Kth smallest element in a row-wise and column-wise sorted 2D array | Set 1

Given an n x n matrix, where every row and column is sorted in non-decreasing order. Find the kth smallest element in the given 2D array.
Example, 

Input:k = 3 and array =
        10, 20, 30, 40
        15, 25, 35, 45
        24, 29, 37, 48
        32, 33, 39, 50 
Output: 20
Explanation: The 3rd smallest element is 20 

Input:k = 7 and array =
        10, 20, 30, 40
        15, 25, 35, 45
        24, 29, 37, 48
        32, 33, 39, 50 
Output: 30

Explanation: The 7th smallest element is 30

Approach: So the idea is to find the kth minimum element. Each row and each column is sorted. So it can be thought as C sorted lists and the lists have to be merged into a single list, the kth element of the list has to be found out. So the approach is similar, the only difference is when the kth element is found the loop ends.
Algorithm:

  1. The idea is to use min heap. Create a Min-Heap to store the elements
  2. Traverse the first row from start to end and build a min heap of elements from first row. A heap entry also stores row number and column number.
  3. Now Run a loop k times to extract min element from heap in each iteration
  4. Get minimum element (or root) from Min-Heap.
  5. Find row number and column number of the minimum element.
  6. Replace root with the next element from same column and min-heapify the root.
  7. Print the last extracted element, which is the kth minimum element

Implementation:

C++

filter_none

edit
close

play_arrow

link
brightness_4
code

// kth largest element in a 2d array sorted row-wise and column-wise 
#include<iostream> 
#include<climits> 
using namespace std; 
  
// A structure to store entry of heap. The entry contains 
// value from 2D array, row and column numbers of the value 
struct HeapNode { 
    int val; // value to be stored 
    int r; // Row number of value in 2D array 
    int c; // Column number of value in 2D array 
}; 
  
// A utility function to swap two HeapNode items. 
void swap(HeapNode *x, HeapNode *y) { 
    HeapNode z = *x; 
    *x = *y; 
    *y = z; 
  
// A utility function to minheapify the node harr[i] of a heap 
// stored in harr[] 
void minHeapify(HeapNode harr[], int i, int heap_size) 
    int l = i*2 + 1; 
    int r = i*2 + 2; 
    int smallest = i; 
    if (l < heap_size && harr[l].val < harr[i].val) 
        smallest = l; 
    if (r < heap_size && harr[r].val < harr[smallest].val) 
        smallest = r; 
    if (smallest != i) 
    
        swap(&harr[i], &harr[smallest]); 
        minHeapify(harr, smallest, heap_size); 
    
  
// A utility function to convert harr[] to a max heap 
void buildHeap(HeapNode harr[], int n) 
    int i = (n - 1)/2; 
    while (i >= 0) 
    
        minHeapify(harr, i, n); 
        i--; 
    
  
// This function returns kth smallest element in a 2D array mat[][] 
int kthSmallest(int mat[4][4], int n, int k) 
    // k must be greater than 0 and smaller than n*n 
    if (k <= 0 || k > n*n) 
    return INT_MAX; 
  
    // Create a min heap of elements from first row of 2D array 
    HeapNode harr[n]; 
    for (int i = 0; i < n; i++) 
        harr[i] = {mat[0][i], 0, i}; 
    buildHeap(harr, n); 
  
    HeapNode hr; 
    for (int i = 0; i < k; i++) 
    
    // Get current heap root 
    hr = harr[0]; 
  
    // Get next value from column of root's value. If the 
    // value stored at root was last value in its column, 
    // then assign INFINITE as next value 
    int nextval = (hr.r < (n-1))? mat[hr.r + 1][hr.c]: INT_MAX; 
  
    // Update heap root with next value 
    harr[0] = {nextval, (hr.r) + 1, hr.c}; 
  
    // Heapify root 
    minHeapify(harr, 0, n); 
    
  
    // Return the value at last extracted root 
    return hr.val; 
  
// driver program to test above function 
int main() 
int mat[4][4] = { {10, 20, 30, 40}, 
                    {15, 25, 35, 45}, 
                    {25, 29, 37, 48}, 
                    {32, 33, 39, 50}, 
                }; 
cout << "7th smallest element is " << kthSmallest(mat, 4, 7); 
return 0; 

chevron_right


Java

filter_none

edit
close

play_arrow

link
brightness_4
code

// Java program for kth largest element in a 2d 
// array sorted row-wise and column-wise 
class GFG{
      
// A structure to store entry of heap.
// The entry contains value from 2D array,
// row and column numbers of the value 
static class HeapNode
{
      
    // Value to be stored 
    int val;
      
    // Row number of value in 2D array 
    int r;
      
    // Column number of value in 2D array 
    int c;
      
    HeapNode(int val, int r, int c)
    {
        this.val = val;
        this.c = c;
        this.r = r;
    }
}
  
// A utility function to swap two HeapNode items. 
static void swap(int i, int min, HeapNode[] arr)
{
    HeapNode temp = arr[i];
    arr[i] = arr[min];
    arr[min] = temp;
}
  
// A utility function to minheapify the node
// harr[i] of a heap stored in harr[] 
static void minHeapify(HeapNode harr[], 
                       int i, int heap_size)
{
    int l = 2 * i + 1;
    int r = 2 * i + 2;
    int min = i;
      
    if (l < heap_size && 
        harr[l].val < harr[i].val)
    {
        min = l;
    }
    if (r < heap_size && 
        harr[r].val < harr[i].val)
    {
        min = r;
    }
      
    if (min != i)
    {
        swap(i, min, harr);
        minHeapify(harr, min, heap_size);
    }
}
  
// A utility function to convert 
// harr[] to a max heap 
static void buildHeap(HeapNode harr[], int n)
{
    int i = (n - 1) / 2;
    while (i >= 0
    {
        minHeapify(harr, i, n);
        i--;
    }
}
  
// This function returns kth smallest
// element in a 2D array mat[][] 
public static int kthSmallest(int[][] mat,
                              int n, int k)
{
      
    // k must be greater than 0 and 
    // smaller than n*n 
    if (k <= 0 || k > n * n)
    {
        return Integer.MAX_VALUE;
    }
      
    // Create a min heap of elements 
    // from first row of 2D array 
    HeapNode harr[] = new HeapNode[n];
      
    for(int i = 0; i < n; i++)
    {
        harr[i] = new HeapNode(mat[0][i], 0, i);
    }
    buildHeap(harr, n);
      
    HeapNode hr = new HeapNode(0, 0, 0);
      
    for(int i = 1; i < k; i++)
    {
          
        // Get current heap root 
        hr = harr[0];
          
        // Get next value from column of root's
        // value. If the value stored at root was
        // last value in its column, then assign
        // INFINITE as next value 
        int nextVal = hr.r < n - 1
                      mat[hr.r + 1][hr.c] :
                      Integer.MAX_VALUE;
                        
        // Update heap root with next value 
        harr[0] = new HeapNode(nextVal, 
                               hr.r + 1, hr.c);
                                 
        // Heapify root 
        minHeapify(harr, 0, n);
    }
      
    // Return the value at last extracted root 
    return hr.val;
}
  
// Driver code
public static void main(String args[])
{
    int mat[][] = { { 10, 20, 30, 40 },
                    { 15, 25, 35, 45 },
                    { 25, 29, 37, 48 },
                    { 32, 33, 39, 50 } };
                      
    int res = kthSmallest(mat, 4, 7);
      
    System.out.print("7th smallest element is "+ res);
}
}
  
// This code is contributed by nikhil kumar sharma

chevron_right


Python3

filter_none

edit
close

play_arrow

link
brightness_4
code

# Program for kth largest element in a 2d array 
# sorted row-wise and column-wise 
from sys import maxsize 
  
# A structure to store an entry of heap. 
# The entry contains a value from 2D array, 
# row and column numbers of the value 
class HeapNode: 
    def __init__(self, val, r, c): 
        self.val = val # value to be stored 
        self.r = r # Row number of value in 2D array 
        self.c = c # Column number of value in 2D array 
  
# A utility function to minheapify the node harr[i] 
# of a heap stored in harr[] 
def minHeapify(harr, i, heap_size): 
    l = i * 2 + 1
    r = i * 2 + 2
    smallest =
    if l < heap_size and harr[l].val < harr[i].val: 
        smallest =
    if r < heap_size and harr[r].val < harr[smallest].val: 
        smallest =
  
    if smallest != i: 
        harr[i], harr[smallest] = harr[smallest], harr[i] 
        minHeapify(harr, smallest, heap_size) 
  
# A utility function to convert harr[] to a max heap 
def buildHeap(harr, n): 
    i = (n - 1) // 2
    while i >= 0
        minHeapify(harr, i, n) 
        i -= 1
  
# This function returns kth smallest element 
# in a 2D array mat[][] 
def kthSmallest(mat, n, k): 
  
    # k must be greater than 0 and smaller than n*n 
    if k <= 0 or k > n * n: 
        return maxsize 
  
    # Create a min heap of elements from 
    # first row of 2D array 
    harr = [0] *
    for i in range(n): 
        harr[i] = HeapNode(mat[0][i], 0, i) 
    buildHeap(harr, n) 
  
    hr = HeapNode(0, 0, 0
    for i in range(k): 
  
        # Get current heap root 
        hr = harr[0
  
        # Get next value from column of root's value. 
        # If the value stored at root was last value 
        # in its column, then assign INFINITE as next value 
        nextval = mat[hr.r + 1][hr.c] if (hr.r < n - 1) else maxsize 
  
        # Update heap root with next value 
        harr[0] = HeapNode(nextval, hr.r + 1, hr.c) 
  
        # Heapify root 
        minHeapify(harr, 0, n) 
  
    # Return the value at last extracted root 
    return hr.val 
  
# Driver Code 
if __name__ == "__main__"
    mat = [[10, 20, 30, 40], 
        [15, 25, 35, 45], 
        [25, 29, 37, 48], 
        [32, 33, 39, 50]] 
    print("7th smallest element is"
            kthSmallest(mat, 4, 7)) 
  
# This code is contributed by 
# sanjeev2552 

chevron_right


Output: 

7th smallest element is 30

Complexity Analysis:

  • Time Complexity: The above solution involves following steps. 
    1. Building a min-heap which takes O(n) time
    2. Heapify k times which takes O(k Logn) time.
  • Space Complexity: O(R), where R is the length of a row, as the Min-Heap stores one row at a time.

The above code can be optimized to build a heap of size k when k is smaller than n. In that case, the kth smallest element must be in first k rows and k columns. 
We will soon be publishing more efficient algorithms for finding the kth smallest element. 
This article is compiled by Ravi Gupta. Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above
 

Attention reader! Don’t stop learning now. Get hold of all the important DSA concepts with the DSA Self Paced Course at a student-friendly price and become industry ready.




My Personal Notes arrow_drop_up

Recommended Posts: