没有用于调用std :: vector的匹配函数 <std::tuple> 推回

Question

I have a sample program containing 6 timepoints using high_resolution_clock::now() from standard chrono header. I take differences b/w each of them resulting in 3 differences and caste them as auto duration1 = std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 ).count(); to microseconds.

I have another variable named durations which is assigned as follows: auto durations = std::make_tuple(duration1,duration2,duration3); containing previous time-point differences.

I have to push this tuple into an vector, so I have introduced std::vector<std::tuple<std::chrono::microseconds,std::chrono::microseconds,std::chrono::microseconds>> list; However on using list.push_back(durations); I get an error as :

prog.cpp: In function 'int main()':
prog.cpp:36:29: error: no matching function for call to 'std::vector<std::tuple<std::chrono::duration<long long int, std::ratio<1ll, 1000000ll> >, std::chrono::duration<long long int, std::ratio<1ll, 1000000ll> >, std::chrono::duration<long long int, std::ratio<1ll, 1000000ll> > > >::push_back(std::tuple<long long int, long long int, long long int>&)'
     list.push_back(durations);

---

I tried to search about std::chrono::microseconds and other std::chrono::duration stuff here but wasn't successful in rectifying the problem.

I know this has something to do with my negligence of type system, but I'm unable to locate that error. Any help would be appreciated, & here is ideone link .

#include <iostream>
#include <chrono>
#include <vector>
#include <tuple>

using namespace std;
using namespace std::chrono;

void function()
{
    long long number = 0;

    for( long long i = 0; i != 2000000; ++i )
    {
       number += 5;
    }
}

int main()
{
    high_resolution_clock::time_point t1 = high_resolution_clock::now();
    high_resolution_clock::time_point t3 = high_resolution_clock::now();
    high_resolution_clock::time_point t5 = high_resolution_clock::now();
    function();
    high_resolution_clock::time_point t2 = high_resolution_clock::now();
    high_resolution_clock::time_point t4 = high_resolution_clock::now();
    high_resolution_clock::time_point t6 = high_resolution_clock::now();

    auto duration1 = std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 ).count();
    auto duration2 = std::chrono::duration_cast<std::chrono::microseconds>( t4 - t3 ).count();
    auto duration3 = std::chrono::duration_cast<std::chrono::microseconds>( t6 - t5 ).count();

    auto durations = std::make_tuple(duration1,duration2,duration3);

    std::vector<std::tuple<std::chrono::microseconds,std::chrono::microseconds,std::chrono::microseconds>> list;
    list.push_back(durations);

    cout << duration1 << " -- "<< duration2 << " -- "<< duration3 << " -- ";
    return 0;
}

Answer 1

You have created a tuple of 3 integers and you're trying to add it to a vector of 3 durations.

I take differences b/w each of them resulting in 3 differences and caste them as auto duration1 = std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 ).count(); to microseconds.

Why are you calling count() on the durations after doing the duration_cast to convert to microseconds?

Just keep the values as microseconds objects and you can add them to the vector:

auto duration1 = std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 );
auto duration2 = std::chrono::duration_cast<std::chrono::microseconds>( t4 - t3 );
auto duration3 = std::chrono::duration_cast<std::chrono::microseconds>( t6 - t5 );

Answer 2

The type of std::chrono::microseconds::count() is not std::chrono::microseconds , it's some signed integral type.

auto duration1 = std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 ).count();
//decltype(duration1) is not std::chrono::microseconds

As such, you can't use your duration*n* variables for a vector expecting microsecond values.

The fix is easy, just defer your call of count until you try and print the contents.

Answer 3

It's simple: don't call count .

std::chrono::microseconds is (in your case) a typedef for the type std::chrono::duration<long long int, std::ratio<1ll, 1000000ll> > . That is also the type you get from doing std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 ) .

However, that is not what you're assigning to duration1 . You're assigning the result of calling the count function on that type. And that returns the number of ticks as a number ( long long int in your standard library's case), and not as duration .

Answer 4

You are getting a type mismatch.

auto duration1 = std::chrono::duration_cast<std::chrono::microseconds>( t2 - t1 ).count()

Actually gives you a long long in your case and not a std::chrono::microseconds . You can fix this by using decltype() and changing

std::vector<std::tuple<std::chrono::microseconds,std::chrono::microseconds,std::chrono::microseconds>> list;

To

std::vector<decltype(durations)> list;

Live Example