Given an array of size n and an integer k, return the count of distinct numbers in all windows of size k.
Example:
Input: arr[] = {1, 2, 1, 3, 4, 2, 3};
k = 4
Output: 3 4 4 3
Explanation:
First window is {1, 2, 1, 3}, count of distinct numbers is 3
Second window is {2, 1, 3, 4} count of distinct numbers is 4
Third window is {1, 3, 4, 2} count of distinct numbers is 4
Fourth window is {3, 4, 2, 3} count of distinct numbers is 3
Input: arr[] = {1, 2, 4, 4};
k = 2
Output: 2 2 1
Explanation:
First window is {1, 2}, count of distinct numbers is 2
First window is {2, 4}, count of distinct numbers is 2
First window is {4, 4}, count of distinct numbers is 1
Source: This problem has appeared in Microsoft Interview Question. Naive Approach: The naive solution is to traverse the given array considering every window in it and keeping a count on the distinct elements of the window.
Algorithm:
For every index i from 0 to len_array(n) – k, i.e n – k, traverse the array from i to i + k. This is the window
Traverse the window, from i to that index and check if the element is present or not.
If the element is not present in the prefix of the array, i.e no duplicate element is present from i to index-1, then increase the count.
Print the count.
Below is the implementation of the above approach:
C++
// Simple C++ program to count distinct
// elements in every window of size k
#include <bits/stdc++.h>
usingnamespacestd;
// Counts distinct elements in window of size k
intcountWindowDistinct(intwin[], intk)
{
intdist_count = 0;
// Traverse the window
for(inti = 0; i < k; i++) {
// Check if element arr[i] exists in arr[0..i-1]
intj;
for(j = 0; j < i; j++)
if(win[i] == win[j])
break;
if(j == i)
dist_count++;
}
returndist_count;
}
// Counts distinct elements in all windows of size k
voidcountDistinct(intarr[], intn, intk)
{
// Traverse through every window
for(inti = 0; i <= n - k; i++)
cout << countWindowDistinct(arr + i, k) << endl;
}
// Driver program
intmain()
{
intarr[] = { 1, 2, 1, 3, 4, 2, 3 }, k = 4;
intn = sizeof(arr) / sizeof(arr[0]);
countDistinct(arr, n, k);
return0;
}
Java
// Simple Java program to count distinct elements in every
// window of size k
importjava.util.Arrays;
classTest {
// Counts distinct elements in window of size k
staticintcountWindowDistinct(intwin[], intk)
{
intdist_count = 0;
// Traverse the window
for(inti = 0; i < k; i++) {
// Check if element arr[i] exists in arr[0..i-1]
intj;
for(j = 0; j < i; j++)
if(win[i] == win[j])
break;
if(j == i)
dist_count++;
}
returndist_count;
}
// Counts distinct elements in all windows of size k
staticvoidcountDistinct(intarr[], intn, intk)
{
// Traverse through every window
for(inti = 0; i <= n - k; i++)
System.out.println(countWindowDistinct(Arrays.copyOfRange(arr, i, arr.length), k));
Time complexity: O(nk2). The time complexity can be improved to O(n*k*log k) by modifying countWindowDistinct() by the use of sorting. The function can further be optimized to use hashing to find distinct elements in a window. With hashing the time complexity becomes O(n*k).
Space Complexity: O(1) No extra space is needed.
Efficient Approach: So, there is an efficient solution using hashing, though hashing requires extra O(n) space but the time complexity will improve. The trick is to use the count of the previous window while sliding the window. To do this a hash map can be used that stores elements of the current window. The hash-map is also operated on by simultaneous addition and removal of an element while keeping track of distinct elements. The problem deals with finding the count of distinct elements in a window of length k, at any step while shifting the window and discarding all the computation done in the previous step, even though k – 1 elements are same from the previous adjacent window. For example, assume that elements from index i to i + k – 1 are stored in a Hash Map as an element-frequency pair. So, while updating the Hash Map in range i + 1 to i + k, reduce the frequency of the i-th element by 1 and increase the frequency of (i + k)-th element by 1. Insertion and deletion from the HashMap takes constant time.
Algorithm:
Create an empty hash map. Let the hash map be hM.
Initialize the count of distinct elements as dist_count to 0.
Traverse through the first window and insert elements of the first window to hM. The elements are used as key and their counts as the value in hM. Also, keep updating dist_count
Print distinct count for the first window.
Traverse through the remaining array (or other windows).
Remove the first element of the previous window.
If the removed element appeared only once, remove it from hM and decrease the distinct count, i.e. do “dist_count–“
else (appeared multiple times in hM), then decrement its count in hM
Add the current element (last element of the new window)
If the added element is not present in hM, add it to hM and increase the distinct count, i.e. do “dist_count++”
Else (the added element appeared multiple times), increment its count in hM
Below is a dry run of the above approach:
Implementation:
C++
// An efficient C++ program to
// count distinct elements in
// every window of size k
#include <iostream>
#include <unordered_map>
usingnamespacestd;
voidcountDistinct(intarr[], intk, intn)
{
// Creates an empty hashmap hm
unordered_map<int, int> hm;
// initialize distinct element count for current window
intdist_count = 0;
// Traverse the first window and store count
// of every element in hash map
for(inti = 0; i < k; i++) {
if(hm[arr[i]] == 0) {
dist_count++;
}
hm[arr[i]] += 1;
}
// Print count of first window
cout << dist_count << endl;
// Traverse through the remaining array
for(inti = k; i < n; i++) {
// Remove first element of previous window
// If there was only one occurrence, then reduce distinct count.
if(hm[arr[i - k]] == 1) {
dist_count--;
}
// reduce count of the removed element
hm[arr[i - k]] -= 1;
// Add new element of current window
// If this element appears first time,
// increment distinct element count
if(hm[arr[i]] == 0) {
dist_count++;
}
hm[arr[i]] += 1;
// Print count of current window
cout << dist_count << endl;
}
}
intmain()
{
intarr[] = { 1, 2, 1, 3, 4, 2, 3 };
intsize = sizeof(arr) / sizeof(arr[0]);
intk = 4;
countDistinct(arr, k, size);
return0;
}
// This solution is contributed by Aditya Goel
Java
// An efficient Java program to count distinct elements in
// Javascript program to count distinct elements in
// every window of size k
functioncountDistinct(arr, k)
{
// Creates an empty hashMap hM
let hM = newMap();
// Traverse the first window and store count
// of every element in hash map
for(let i = 0; i < k; i++)
{
if(hM.has(arr[i]))
hM.set(arr[i], hM.get(arr[i])+1)
else
hM.set(arr[i], 1);
}
// Print count of first window
document.write(hM.size + "<br/>");
// Traverse through the remaining array
for(let i = k; i < arr.length; i++) {
// Remove first element of previous window
// If there was only one occurrence
if(hM.get(arr[i - k]) == 1) {
hM.delete(arr[i - k]);
}
else// reduce count of the removed element
hM.set(arr[i - k], hM.get(arr[i - k]) - 1);
// Add new element of current window
// If this element appears first time,
// set its count as 1,
if(hM.has(arr[i]))
hM.set(arr[i], hM.get(arr[i])+1)
else
hM.set(arr[i], 1);
// Print count of current window
document.write(hM.size + "<br/>");
}
}
// Driver code
let arr = [ 1, 2, 1, 3, 4, 2, 3 ];
let size = arr.length;
let k = 4;
countDistinct(arr, k, size);
// This code is contributed by splevel62.
</script>
Output
3
4
4
3
Complexity Analysis:
Time complexity O(n). A single traversal of the array is required.
Space Complexity O(n). Since the hashmap requires linear space.
This article is contributed by Piyush. Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.
We use cookies to ensure you have the best browsing experience on our website. By using our site, you
acknowledge that you have read and understood our
Cookie Policy &
Privacy Policy