3

I'm trying to change a a tuple of vectors into a vector of tuples (and vice-versa). I'm having trouble calling the tuple_transpose function. When I call it with one argument I get a no matching function call error:

prog.cpp: In function ‘int main()’:
prog.cpp:44:24: error: no matching function for call to ‘tuple_transpose(std::tuple >, std::vector > >&)’
prog.cpp:44:24: note: candidates are:
prog.cpp:30:6: note: template typename transpose::type tuple_transpose(std::tuple >...>&, seq)
prog.cpp:30:6: note: template argument deduction/substitution failed:
prog.cpp:44:24: note: candidate expects 2 arguments, 1 provided
prog.cpp:36:6: note: template typename transpose::type tuple_transpose(std::tuple >...>&)
prog.cpp:36:6: note: template argument deduction/substitution failed:
prog.cpp: In substitution of ‘template typename transpose::type tuple_transpose(std::tuple >...>&) [with T = {int, bool}]’:
prog.cpp:44:24: required from here
prog.cpp:36:6: error: no type named ‘type’ in ‘struct transpose >, std::vector > >&>’

#include <vector>
#include <tuple>
#include <type_traits>

template <typename... T>
struct transpose {};

template <typename... T>
struct transpose<std::tuple<std::vector<T>...>>
{
    using type = std::vector<std::tuple<T...>>;
};

template <typename... T>
struct transpose<std::vector<std::tuple<T...>>>
{
    using type = std::tuple<std::vector<T>...>;
};

// Indicies from Andy Prowl's answer
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 <typename... T, int... Is>
auto tuple_transpose(std::tuple<std::vector<T>...>& var, seq<Is...>) -> typename transpose<decltype(var)>::type
{
    return { std::make_tuple(std::get<Is>(var)...) };
}

template <typename... T>
auto tuple_transpose(std::tuple<std::vector<T>...>& var) -> typename transpose<decltype(var)>::type
{
    return tuple_transpose(var, gen_seq<sizeof...(T)>{});
}

int main()
{
    std::tuple<std::vector<int>, std::vector<bool>> var;
    tuple_transpose(var); // error
    ...
}

Here is a demo with the error in it: http://ideone.com/7AWiQQ#view_edit_box

What am I doing wrong and how can I fix it? Thanks.

2
  • 2
    What do you mean "change a a tuple of vectors into a vector of tuples". Transpose? What if vectors have different lengths? The current error is because decltype(var) is a reference type, but there are other issues with your code, most notably complete absence of conversion logic. I see only unpacking and repacking a tuple. Commented May 15, 2013 at 23:20
  • Your idea with expanding the vector in a braced-init-list cannot work as the size of the vector is only known at run-time. (See Andy Prowl's answer.) Commented May 15, 2013 at 23:26

1 Answer 1

3

If you are assuming vectors of identical sizes, this should do the job:

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 <typename... T, int... Is>
auto transpose(std::tuple<std::vector<T>...>& var, seq<Is...>)
    -> std::vector<std::tuple<T...>>
{
    std::vector<std::tuple<T...>> result;
    for (std::size_t i = 0; i < std::get<0>(var).size(); i++)
    {
        std::tuple<T...> t = std::make_tuple(std::get<Is>(var)[i]...);
        result.push_back(t);
    }

    return result;
}

template <typename... T, int... Is>
auto transpose(std::tuple<std::vector<T>...>& var)
    -> std::vector<std::tuple<T...>>
{
    return transpose(var, gen_seq<sizeof...(T)>());
}

And here is how you could test it:

#include <iostream>
#include <iomanip>

int main()
{
    std::vector<int> vi = {42, 1729, 6};
    std::vector<bool> vb = {true, false, false};
    std::vector<std::string> vs = {"Hi", "Hey", "Ho"};

    auto t = make_tuple(vi, vb, vs);
    auto v = transpose(t);

    std::cout << std::boolalpha;
    for (auto const& t : v)
    {
        std::cout << "(";
        std::cout << std::get<0>(t);
        std::cout << ", " << std::get<1>(t);
        std::cout << ", " << std::get<2>(t);
        std::cout << ")" << std::endl;
    }
}

Finally, a live example.

Sign up to request clarification or add additional context in comments.

1 Comment

I'd be tempted to use vector::at to remain defined behaviour if the vectors are different sizes.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.