Find k closest numbers in an unsorted array

Given an unsorted array and two numbers x and k, find k closest values to x.

Examples:

Input : arr[] = {10, 2, 14, 4, 7, 6}, x = 5, k = 3 
Output : 4 6 7
Three closest values of x are 4, 6 and 7.

Input : arr[] = {-10, -50, 20, 17, 80}, x = 20, k = 2
Output : 17, 20

A simple solution is to sort the array. Then apply the method discussed to k closest values in a sorted array.



Time Complexity : O(n Log n)

A better solution is to use Heap Data Structure
1) Make a max heap of differences with first k elements.
2) For every element starting from (k+1)-th element, do following.
…..a) Find difference of current element with x.
…..b) If difference is more than root of heap, ignore current element.
…..c) Else insert the current element to the heap after removing the root.
3) Finally the heap has k closest elements.

C++

filter_none

edit
close

play_arrow

link
brightness_4
code

// C++ program to find k closest elements
#include <bits/stdc++.h>
using namespace std;
  
void printKclosest(int arr[], int n, int x,
                   int k)
{
    // Make a max heap of difference with
    // first k elements.
    priority_queue<pair<int, int> > pq;
    for (int i = 0; i < k; i++)
        pq.push({ abs(arr[i] - x), i });
  
    // Now process remaining elements.
    for (int i = k; i < n; i++) {
  
        int diff = abs(arr[i] - x);
  
        // If difference with current
        // element is more than root,
        // then ignore it.
        if (diff > pq.top().first)
            continue;
  
        // Else remove root and insert
        pq.pop();
        pq.push({ diff, i });
    }
  
    // Print contents of heap.
    while (pq.empty() == false) {
        cout << arr[pq.top().second] << " ";
        pq.pop();
    }
}
  
// Driver program to test above functions
int main()
{
    int arr[] = { -10, -50, 20, 17, 80 };
    int x = 20, k = 2;
    int n = sizeof(arr) / sizeof(arr[0]);
    printKclosest(arr, n, x, k);
    return 0;
}

chevron_right






                        filter_none









Python3














                                    filter_none




                                    edit

                                    close




                                    play_arrow




                                    link

                                    brightness_4

                                    code
                                




















# Python3 program to find k closest elements  


import math 


import sys 


from queue import PriorityQueue 


def printKclosest(arr,n,x,k): 


  


    # Make a max heap of difference with  


    # first k elements.  


    pq = PriorityQueue() 


    for i in range(k): 


        pq.put((-abs(arr[i]-x),i)) 


  


    # Now process remaining elements 


    for i in range(k,n): 


        diff = abs(arr[i]-x) 


        p,pi = pq.get() 


        curr = -p 


  


        # If difference with current  


        # element is more than root,  


        # then put it back.  


        if diff>curr: 


            pq.put((-curr,pi)) 


            continue


        else: 


  


            # Else remove root and insert 


            pq.put((-diff,i)) 


              


    # Print contents of heap. 


    while(not pq.empty()): 


        p,q = pq.get() 


        print("{} ".format(arr[q]),end = "") 


  


# Driver program to test above functions  


if __name__=='__main__': 


    arr = [-10,-50,20,17,80] 


    x,k = 20,2


    n = len(arr) 


    printKclosest(arr, n, x, k) 


  


# This code is contributed by Vikash Kumar 37 



















                        chevron_right







                        filter_none











Output:


17 20





Time Complexity : O(n Log k)


GeeksforGeeks has prepared a complete interview preparation course with premium videos, theory, practice problems, TA support and many more features. Please refer Placement 100 for details 





    

        My Personal Notes
        arrow_drop_up
    

    

        
        

            
            
        

    


Recommended Posts:
kartik
Check out this Author's contributed articles.
If you like GeeksforGeeks and would like to contribute, you can also write an article using contribute.geeksforgeeks.org or mail your article to contribute@geeksforgeeks.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.
Please Improve this article if you find anything incorrect by clicking on  the "Improve Article" button below.