简体   繁体   English

通过索引在运行时访问std :: tuple元素的最佳方法

[英]Optimal way to access std::tuple element in runtime by index

I have function at designed to access std::tuple element by index specified in runtime 我有一个函数at旨在通过运行时指定的索引访问std :: tuple元素

template<std::size_t _Index = 0, typename _Tuple, typename _Function>
inline typename std::enable_if<_Index == std::tuple_size<_Tuple>::value, void>::type
for_each(_Tuple &, _Function)
{}

template<std::size_t _Index = 0, typename _Tuple, typename _Function>
inline typename std::enable_if < _Index < std::tuple_size<_Tuple>::value, void>::type
    for_each(_Tuple &t, _Function f)
{
    f(std::get<_Index>(t));
    for_each<_Index + 1, _Tuple, _Function>(t, f);
}

namespace detail { namespace at {

template < typename _Function >
struct helper
{
    inline helper(size_t index_, _Function f_) : index(index_), f(f_), count(0) {}

    template < typename _Arg >
    void operator()(_Arg &arg_) const
    {
        if(index == count++)
            f(arg_);
    }

    const size_t index;
    mutable size_t count;
    _Function f;
};

}} // end of namespace detail

template < typename _Tuple, typename _Function >
void at(_Tuple &t, size_t index_, _Function f)
{
    if(std::tuple_size<_Tuple> ::value <= index_)
        throw std::out_of_range("");

    for_each(t, detail::at::helper<_Function>(index_, f));
}

It has linear complexity. 它具有线性复杂性。 How can i achive O(1) complexity? 我怎样才能实现O(1)的复杂性?

Assuming you pass something similar to a generic lambda, ie a function object with an overloaded function call operator: 假设你传递类似于泛型lambda的东西,即带有重载函数调用操作符的函数对象:

#include <iostream>

struct Func
{
    template<class T>
    void operator()(T p)
    {
        std::cout << __PRETTY_FUNCTION__ << " : " << p << "\n";
    }
};

The you can build an array of function pointers: 你可以构建一个函数指针数组:

#include <tuple>

template<int... Is> struct seq {};
template<int N, int... Is> struct gen_seq : gen_seq<N-1, N-1, Is...> {};
template<int... Is> struct gen_seq<0, Is...> : seq<Is...> {};

template<int N, class T, class F>
void apply_one(T& p, F func)
{
    func( std::get<N>(p) );
}

template<class T, class F, int... Is>
void apply(T& p, int index, F func, seq<Is...>)
{
    using FT = void(T&, F);
    static constexpr FT* arr[] = { &apply_one<Is, T, F>... };
    arr[index](p, func);
}

template<class T, class F>
void apply(T& p, int index, F func)
{
    apply(p, index, func, gen_seq<std::tuple_size<T>::value>{});
}

Usage example: 用法示例:

int main()
{
    std::tuple<int, double, char, double> t{1, 2.3, 4, 5.6};
    for(int i = 0; i < 4; ++i) apply(t, i, Func{});
}

clang++ also accepts an expansion applied to a pattern that contains a lambda expression: clang ++还接受应用于包含lambda表达式的模式的扩展:

static FT* arr[] = { [](T& p, F func){ func(std::get<Is>(p)); }... };

(although I've to admit that looks really weird) (虽然我承认看起来很奇怪)

g++4.8.1 rejects this. g ++ 4.8.1拒绝这一点。

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM