for user defined struct, as I understand, it's easy. Just overload the operator <. However, for int/float etc.., do I really need to overload operator < for int? Here is what I tried:
#include <iostream>
#include <algorithm>
#include <vector>
using namespace std;
bool comp(const int& a, const int& b)
{
return a<b?false:true;
}
int main ()
{
int myints[] = {10,20,30,5,15};
vector<int> v(myints,myints+5);
vector<int>::iterator it;
make_heap(v.begin(), v.end(), comp);
cout << "initial min heap : " << v.front() << endl;
for (unsigned i=0; i<v.size(); i++) cout << " " << v[i];
cout<<endl;
pop_heap (v.begin(),v.end());
v.pop_back();
for (unsigned i=0; i<v.size(); i++) cout << " " << v[i];
cout<<endl;
}
the results are:
initial min heap : 5
5 10 30 20 15
30 10 15 20
now pop_heap, push_heap won't maintain the min-heap correctly? is there any easier way to achieve this? Thanks!
Edit: sorry, I didn't check the manual carefully. yes, passing comp to pop_heap or push_heap should do the trick. However, what do you mean, I should not use an external comparator? If it's not the right way, what's the common way to achieve this?
Use std::greater<int>()
as the comparator(to all of make_heap
, push_heap
, pop_heap
). The ()
are important - std::greater<int>
is a functor class not a function, so you need an instance of it.