgill: @JavaCoder-1337: Ok, so you've (a) moved the goalposts, invalidating the existing answers, and (b) continued to fail to provide sample data that demonstrates your problem. At this point we're having an increasingly difficult time replicating your problem, since we don't have all the information you do.

Answer 3

Arrays aren't well suited for your job. Not only that they don't don't have methods like intersection or removeAll, they are hard to change, especially if you want to reduce their size.

To avoid this hassle, I mark a Point as removed by setting x and y to -1, one coordinate would of course be enough, but for reasons of symmetry ...

import java.util.*;
import java.awt.Point;
public class RandPoints
{
 Random r = new Random ();
 Point [][] arr;
 public void removeDuplicatesFrom2 (Point [] pa, Point [] pb) {
 for (Point p : pb) {
 if (p.x != - 1 && Arrays.asList (pa).contains (p)) {
 // System.out.println ("dup: " + p.x + " " + p.y);
 p.x = -1;
 p.y = -1;
 }
 }
 }
 // create a random point in range (0..19)(0..19)
 public Point randpoint () {
 return new Point (r.nextInt (20), r.nextInt (20));
 }
 // create 10 (default) arrays of size 100 (default)
 public void init (int arrsize, int arrcount) 
 {
 arr = new Point [arrcount][];
 for (int c = 0; c < arrcount; ++c)
 {
 Point [] points = new Point [arrsize];
 for (int s = 0; s < arrsize; ++s)
 {
 Point p = randpoint ();
 points[s] = p; 
 }
 arr[c] = points;
 }
 }
 public void unify () 
 {
 for (Point[] pac0 : arr) {
 for (Point[] pac1 : arr) 
 {
 if (pac0 != pac1)
 removeDuplicatesFrom2 (pac0, pac1); 
 } 
 }
 }
 /**
 fill arrays with duplicates by random, 
 show, delete and show again.
 */
 public RandPoints (int arrsize, int arrcount)
 {
 init (arrsize, arrcount);
 show ();
 unify (); 
 System.out.println ();
 show ();
 }
 public static void main (String args[])
 {
 int arrsize = 100;
 int arrcount = 10;
 if (args.length == 2)
 {
 arrsize = Integer.parseInt (args[0]);
 arrcount = Integer.parseInt (args[1]);
 }
 new RandPoints (arrsize, arrcount);
 }
 // visible feedback is always welcome while debugging/testing
 public void show ()
 {
 for (Point[] pa: arr) {
 for (Point point: pa)
 if (point.x != -1 && point.y != -1)
 System.out.print (point.x + ", " + point.y + "\t");
 System.out.println ();
 }
 }
}

I modified the code a bit, to make the maximum range for the values (20) configurable too.

Starting it with

java RandPoints 16 10 10
3, 0 9, 0 6, 9 2, 3 6, 9 7, 4 9, 9 2, 5 8, 7 3, 3 9, 5 3, 7 0, 5 7, 6 0, 4 8, 1 
6, 1 2, 7 2, 5 6, 7 0, 7 5, 8 4, 2 1, 9 8, 4 5, 7 0, 2 3, 1 1, 9 2, 1 8, 0 1, 7 
5, 4 9, 7 9, 3 7, 3 1, 2 9, 6 0, 4 6, 0 3, 0 7, 7 1, 3 1, 1 5, 3 3, 8 1, 0 4, 9 
4, 7 8, 9 4, 0 0, 2 8, 7 5, 8 7, 0 1, 4 4, 9 8, 2 6, 9 9, 6 2, 1 1, 9 0, 8 6, 5 
6, 8 9, 6 1, 0 6, 9 4, 0 5, 1 2, 9 7, 3 5, 1 2, 5 6, 9 0, 9 7, 4 8, 1 5, 5 3, 4 
5, 9 0, 4 5, 4 2, 2 2, 6 7, 1 2, 0 6, 1 0, 4 9, 8 5, 7 5, 5 4, 6 9, 0 2, 8 8, 5 
8, 2 4, 2 0, 8 1, 1 0, 3 3, 4 1, 8 3, 1 6, 6 4, 1 3, 6 6, 0 1, 7 4, 8 1, 6 1, 1 
6, 2 1, 3 2, 4 0, 8 9, 0 3, 0 1, 1 3, 7 6, 2 2, 4 0, 9 3, 6 7, 2 1, 2 5, 0 8, 2 
2, 3 5, 6 7, 9 3, 0 9, 3 2, 6 4, 8 8, 7 9, 4 5, 3 0, 3 3, 0 5, 5 1, 4 6, 4 5, 2 
3, 2 4, 9 6, 9 4, 7 7, 1 0, 4 5, 8 7, 2 5, 2 5, 5 2, 1 9, 8 4, 9 6, 6 7, 7 0, 3 
3, 0 9, 0 6, 9 2, 3 6, 9 7, 4 9, 9 2, 5 8, 7 3, 3 9, 5 3, 7 0, 5 7, 6 0, 4 8, 1 
6, 1 2, 7 6, 7 0, 7 5, 8 4, 2 1, 9 8, 4 5, 7 0, 2 3, 1 1, 9 2, 1 8, 0 1, 7 
5, 4 9, 7 9, 3 7, 3 1, 2 9, 6 6, 0 7, 7 1, 3 1, 1 5, 3 3, 8 1, 0 4, 9 
4, 7 8, 9 4, 0 7, 0 1, 4 8, 2 0, 8 6, 5 
6, 8 5, 1 2, 9 5, 1 0, 9 5, 5 3, 4 
5, 9 2, 2 2, 6 7, 1 2, 0 9, 8 4, 6 2, 8 8, 5 
0, 3 1, 8 6, 6 4, 1 3, 6 4, 8 1, 6 
6, 2 2, 4 6, 2 2, 4 7, 2 5, 0 
5, 6 7, 9 9, 4 6, 4 5, 2 
3, 2