# Queries to insert, delete one occurrence of a number and print the least and most frequent element

Given Q queries of type 1, 2, 3 and 4 as described below.

**Type-1:**Insert a number to the list.**Type-2:**Delete only one occurrence of a number if exists.**Type-3:**Print the least frequent element, if multiple elements exist then print the greatest among them.**Type-4:**Print the most frequent element, if multiple elements exist then print the smallest among them.

The task is to write a program to perform all the above queries.

**Examples:**

Input:

Query1: 1 6

Query2: 1 6

Query3: 1 7

Query4: 3

Query5: 1 7

Query6: 2 7

Query7: 1 7

Query8: 3

Query9: 4

Output:

7

7

6While answering Query4, the frequency of 6 is 2 and that of

7 is 1, hence the least frequent element is 7.

In Query8, the least frequent element is 6 and 7, so print the largest.

In Query9, the most frequent element is 6 and 7, so print the smallest.

A **naive approach** is to use any Data-Structures(array, vector, ..) and store all the elements. Using a hash-table, the frequency of every element can be stored. While processing the Query of type-2, delete one occurrence of that element from the DS in which the element has been stored. The queries of type-3 and type-4 can be answered by traversing the hash-table. The time complexity will be **O(N)** per query, where N is the number of elements till then in the DS.

An **efficient approach** will be to use set container to answer every query. Using two sets, one hash-table the above problem can be solved in **O(log n)** per query. Two sets *s1* and *s2* are used, one stores the * {num, frequency}*, while the other stores the

*. A hash-map is used which stores the frequency of each number. Design the set s2 using operator overloading such that it is sorted in ascending order of the first elements. If the first element appears to be same of one or more element, the set will be sorted in descending order of the second elements. The user-defined operating-overloading function thus will be:*

**{frequency, number}**bool operator<(pr a, pr b) { if(a.first == b.first) return a.second > b.second; return a.first < b.first; }Note:Operator overloading only works with user-defined data-types.pris a struct which has first and second as two integers.

Below is the algorithm to solve query of every type:

**Type1:**Check using hash-table if the element exists. If it does not exist, then mark the number in hash-table. Insert**{num, 1}**in set s1 and**{1, num}**in set2 using insert(). If it exists previously, then get the frequency from hash-table and delete**{num, frequency}**from set1 and**{frequency, num}**from set2 using find() and erase() function. Insert**{num, frequency+1}**in set1 and**{frequency+1, num}**in set2. Also, increase the count in hash-table.**Type2:**Follow the same process as above in query type1. Ony difference is to decrease the count in hash-table and insert**{num, frequency-1}**in set1 and**{frequency-1, num}**in set2.**Type3:**Print the beginning element which can be obtained using begin() function, as the set has been designed in such a way that begin() returns the least frequent element. If there are more than one, then it returns the largest.**Type4:**Print the last element in the set which can be obtained using rbegin() function in set.

Below is the implementation of the above approach:

`// C++ program for performing ` `// Queries of insert, delete one ` `// occurrence of a number and ` `// print the least and most frequent element ` `#include <bits/stdc++.h> ` `using` `namespace` `std; ` ` ` `// user-defined data-types ` `struct` `pr { ` ` ` `int` `first; ` ` ` `int` `second; ` `}; ` ` ` `// user-defined function to ` `// design a set ` `bool` `operator<(pr a, pr b) ` `{ ` ` ` `if` `(a.first == b.first) ` ` ` `return` `a.second > b.second; ` ` ` `return` `a.first < b.first; ` `} ` ` ` `// declare a user-defined set ` `set<pr> s1, s2; ` ` ` `// hash map ` `unordered_map<` `int` `, ` `int` `> m; ` ` ` `// Function to process the query ` `// of type-1 ` `void` `type1(` `int` `num) ` `{ ` ` ` ` ` `// if the element is already there ` ` ` `if` `(m[num]) { ` ` ` ` ` `// get the frequency of the element ` ` ` `int` `cnt = m[num]; ` ` ` ` ` `// returns an iterator pointing to ` ` ` `// position where the pair is ` ` ` `auto` `it1 = s1.find({ num, cnt }); ` ` ` `auto` `it2 = s2.find({ cnt, num }); ` ` ` ` ` `// deletes the pair from sets ` ` ` `s1.erase(it1); ` ` ` `s2.erase(it2); ` ` ` ` ` `// re-insert the pair by increasing ` ` ` `// frequency ` ` ` `s1.insert({ num, m[num] + 1 }); ` ` ` `s2.insert({ m[num] + 1, num }); ` ` ` `} ` ` ` ` ` `// if the element is not there in the list ` ` ` `else` `{ ` ` ` ` ` `// insert the element with frequency 1 ` ` ` `s1.insert({ num, 1 }); ` ` ` `s2.insert({ 1, num }); ` ` ` `} ` ` ` ` ` `// increase the count in hash-table ` ` ` `m[num] += 1; ` `} ` ` ` `// Function to process the query ` `// of type-2 ` `void` `type2(` `int` `num) ` `{ ` ` ` `// if the element exists ` ` ` `if` `(m[num]) { ` ` ` ` ` `// get the frequency of the element ` ` ` `int` `cnt = m[num]; ` ` ` ` ` `// returns an iterator pointing to ` ` ` `// position where the pair is ` ` ` `auto` `it1 = s1.find({ num, cnt }); ` ` ` `auto` `it2 = s2.find({ cnt, num }); ` ` ` ` ` `// deletes the pair from sets ` ` ` `s1.erase(it1); ` ` ` `s2.erase(it2); ` ` ` ` ` `// re-insert the pair by increasing ` ` ` `// frequency ` ` ` `s1.insert({ num, m[num] - 1 }); ` ` ` `s2.insert({ m[num] - 1, num }); ` ` ` ` ` `// decrease the count ` ` ` `m[num] -= 1; ` ` ` `} ` `} ` ` ` `// Function to process the query ` `// of type-3 ` `int` `type3() ` `{ ` ` ` `// if the set is not empty ` ` ` `// return the first element ` ` ` `if` `(!s1.empty()) { ` ` ` `auto` `it = s2.begin(); ` ` ` `return` `it->second; ` ` ` `} ` ` ` ` ` `else` ` ` `return` `-1; ` `} ` ` ` `// Function to process the query ` `// of type-4 ` `int` `type4() ` `{ ` ` ` ` ` `// if the set is not empty ` ` ` `// return the last element ` ` ` `if` `(!s1.empty()) { ` ` ` `auto` `it = s2.rbegin(); ` ` ` `return` `it->second; ` ` ` `} ` ` ` ` ` `else` ` ` `return` `-1; ` `} ` `// Driver Code ` `int` `main() ` `{ ` ` ` ` ` `// Queries ` ` ` ` ` `// inserts 6, 6 and 7 ` ` ` `type1(6); ` ` ` `type1(6); ` ` ` `type1(7); ` ` ` ` ` `// print the answer to query of type3 ` ` ` `cout << type3() << endl; ` ` ` ` ` `// inserts 7 ` ` ` `type1(7); ` ` ` ` ` `// deletes one occurence of 7 ` ` ` `type2(7); ` ` ` ` ` `// inserts 7 ` ` ` `type1(7); ` ` ` ` ` `// print the answer to query of type3 ` ` ` `cout << type3() << endl; ` ` ` ` ` `// print the answer to query of type4 ` ` ` `cout << type4() << endl; ` ` ` ` ` `return` `0; ` `} ` |

*chevron_right*

*filter_none*

**Output:**

7 7 6

**Time Complexity:** O(log N) per query.

**Auxiliary Space:** O(N)

## Recommended Posts:

- Most frequent element in an array
- Smallest subarray with all occurrences of a most frequent element
- Design a data structure that supports insert, delete, search and getRandom in constant time
- Print characters having odd frequencies in order of occurrence
- Print the last occurrence of elements in array in relative order
- Group multiple occurrence of array elements ordered by first occurrence
- How to delete last element from a set in C++
- How to delete last element from a map in C++
- How to delete last element from a List in C++ STL
- How to delete an element from the Set by passing its value in C++
- Queries on insertion of an element in a Bitonic Sequence
- Number of comparisons in each direction for m queries in linear search
- Minimum number of operations required to delete all elements of the array
- Number of elements greater than K in the range L to R using Fenwick Tree (Offline queries)
- Find top k (or most frequent) numbers in a stream

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.