How to create int array containing (quasi) random values in a given range but with a fixed % of duplicates?

Question

I am trying to create (in Java) an integer array containing a specified number of integers, k, (for example, 50, 500, 1000, etc.) that does not use library functions or Collections, but contains a random assortment of numbers in the range (i.e. from 1 to k), with a specified percentage of duplicates.

I have figured out how to implement the shuffle, but I am not sure how to best implement the dups threshold. What I have so far:

static void shuffle(int[] array) {
int n = array.length;
for (int i = 0; i < array.length; i++) {
    // Get a random index of the array past i.
    int random = i + (int) (Math.random() * (n - i));
    // Swap the random element with the present element.
    int randomElement = array[random];
    array[random] = array[i];
    array[i] = randomElement;
}
}

Then, using this method:

    //Create an ascending ordered array of size k to be shuffled
    int [] tempInit = new int [filesizes[i]];

    for(int k = 0; k < filesizes[i]; k++)
    {

        tempInit[k] = k+1;

    }

    //Shuffle the ascending array                                                           
    shuffle(tempInit);

    for(int k = 0; k < tempInit.length; k++)
    {

        System.out.println(tempInit[k]);
    }

One way I imagine it could work would be to use the percentage of duplicates required (let's say it's 20%) and then randomly select a fixed number of integers, and loop through the array and replace every element (if not equal to one of the fixed numbers) with one of these fixed numbers. I'm not sure what sort of logic would make sense to select these fixed numbers, or the amount of fixed numbers to use for replacement purposes.

henorek · Accepted Answer · 2015-12-04 22:38:17Z

1

Try this! For 80% for number 1 and 5% for each other up to 5:

Map<Double, Integer> map = new LinkedHashMap<>();
map.put(0.8, 1);
map.put(0.85, 2);
map.put(0.9, 3);
map.put(0.95, 4);
map.put(1, 5);
Random random = new Random();
double result = random.nextDouble();
int value = map.entrySet().stream().filter(entry -> entry.getKey() < result).reduce(0, Math::min);

answered Dec 4, 2015 at 22:38

henorek

5021 gold badge6 silver badges19 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

ymediceacuerdate · Accepted Answer · 2015-12-04 22:53:35Z

edit: solved this as follows:

double percentDupes= 0.2;

//Create an ascending ordered array of size k to be shuffled
int [] tempSorted = new int [filesizes[i]];

for(int k = 0; k < filesizes[i]; k++)
{

    tempSorted[k] = k+1;

}

//Shuffle the ascending array                                                           
shuffleArray(tempSorted);


//Figure out the proportion of the array to replace with duplicate values of elements already in the array 
Double proportion =  tempSorted.length*percentDupes;
int ratio = proportion.intValue();

//Instantiate a new array of size "ratio"
int [] tempShuffled = new int[ratio];

//Fill the new, smaller array with randomly selected elements from the original, sorted array               
for(int b = 0; b< tempShuffled.length; b++)
{

    int randNum = i + (int) (Math.random() * (tempSorted.length - i));  //Select a random element of the array to have as a duplicate               
    tempShuffled[b] = tempSorted[randNum];  //tempSorted was previously shuffled  
}

//Shuffle this sub-array 
shuffleArray(tempShuffled);     

//Loop through (shuffled) original array and, if the values don't match, replace a non-dup with a dup value 

for(int c= 0; c<tempShuffled.length; c++ )
{
    if(tempSorted[c] != tempShuffled[c])
    {

        tempSorted[c] = tempShuffled[c];
    }
}


//Final output to print to file                                             
for(int k = 0; k < tempSorted.length; k++)
{

    System.out.println(tempSorted[k]);
}

Collectives™ on Stack Overflow

How to create int array containing (quasi) random values in a given range but with a fixed % of duplicates?

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related