Suppose we have a std::vector<T> v and T is a class with a member function U T::f() const, where U is convertible to double. I want to create a std::discrete_distribution<> d(/* v[0].f(), ..., v[v.size() - 1].f() */). Is this possible without creating a new container containing the values v[0].f(), ..., v[v.size() - 1].f()?
Create a discrete distribution from transformed container elements
Viewed 63
2 Answers
Something like boost::transform_iterator can do that. You'll just need to create those from the vector's iterators.
If you can add a operator double to T then you can have that return f() and then use the iterator constructor of std::discrete_distribution like
struct Foo
{
operator double() const { return f(); }
double f() const { return 1.0; }
};
int main()
{
std::vector<Foo> foo(10);
std::discrete_distribution dd{foo.begin(), foo.end()};
}
Related