I was having this discussion with my friend who had this question asked to him in the Interview. The Question goes like this. Write a Function which takes in a byte array(2 dimensional) as input along with an Integer n, The initial assumption is all the elements of M*N byte array is zero and the problem is to fill 'n' Byte array elements with value 1, For instance if M=5 and N=5 and the n value is 10 the Byte array should've 10/25 elements to be 1 and rest of the 15 values to be 0. The values filled should be random and one cell in byte array should be filled only once. I was fascinated to try solving this on my own. I've attached the code I've come up with so far.
public Boolean ByteArrayFiller(int a,int b, int n)
{
int count = n;
int iLocalCount = 0;
byte[,] bArray= new byte[a,b];
for (int i = 0; i <a; i++)
for (int j = 1; j <b; j++)
bArray[i, j] = 0;
Random randa= new Random();
int iRandA = randa.Next(a);
int iRandB = randa.Next(b);
while (iLocalCount < n)
{
if (bArray[iRandA, iRandB] == 0)
{
bArray[iRandA, iRandB] = 1;
iLocalCount++;
}
iRandA = randa.Next(a);
iRandB = randa.Next(b);
continue;
}
//do
//{
// //iRandA = randa.Next(a);
// //iRandB = randa.Next(b);
// bArray[iRandA,iRandB]=1;
// iLocalCount++;
//} while (iLocalCount<=count && bArray[iRandA,iRandB]==0);
return true;
}
The code i wrote is in C# but it's straight forward to understand. It's able to do the purpose of the question( I did some trials runs and results came out correctly) perfectly but I have used Random object in C#(Equivalent to Math.Rand in Java) to fill up the byte array and I keep thinking if Rand returns the same values for a and b. There is a good chance for this to go indefinitely. Is that the purpose of the question? or Does the solution that i came up for this question is good enough!
I am curious to see how experts here solve this problem? I am just looking for new ideas to expand my horizon. Any pointers would be greatly appreciated. Thanks for taking the time to read this post!
A while loop trying random locations until it finds a good one is generally a very bad approach. If n = M*N, then the last one will have a probability of 1/(M*N) of finding a match. If M*N are sufficiently large, this can be extremely inefficient.
If M*N is not too large, I would create a temporary array of M*N size, fill it with the numbers 0 through (M*N)-1, and then permutate it - i.e. you walk through it and swap the current value with that of a random other value.
Then you go to the first n elements in your array and set the appropriate cell. (row = value / columns, col = value % columns).
I would treat the array, logically, as a one-dimensional array. Fill the first n positions with the prescribed value, and then shuffle the array.
Given a byte array, and the number of rows and columns in the array, and assuming that the array is already filled with 0:
int NumElements = NumRows * NumCols;
for (int i = 0; i < NumElementsToFill; ++i)
{
int row = i / NumRows;
int col = i % NumCols;
array[row, col] = 1;
}
// Now shuffle the array
Random rnd = new Random();
for (int i = 0; i < NumElements; ++i)
{
int irow = i / NumRows;
int icol = i % NumCols;
int swapWith = rnd.Next(i+1);
int swapRow = swapWith / NumRows;
int swapCol = swapWith % NumCols;
byte temp = array[irow, icol];
array[irow, icol] = array[swapRow, swapCol];
array[swapRow, swapCol] = temp;
}
The key here is converting the one-dimensional index into row/col values. I used / and %. You could also use Math.DivRem. Or create Action methods that do the get and set for you.
Choose a number, which is larger than both N and M and is prime (or co-prime to both N and M). Let's call this number p.
Loop until you've set x numbers:
Generate a random number less than N*M. Call this number `l`.
Then the next place to put the number will be `p*l%(N*M)`, if that position hasn't been set.
A downside to this approach is that if the array is filling up, you'll have more collisions.
Bascially, you need to choose n unique random numbers from range [0, p) (where p = M * N), and map them to positions of 2-dimensional array.
Naive approaches are 1) generate non-unique numbers with retry 2) fill an array with numbers from 0 to p-1, shuffle it and take first n numbers (takes O(p) time, O(p) memory).
Another approach is to choose them with the following algorithm (O(n2) time, O(n) memory, code in Java):
public Set<Integer> chooseUniqueRandomNumbers(int n, int p) {
Set<Integer> choosen = new TreeSet<Integer>();
Random rnd = new Random();
for (int i = 0; i < n; i++) {
// Generate random number from range [0, p - i)
int c = rnd.nextInt(p - i);
// Adjust it as it was choosen from range [0, p) excluding already choosen numbers
Iterator<Integer> it = choosen.iterator();
while (it.hasNext() && it.next() <= c) c++;
choosen.add(c);
}
return choosen;
}
Mapping of generated numbers to positions of 2-dimensional array is trivial.
Related
I'm kind of new to C# and I'm trying to make a sudoku game. I'm using a for loop to loop through the column arrays and add a random number to the spots but I'm using Random and Next() so it allows the numbers to repeat. But because I can't have more than one of a number in each column (array) for it to function, how can I replace the repeated number with another number that isn't repeated? I don't know how to do this.
Also the columns are stored in other arrays that I've called rows just so that I can say Row1[0][2] to access position 3 in column 1 for example.
Here's the method & for loop I'm using to replace add the numbers to the arrays:
void populateColumns(int[][] arr) // goes through column arrays and replaces the numbers with random ones from Num()
{
int i;
int j = 0;
for (int l = 0; l <= 8; l++)
{
for (i = 0; i <= 8; i++)
{
arr[j][i] = Num();
int currentPos = arr[j][i];
Console.Write(currentPos);
}
j = l;
}
}
My task is to make a organized shuffle, from source all odd numbers will go to left and even number will go to right.
I have done that much like this, and it is good for normal scenario:
public static string ShuffleChars(string source, int count)
{
if (string.IsNullOrWhiteSpace(source) || source.Length == 0)
{
throw new ArgumentException(null);
}
if (count < 0)
{
throw new ArgumentException(null);
}
for (int i = 0; i < count; i++)
{
source = string.Concat(source.Where((item, index) => index % 2 == 0)) +
string.Concat(source.Where((item, index) => index % 2 != 0));
}
return source;
}
Now the problem is, what if the count is int.MaxValue or a other huge number in millions, it will loop trough a lot. How can I optimize the code in terms of speed and resource consumption?
You should be able to determine by the string's length how many iterations it will take before it's back to it's original sort order. Then take the modulus of the iteration count and the input count, and only iterate that many times.
For example, a string that is three characters will be back to it's original sort order in 2 iterations. If the input count was to do 11 iterations, we know that 11 % 2 == 1, so we only need to iterate one time.
Once you determine a formula for how many iterations it takes to reach the original sort order for any length of string, you can always reduce the number of iterations to that number or less.
Coming up with a formula will be tricky, however. A string with 14 characters takes 12 iterations until it matches itself, but a string with 15 characters only takes 4 iterations.
Therefore, a shortcut might be to simply start iterating until we reach the original sort order (or the specified count, whichever comes first). If we reach the count first, then we return that answer. Otherwise, we can determine the answer from the idea in the first paragraph - take the modulus of the input count and the iteration count, and return that answer.
This would require that we store the values from our iterations (in a dictionary, for example) so we can retrieve a specific previous value.
For example:
public static string ShuffleChars(string source, int count)
{
string s = source;
var results = new Dictionary<int, string>();
for (int i = 0; i < count; i++)
{
s = string.Concat(s.Where((item, index) => index % 2 == 0)) +
string.Concat(s.Where((item, index) => index % 2 != 0));
// If we've repeated our original string, return the saved
// value of the input count modulus the current iteration
if (s == source)
{
return results[count % (i + 1) - 1];
}
// Otherwise, save the value for later
else
{
results[i] = s;
}
}
// If we get here it means we hit the requested count before
// ever returning to the original sort order of the input
return s;
}
Instead of creating new immutable strings on each loop, you could work with a mutable array of characters (char[]), and swap characters between places. This would be the most efficient in terms of memory consumption, but doing the swaps on a single array could be quite tricky. Using two arrays is much easier, because you can just copy characters from one array to the other, and at the end of each loop swap the two arrays.
One more optimization you could do is to work with the indices of the char array, instead of its values. I am not sure if this will make any difference in practice, since in modern 64 bit machines both char and int types occupy 8 bytes (AFAIK). It will surely make a difference on 32 bit machines though. Here is an implementation, with all these ideas put together:
public static string ShuffleChars(string source, int count)
{
if (source == null) throw new ArgumentNullException(nameof(source));
if (count < 0) throw new ArgumentOutOfRangeException(nameof(count));
// Instantiate the two arrays
int[] indices = new int[source.Length];
int[] temp = new int[source.Length];
// Initialize the indices array with incremented numbers
for (int i = 0; i < indices.Length; i++)
indices[i] = i;
for (int k = 0; k < count; k++)
{
// Copy the odds to the temp array
for (int i = 0, j = 0; j < indices.Length; i += 1, j += 2)
temp[i] = indices[j];
// Copy the evens to the temp array
int lastEven = (indices.Length >> 1 << 1) - 1;
for (int i = indices.Length - 1, j = lastEven; j >= 0; i -= 1, j -= 2)
temp[i] = indices[j];
// Swap the two arrays, using value tuples
(indices, temp) = (temp, indices);
}
// Map the indices to characters from the source string
return String.Concat(indices.Select(i => source[i]));
}
This question already has answers here:
How do I check if my array has repeated values inside it?
(8 answers)
Closed 5 years ago.
I would like to check the array I passed in the shuffle method for any duplicates.
I would like to also generate a random number that is the size of a.length.
The thing is that I can't figure out how to check if the array has duplicates and if it does it would generate another number until it is unique from the rest.
public int[] Shuffle(int[] a)
{
//check if the array has duplicates
for (int i = 0; i < a.Length; i++)
{
int curValue = random.Next(a.Length);
if(a.Contains(curValue))
{
curValue = random.Next(a.Length);
}
else
{
a[i] = curValue;
}
}
for (int i = 0; i < a.Length; i++)
{
int r = random.Next(a.Length);
int t = a[r];
a[r] = a[i];
a[i] = t;
}
return a;
}
Can anyone help me?
Here is a function that does what you ask as I presume. Include using System.Linq to run it. I went on the basis of your text, because the code did not make your question clear to me. If you meant something differently please clarify.
static int[] Shuffle(int[] a)
{
Random rnd = new Random();
//Remove duplicates from array
int[] distinct = a.Distinct().ToArray();
//Add the same amount of unique numbers that have been removed as duplicates
int len = a.Length - distinct.Length;
int[] newNumbers = new int[len];
int i = 0;
while(i < len)
{
newNumbers[i] = rnd.Next(a.Length); //NOTE: here i put in the length of array a, but with an int array this will always resolve in a shuffled array containing all digits from 0 to the length-1 of the array.
if (!distinct.Contains(newNumbers[i]) && !newNumbers.Take(i).Contains(newNumbers[i]))
{
++i;
}
}
//Randomize the array
int[] b = a.OrderBy(x => rnd.Next()).ToArray();
//Concatenate the two arrays and return it (shuffled numbers and new unique numbers)
return distinct.Concat(newNumbers).ToArray();
}
As I said in the note, if you remain with an integer array and only allow adding of new random replacement number lower than the length of the array, you and up with a shuffled array of all numbers from 0 to the array length.
Since a new array for newNumbers is generated, you also need to check if a newly generated number is not in that array.
i'm trying to get X number of random number (where X is a variable) between 0 and 100 and add them to a row in a DataGridView. I'm using the code below, but the problem is i need it to be impossible to have the same number twice. Is there a way make sure i get unique random numbers?
int X = 20;
Random random = new Random();
int randomrow = random.Next(0, 100);
for (int i = 0; i < X; i++)
{
int randomNumber = random.Next(0, 100);
data.Rows[randomrow][3] = randomNumber;
}
Thanks for the help!
Split your problem in two:
Create a list of X random, unique numbers. There are numerous ways to do that:
a. Create a list of all numbers 0-100 and then shuffle them. OR
b. Keep a hash set of all the numbers you already created (in addition to the list) and only add a new one if it has not been added before.
Afterwards, loop through the list and the data rows simultaneously and insert the values into the rows.
Here's the simple way to do it without creating a shuffle method for the List (though there are examples of that). Basically create a list of all possible integers, populate the list, then sort by new GUIDs.
int X = 20;
var RowNumbers = new List<int>();
for (int i = 0; i < X; i++)
RowNumbers.Add(i);
foreach (int i in RowNumbers.OrderBy(f => Guid.NewGuid()))
{
data.Rows[i][3] = i;
}
You would need to compare the numbers you have already used to the next one you get from random.Next. If you have already used it, pick another. I also like Heinzi's answer.
Here is algorithm without shuffling and previous result using:
var max = 100; // max available value
var count = 20; // number of random numbers
var random = new Random();
var rest = (double)max;
for (int i = 0; i < max; i++, rest--)
{
if (count / rest > random.NextDouble())
{
Console.WriteLine(i); // here is your random value
count--;
if (count == 0)
{
break;
}
}
}
Ok, so I am writing a Minesweeper app in WinForms. This is my first C# app so it has been a huge learning process. I want to randomly generate two numbers to put into two separate arrays, mineX & mineY. These two arrays are the coordinates for where all the mines are going to be placed on the board.
Here is the issue, I want to make sure that when it randomly generates these numbers, that X,Y will never be the same. I've done the research and know how to generate a single number and make each one unique. My issue is I want the combination of two random numbers to be unique.
Here is my code for the placeMines method.
private void placeMines()
{
Random rnd = new Random();
for (int i = 0; i < MINE_COUNT; i++)
{
bombX[i] = rnd.Next(0, BOARD_X);
bombY[i] = rnd.Next(0, BOARD_Y)'
}
}
This seems to be a simple issue to resolve, but I can't figure it out for the life of me.
Create a list of possible mine positions.
Choose a random item from the list, remove the item.
Repeat 2 until there're enough mines.
Sample code:
const int BOARD_X = 10;
const int BOARD_Y = 6;
const int MINE_COUNT = 30;
List<int> positions = Enumerable.Range(0, BOARD_X * BOARD_Y).ToList();
var rnd = new Random();
for (int i = 0; i < MINE_COUNT; i++) {
int index = rnd.Next(positions.Count);
int pos = positions[index];
positions.RemoveAt(index);
int x = pos % BOARD_X, y = pos / BOARD_X;
Console.WriteLine("({0}, {1})", x, y);
}
NOTE: If the number of positions is big, you can optimize algorithm for space by using bit vectors. I'll leave this as an exercise.
Generate all positions (let it be struct with two int fields x, y) in array and shuffle this array.
Random _rand = new Random();
Position[] allPositions = GenerateAllPositions();
for (int i=allPositions.Length-1; i>=1; i--)
{
int j = _rand.Next(0, i+1);
// now swap
Position tempPosition = allPositions[i];
allPositions[i] = allPositions[j];
allPositions[j] = tempPosition;
}
Now you can have temp int variable lastQueriedPositionIndex = 0, increase it by one when you pick any position. If lastQueriedPositionIndex == allPositions.Length make lastQueriedPositionIndex=0 and invoke this "swap" routine again. This is O(n) worst case solution instead of O(n^2) solution proposed by Athari - however in small board dimensions you won't see any diffrence.