It isn't all to hard to do what you are looking for:
main.cpp
#include <iostream>
#include <vector>
#include <algorithm>
#include <random>
#include "Point.h"
int main() {
std::random_device rd; // Random Device: Used To Seed Mersenne Random Generator
std::mt19937 gen; // Mersenne Twister
gen.seed( rd() ); // Seed The Generator
std::uniform_int_distribution<> dist(0, 100); // Uniform Int Distribution between [a, max]
// Point<int>
std::vector<Point<int>> points;
points.reserve( NUM_POINTS );
for ( std::size_t i = 0; i < NUM_POINTS; i++ ) {
// Instead of creating a temporary stack copy each iteration
// I chose to use the constructor directly and instead of
// push_back, I'm using emplace_back.
// Point<int> p( dist( gen ), dist( gen ) );
// points.push_back( p );
points.emplace_back( Point<int>( dist( gen ), dist( gen ) ) );
}
std::cout << "Showing 10 points of type int with random (x,y):\n";
for ( auto& p : points ) {
std::cout << p;
}
std::cout << std::endl;
// Point<float>
std::vector<Point<float>> points2;
points.reserve( NUM_POINTS );
std::uniform_real_distribution<float> dist2( 0, 100.0f );
for ( std::size_t i = 0; i < NUM_POINTS; i++ ) {
// Instead of creating a temporary stack copy each iteration
// I chose to use the constructor directly and instead of
// push_back, I'm using emplace_back.
// Point<float> p( dist( gen ), dist( gen ) );
// points2.push_back( p );
points2.emplace_back( Point<float>( dist( gen ), dist( gen ) ) );
}
std::cout << "Showing 10 points of type float with random (x,y):\n";
for ( auto& p : points2 ) {
std::cout << p;
}
std::cout << std::endl;
// Sorting the containers:
std::sort( points.begin(), points.end() );
std::sort( points2.begin(), points2.end() );
std::cout << "Showing the sorted points with type int (x,y):\n";
for ( auto& p : points ) {
std::cout << p;
}
std::cout << std::endl;
std::cout << "Showing the sorted points with type float (x,y):\n";
for ( auto& p : points2 ) {
std::cout << p;
}
std::cout << std::endl;
std::cout << std::endl;
system( "PAUSE" );
return 0;
}
Point.h
#ifndef POINT_H
#define POINT_H
#include <iostream>
#include <tuple> // std::tie
const std::size_t NUM_POINTS { 10 };
// Need class Point prototype for operator<< declaration
template<class> class Point;
// Need operator<< declaration for class template Point's friend declaration
template<class T>
std::ostream& operator<<( std::ostream& out, const Point<T>& );
// Class Declaration & Definition
template<class T>
class Point {
public:
T _x;
T _y;
Point() : _x( 0 ), _( 0 ) {}
Point( T x, T y ) : _x( x ), _y( y ) {}
Point( T& x, T& y ) : _x( x ), _y( y ) {}
Point( T* x, T* y ) : _x( *x ), _y( *y ) {}
// friend prototype: notice the extra <> in this declaration
// It tells the compiler that this friend function will be a specialization of this class template
friend std::ostream& operator<< <>( std::ostream& out, const Point<T>& p );
// operator< for comparison
bool operator<( Point<T>& p ) {
// std::tie makes it real easy to compare a (set) of values.
return std::tie( _x, _y ) < std::tie( p._x, p._y );
}
// operator> for comparison
bool operator<( Point<T>& p ) {
return !(*this < p );
}
// operator== for comparison
bool operator==( Point<T>& p ) {
return (this->_x == p._x && this->y == p._y );
}
};
// operator<< definition
template<class T>
std::ostream& operator<<( std::ostream& out, const Point<T>& p ) {
return out << "(" << p._x << "," << p._y << ")\n";
}
#endif // !POINT_H
As for the implementation of the class template Point<T> you can refer to the comments in the header file.
For the details of the main function I will go over some of those details.
For generating your random values I would high suggest staying away from random() or any of its related deprecated or soon to be functions. I would begin by learning and using Pseudo Random Generators that can be found in the standard library along with different types of distributions: these can all be found in <random> header file. You can use std::default_random_engine() but I prefer to use std::random_device We can use that to SEED the engine(generator) of our choosing. One of the more popularly used engines or generators is known as the Mersenne Twister which is std::mt19937 and there is a 65bit version of it too. It is quite simple to do.
{
std::random_device rd; // create an instance of our device to seed with
std::mt19937 gen; // create an instance of our generator (engine)
gen.seed( rd() ); // This seeds the generator (engine)
// Now we need a distribution along with its data type
// there are different versions of these distributions for different types
// Some are for integral types while others are for floating point types
// Here we want a uniform distribution for int so we default the template
std::uniform_int_distribution<> dist(0, 100); //random from [0,100]
// otherwise we could of done
std::uniform_int_distribution<unsigned int> dist2( 0, 50 ); // random from [0, 50]
// There are other types of distributions
std::normal_distribution<> a;
std::poisson_distribution<> b;
// etc.
// If the distributions say "real" they are floating point types
std::uniform_real_distribution<float> f;
std::uniform_real_distribution<double> d;
// Just as there are different distributions there also other
// generators or engines beside the mersenne twister.
// There is another way besides using `random_device` to seed the generator
// you can use <chrono> header to use `std::chrono::high_resolution_clock
// to seed the generator
// You can also seed by const value
// and you can use std::seed_seq;
}
You can find all the information that you need for doing Pseudo Random Generators & Distributions from this web page.
So now that we have our random generators up and working the next step is we declare a std::vector<Point<int>> then we use its reserve function and set it with our const NUM_POINTS. Then we go through a for loop for NUM_POINTS iterations and we populate our container with a random (x,y) set of values.
Then we display the results using a ranged base for loop.
I repeat the above process to show it being done with floats. I did it this way to show the usefulness of templates.
After that I finally sort the containers simply by calling std::sort( begin, end ) using the vector's iterators. Then I go back and use a ranged base for loop to show both of the sorted vectors.
using std::sort works very easy since we defined an overloaded operator<() for our class and we used std::tie to easily compare them. This shows you the power of being of the standard library by bringing together a bunch of parts like set of Legos!
