Conversion of variant, vector<variant> and vector<vector<variant>> to the equivalent type of my choice

Question

My question concerns the mix of boost::variant conversions with std::vector in C++. In my project, I use variants as output of SQL. I will always just use 1 type of variant. What I am trying to do is to have an easy conversion from variant, vector of variant or 2xvector of variant to the equivalent type of my choice. Naively, I need something along the lines of:

std::vector < int > my_variable = convert(some_variant_vector)

First here is my setup:

#include "boost\variant.hpp"
#include "boost\date_time\gregorian\gregorian.hpp"
typedef boost::variant< int, std::string, boost::gregorian::date> _var;
typedef std::vector<_var> _vec_var; 
typedef std::vector<_vec_var> _vec2_var;

I would like to have a simple way to convert a _var (or _vec_var or _vec2_var) into a int/string/date depending on my needs. From the following post I know that my answer should look something like this:

template<typename T>
struct converter_visitor : public boost::static_visitor<T>
{
    const _var &converter;
    converter_visitor(const _var &r) : converter(r) {}
    T operator()(const _var &) const{
        return boost::get<_var>(converter);
    }

    const _vec_var &v_converter;            // case of vector<>
    converter_visitor(const _vec_var &r) : v_converter(r) {}
    T operator()(const _vec_var &) const{
        T ans;
        ans.reserve(_cont.size());
        for (int i = 0; i < _cont.size(); ++i)
            ans.push_back(boost::get<T>(v_converter[i]));
        return ans;
    }

    const _vec2_var & v2_converter;         // case of vector<vector>
    converter_visitor(const _vec2_var &r) : v2_converter(r) {}
    T operator()(const _vec2_var &) const {
        T ans;
        ans.reserve(v2_converter.size(), v2_converter[0].size());
        for (int i = 0; i < _cont.size(); ++i)
        {
            for (size_t j = 0; j < v2_converter[0].size(); j++)
            {
                ans.push_back(boost::get<T>(v2_converter[i][j]));
            }
        }
        return ans;
    }
};

int main()
{
    _var variable = 1;
    int integer_conversion;
    boost::apply_visitor(converter_visitor(test), integer_conversion);
    return 0;
}

Unfortunately I have been stuck in a while with this as it is not working as intended and does not even compile.

Answer 1

I found a solution to your problem, as I said in the comment, I did the following:

I wrote a bool_visitor<T> wich returns true if a boost::variant holds T and false otherwise;

here is the visitor:

template<typename T>
struct bool_visitor : public boost::static_visitor<bool> {

    inline constexpr bool operator()(T val) const
    {
        return false;
    }

    template<typename K>

    typename std::enable_if<!std::is_convertible<K, T>::value, bool>::type
    operator()(K val) const
    {
        return true;
    }

};

with this visitor we are able to remove all variants of a vector which doesnt hold a specific type. After this we use std::transform to transform all the variants to the type they hold. We use SFINAE to check if the given vector is a nested vector(vector) or not, if so, we flatten the vector first, with a flatten function:

template <typename FromIter, typename ToIter>
void flatten (FromIter start, FromIter end, ToIter dest) {
    while (start != end) {
        dest = std::copy(start->begin(), start->end(), dest);
        ++start;
    }
}

to check if the vector is nested we need a is_vector trait:

template<typename>
struct is_vector : std::false_type {};

template<typename T, typename A>
struct is_vector<std::vector<T,A>> : std::true_type {};

now we can use SFINAE to enable the function we want on a given vector:

template<typename T, typename K>
typename
std::enable_if_t<
        !is_vector<K>::value,
        std::vector<T>>
get_vec_of(std::vector<K>& input_vec){
    //delete all variants not holding T
    auto it = std::remove_if(input_vec.begin(), input_vec.end(),
                             [](auto item){
                                 return boost::apply_visitor(bool_visitor<T>(), item);
                             });
    //input_vec.erase(it, input_vec.end());

    //create new vector of T
    std::vector<T> return_vec;

    //transform all variants holding T to T and put them in the return_vec
    std::transform(input_vec.begin(), it, std::back_inserter(return_vec),
                   [](auto item){
                       //this can never throw because all variants in the vector are holding T
                       return boost::get<T>(item);
                   });

    return return_vec;
}

template<typename T, typename K>
typename
std::enable_if_t<
        is_vector<K>::value,
        std::vector<T>>
get_vec_of(std::vector<K>& input_vec){
    std::vector<typename K::value_type> flatten_vec;
    flatten(input_vec.begin(), input_vec.end(), std::back_inserter(flatten_vec));
    return get_vec_of<T>(flatten_vec);
};

This solution changes the order of the elements inside the given vector. If this is bad for you, you maybe should copy the vector first.

here is a demo on how to use this function, and how it works.

Answer 2

I would suggest the following to get the types you want:

template<typename T>
class converter_visitor : public boost::static_visitor<>
{
public: 
    std::vector<T>& vec;

    converter_visitor(std::vector<T>& r) : vec(r) {}

    // only push back values of specific types...
    void operator()(const T& u) const {
        vec.push_back(u);
    }

    // ignore other types...
    void operator()(...) const {}  
};

template<typename T>
converter_visitor<T> make_visitor(std::vector<T>& r) { return converter_visitor<T>(r); }

and then funnel that to a recursive filter func that can handle nested vectors:

template<typename T,typename U>
void filter(std::vector<T>& result,const U& var) {
    boost::apply_visitor( make_visitor(result), var );
}

template<typename T,typename U>
void filter(std::vector<T>& result,const std::vector<U>& cont) {
    std::for_each(cont.begin(),cont.end(),[&](const U& c) {
        filter(result,c);
    });
}

then you can do:

_var v = 314;
std::vector<int> result; 
filter(result,v);
print(result);

result: 314

_vec_var v;
v.push_back(2);
v.push_back(3);
v.push_back("hello");
v.push_back(5);
v.push_back(7);
v.push_back("world");

std::vector<int> result;         
filter(result,v);
print(result);

std::vector<std::string> result2;         
filter(result2,v);
print(result2);

result1: 2 3 5 7
result2: hello world

_vec_var v1;
v1.push_back(11);
v1.push_back(13);
v1.push_back("see ya");

_vec_var v2;
v2.push_back(17);
v2.push_back(19);
v2.push_back("later");

_vec2_var vv;
vv.push_back(v1);
vv.push_back(v2);

std::vector<int> result;         
filter(result,vv);
print(result);

std::vector<std::string> result2;         
filter(result2,vv);
print(result2);

result1: 11 13 17 19
result2: see ya later

see live demo here