Reputation: 23
I've searched the Internet and known how to delete an element (with std::erase
) and finding duplicates of an element to then delete it (vec.erase(std::unique(vec.begin(), vec.end()),vec.end());
). But all methods only delete either an element or its duplicates.
I want to delete both.
For example, using this vector:
std::vector<int> vec = {2,3,1,5,2,2,5,1};
I want output to be:
{3}
My initial idea was:
void removeDuplicatesandElement(std::vector<int> &vec)
{
std::sort(vec.begin(), vec.end());
int passedNumber = 0; //To tell amount of number not deleted (since not duplicated)
for (int i = 0; i != vec.size(); i = passedNumber) //This is not best practice, but I tried
{
if (vec[i] == vec[i+1])
{
int ctr = 1;
for(int j = i+1; j != vec.size(); j++)
{
if (vec[j] == vec[i]) ctr++;
else break;
}
vec.erase(vec.begin()+i, vec.begin()+i+ctr);
}
else passedNumber++;
}
}
And it worked. But this code is redundant and runs at O(n^2), so I'm trying to find other ways to solve the problem (maybe an STL function that I've never heard of, or just improve the code).
Upvotes: 0
Views: 346
Reputation: 177
From what you have searched, we can look in the vector for duplicated values, then use the Erase–remove idiom to clean up the vector.
#include <vector>
#include <algorithm>
#include <iostream>
void removeDuplicatesandElement(std::vector<int> &vec)
{
std::sort(vec.begin(), vec.end());
if (vec.size() < 2)
return;
for (int i = 0; i < vec.size() - 1;)
{
// This is for the case we emptied our vector
if (vec.size() < 2)
return;
// This heavily relies on the fact that this vector is sorted
if (vec[i] == vec[i + 1])
vec.erase(std::remove(vec.begin(), vec.end(), (int)vec[i]), vec.end());
else
i += 1;
}
// Since all duplicates are removed, the remaining elements in the vector are unique, thus the size of the vector
// But we are not returning anything or any reference, so I'm just gonna leave this here
// return vec.size()
}
int main()
{
std::vector<int> vec = {2, 3, 1, 5, 2, 2, 5, 1};
removeDuplicatesandElement(vec);
for (auto i : vec)
{
std::cout << i << " ";
}
std::cout << "\n";
return 0;
}
Output: 3
Time complexity: O(n)
Upvotes: 1
Reputation: 453
You can do it using STL maps as follows:
#include <iostream>
#include <vector>
#include <unordered_map>
using namespace std;
void retainUniqueElements(vector<int> &A){
unordered_map<int, int> Cnt;
for(auto element:A) Cnt[element]++;
A.clear(); //removes all the elements of A
for(auto i:Cnt){
if(i.second == 1){ // that if the element occurs once
A.push_back(i.first); //then add it in our vector
}
}
}
int main() {
vector<int> vec = {2,3,1,5,2,2,5,1};
retainUniqueElements(vec);
for(auto i:vec){
cout << i << " ";
}
cout << "\n";
return 0;
}
Output:
3
Time Complexity of the above approach: O(n)
Space Complexity of the above approach: O(n)
Upvotes: 1
Reputation: 52471
Something like this, perhaps:
void removeDuplicatesandElement(std::vector<int> &vec) {
if (vec.size() <= 1) return;
std::sort(vec.begin(), vec.end());
int cur_val = vec.front() - 1;
auto pred = [&](const int& val) {
if (val == cur_val) return true;
cur_val = val;
// Look ahead to the next element to see if it's a duplicate.
return &val != &vec.back() && (&val)[1] == val;
};
vec.erase(std::remove_if(vec.begin(), vec.end(), pred), vec.end());
}
This relies heavily on the fact that std::vector
is guaranteed to have contiguous storage. It won't work with any other container.
Upvotes: 2