Intersection of n sets

Given n sets of integers of different sizes. Each set may contain duplicates also. How to find the intersection of all the sets. If an element is present multiple times in all the sets, it should be added that many times in the result.

For example, consider there are three sets {1,2,2,3,4} {2,2,3,5,6} {1,3,2,2,6}. The intersection of the given sets should be {2,2,3}

The following is an efficient approach to solve this problem.

1. Sort all the sets.
2. Take the smallest set, and insert all the elements, and their frequencies into a map.
3. For each element in the map do the following
…..a. If the element is not present in any set, ignore it
…..b. Find the frequency of the element in each set (using binary search). If it less than the frequency in the map, update the frequency
…..c. If the element is found in all the sets, add it to the result.

Here is the C++ implementation of the above approach.





// C++ program to find intersection of n sets
#include <iostream>
#include <vector>
#include <algorithm>
#include <map>
using namespace std;
// The main function that receives a set of sets as parameter and
// returns a set containing intersection of all sets
vector <int> getIntersection(vector < vector <int> > &sets)
    vector <int> result;  // To store the reaultant set
    int smallSetInd = 0;  // Initialize index of smallest set
    int minSize = sets[0].size(); // Initialize size of smallest set
    // sort all the sets, and also find the smallest set
    for (int i = 1 ; i < sets.size() ; i++)
        // sort this set
        sort(sets[i].begin(), sets[i].end());
        // update minSize, if needed
        if (minSize > sets[i].size())
            minSize = sets[i].size();
            smallSetInd = i;
    map<int,int> elementsMap;
    // Add all the elements of smallest set to a map, if already present,
    // update the frequency
    for (int i = 0; i < sets[smallSetInd].size(); i++)
        if (elementsMap.find( sets[smallSetInd][i] ) == elementsMap.end())
            elementsMap[ sets[smallSetInd][i] ] = 1;
            elementsMap[ sets[smallSetInd][i] ]++;
    // iterate through the map elements to see if they are present in
    // remaining sets
    map<int,int>::iterator it;
    for (it = elementsMap.begin(); it != elementsMap.end(); ++it)
        int elem = it->first;
        int freq = it->second;
        bool bFound = true;
        // Iterate through all sets
        for (int j = 0 ; j < sets.size() ; j++)
            // If this set is not the smallest set, then do binary search in it
            if (j != smallSetInd)
                // If the element is found in this set, then find its frequency
                if (binary_search( sets[j].begin(), sets[j].end(), elem ))
                   int lInd = lower_bound(sets[j].begin(), sets[j].end(), elem)
                                                            - sets[j].begin();
                   int rInd = upper_bound(sets[j].begin(), sets[j].end(), elem)
                                                            - sets[j].begin();
                   // Update the minimum frequency, if needed
                   if ((rInd - lInd) < freq)
                       freq = rInd - lInd;
                // If the element is not present in any set, then no need 
                // to proceed for this element.
                    bFound = false;
        // If element was found in all sets, then add it to result 'freq' times
        if (bFound)
            for (int k = 0; k < freq; k++)
    return result;
// A utility function to print a set of elements
void printset(vector <int> set)
    for (int i = 0 ; i < set.size() ; i++)
        cout<<set[i]<<" ";
// Test case
void TestCase1()
    vector < vector <int> > sets;
    vector <int> set1;
    vector <int> set2;
    vector <int> set3;
    vector <int> r = getIntersection(sets);
// Driver program to test above functions
int main()
    return 0;



1 1 5

Time Complexity: Let there be ‘n’ lists, and average size of lists be ‘m’. Sorting phase takes O( n* m *log m) time (Sorting n lists with average length m). Searching phase takes O( m * n * log m) time . ( for each element in a list, we are searching n lists with log m time). So the overall time complexity is O( n*m*log m ).

Auxiliary Space: O(m) space for storing the map.

This article is compiled by Ravi Chandra Enaganti. Please write comments if you find anything incorrect, or you want to share more information about the topic discussed above.

My Personal Notes arrow_drop_up
Article Tags :
Practice Tags :

Be the First to upvote.

Please write to us at to report any issue with the above content.