iterator adapter to iterate just the values in a map? - c++

I'm just getting back into C++ after a couple of years of doing a lot of C#, and recently Objective C.
One thing I've done before is to roll my own iterator adapter for std::map that will deref to just the value part, rather than the key-value pair. This is quite a common and natural thing to do. C# provides this facility with its Keys and Values properties of its Dictionary class. Objective-C's NSDictionary, similarly, has allKeys and allValues.
Since I've been "away", Boost has acquired the Range and ForEach libraries, which I am now using extensively. I wondered if between the two there was some facility to do the same, but I haven't been able to find anything.
I'm thinking of knocking something up using Boost's iterator adapters, but before I go down that route I thought I'd ask here if anyone knows of such a facility in Boost, or somewhere else ready made?

Replacing the previous answer, in case anybody else finds this like I did. As of boost 1.43, there are some commonly used range adaptors provided. In this case, you want boost::adaptors::map_values. The relevant example:
http://www.boost.org/doc/libs/1_46_0/libs/range/doc/html/range/reference/adaptors/reference/map_values.html#range.reference.adaptors.reference.map_values.map_values_example

I don't think there's anything out of the box. You can use boost::make_transform.
template<typename T1, typename T2> T2& take_second(const std::pair<T1, T2> &a_pair)
{
return a_pair.second;
}
void run_map_value()
{
map<int,string> a_map;
a_map[0] = "zero";
a_map[1] = "one";
a_map[2] = "two";
copy( boost::make_transform_iterator(a_map.begin(), take_second<int, string>),
boost::make_transform_iterator(a_map.end(), take_second<int, string>),
ostream_iterator<string>(cout, "\n")
);
}

There is a boost range adaptor for exactly this purpose.
See http://www.boost.org/doc/libs/1_53_0/libs/range/doc/html/range/reference/adaptors/reference/map_values.html
(This example cribbed from there)
int main(int argc, const char* argv[])
{
using namespace boost::assign;
using namespace boost::adaptors;
std::map<int,int> input;
for (int i = 0; i < 10; ++i)
input.insert(std::make_pair(i, i * 10));
boost::copy(
input | map_values,
std::ostream_iterator<int>(std::cout, ","));
return 0;
}

Continuing David's answer, there's another possibility to put the boile by creating a derived class from boost::transform_iterator. I'm using this solution in my projects:
namespace detail
{
template<bool IsConst, bool IsVolatile, typename T>
struct add_cv_if_c
{
typedef T type;
};
template<typename T>
struct add_cv_if_c<true, false, T>
{
typedef const T type;
};
template<typename T>
struct add_cv_if_c<false, true, T>
{
typedef volatile T type;
};
template<typename T>
struct add_cv_if_c<true, true, T>
{
typedef const volatile T type;
};
template<typename TestConst, typename TestVolatile, typename T>
struct add_cv_if: public add_cv_if_c<TestConst::value, TestVolatile::value, T>
{};
} // namespace detail
/** An unary function that accesses the member of class T specified in the MemberPtr template parameter.
The cv-qualification of T is preserved for MemberType
*/
template<typename T, typename MemberType, MemberType T::*MemberPtr>
struct access_member_f
{
// preserve cv-qualification of T for T::second_type
typedef typename detail::add_cv_if<
std::tr1::is_const<T>,
std::tr1::is_volatile<T>,
MemberType
>::type& result_type;
result_type operator ()(T& t) const
{
return t.*MemberPtr;
}
};
/** #short An iterator adaptor accessing the member called 'second' of the class the
iterator is pointing to.
*/
template<typename Iterator>
class accessing_second_iterator: public
boost::transform_iterator<
access_member_f<
// note: we use the Iterator's reference because this type
// is the cv-qualified iterated type (as opposed to value_type).
// We want to preserve the cv-qualification because the iterator
// might be a const_iterator e.g. iterating a const
// std::pair<> but std::pair<>::second_type isn't automatically
// const just because the pair is const - access_member_f is
// preserving the cv-qualification, otherwise compiler errors will
// be the result
typename std::tr1::remove_reference<
typename std::iterator_traits<Iterator>::reference
>::type,
typename std::iterator_traits<Iterator>::value_type::second_type,
&std::iterator_traits<Iterator>::value_type::second
>,
Iterator
>
{
typedef boost::transform_iterator<
access_member_f<
typename std::tr1::remove_reference<
typename std::iterator_traits<Iterator>::reference
>::type,
typename std::iterator_traits<Iterator>::value_type::second_type,
&std::iterator_traits<Iterator>::value_type::second
>,
Iterator
> baseclass;
public:
accessing_second_iterator():
baseclass()
{}
// note: allow implicit conversion from Iterator
accessing_second_iterator(Iterator it):
baseclass(it)
{}
};
This leads to even cleaner code:
void run_map_value()
{
typedef map<int, string> a_map_t;
a_map_t a_map;
a_map[0] = "zero";
a_map[1] = "one";
a_map[2] = "two";
typedef accessing_second_iterator<a_map_t::const_iterator> ia_t;
// note: specify the iterator adaptor type explicitly as template type, enabling
// implicit conversion from begin()/end()
copy<ia_t>(a_map.begin(), a_map.end(),
ostream_iterator<string>(cout, "\n")
);
}

Related

Is there a more idiomatic way to specialise behaviour using flags passed via template?

Apologises for the ambiguous title.
Here is my code:
struct LowHigh
{
};
struct HighLow
{
};
template < class LookupScheme>
struct ladder_base
{
using value_type = price_depth;
using ladder_type = std::vector< value_type >;
template < class T >
struct lookup;
template <>
struct lookup< LowHigh >
{
static constexpr auto func = std::upper_bound< ladder_type::iterator, value_type >;
};
template <>
struct lookup< HighLow >
{
static constexpr auto func = std::lower_bound< ladder_type::iterator, value_type >;
};
void
insert(value_type v)
{
auto iter = lookup< LookupScheme >::func(std::begin(data_), std::end(data_), v);
data_.insert(iter, std::move(v));
}
protected:
std::vector< value_type > data_;
};
} // namespace detail
struct bid_ladder : detail::ladder_base< detail::HighLow >
{
};
struct offer_ladder : detail::ladder_base< detail::LowHigh >
{
};
I'm specialising lookup::func depending on the scheme passed as a template type. There are currently only two possible schemes: LowHigh & HighLow. This has the effect of determining how the underlying vector is sorted.
Is there a more idiomatic/cleaner way to express this logic?
These algorithms take a comparison object as their last parameter - so you can use that to your advantage.
template < class Compare >
struct ladder_base
{
using value_type = price_depth;
using ladder_type = std::vector< value_type >;
void
insert(value_type v)
{
auto iter = std::upper_bound(data_.begin(), data_.end(), v, Compare{} );
data_.insert(iter, std::move(v));
}
protected:
std::vector< value_type > data_;
};
And then use ladder_base<std::less<>> or ladder_base<std::greater<>>, depending on which sort order you want.
Note that std::lower_bound and std::upper_bound are not antonyms, so your original wasn't really correct. lower_bound gives you the first element >= x and upper_bound gives you the first element > x. So changing from one to the other doesn't change your sort order (both require increasing order), only the comparison object affects that.
For instance:
std::vector<int> v = {1, 3, 5, 7};
auto i = std::lower_bound(v.begin(), v.end(), 3); // this is the 3
auto j = std::upper_bound(v.begin(), v.end(), 3); // this is the 5
Note that the vector is sorted in increasing order, but both calls are perfectly well-formed. If you wanted a reverse sort, you'd have to pass std::greater{} in as the comparison object (as I'm showing).
But either way, you want to use std::upper_bound - regardless of sort order.
The idiomatic way of doing that type of stuff is to have the template parameter contain the code to invoke directly, instead of indirecting through a tag type like you are doing.
Mind you, at that point, you could always just pass std::upper_bound directly as a template parameter.
Furthermore, since this is tagged c++20, you would also ideally use a concept, to constrain the types that can be passed to ladder_base.
#include <concepts>
#include <vector>
using price_depth = int;
template<typename T>
concept LookupScheme = requires (const T& x, const std::vector<price_depth>& v) {
{x(v.begin(), v.end(), price_depth{})} -> std::same_as<decltype(v.begin())>;
};
namespace detail {
struct LowHigh {
template<typename ForwardIt, typename T>
decltype(auto) operator()(ForwardIt first, ForwardIt last, const T& value ) const {
return std::upper_bound(first, last, value);
}
};
struct HighLow {
template<typename ForwardIt, typename T>
decltype(auto) operator()(ForwardIt first, ForwardIt last, const T& value ) const {
return std::lower_bound(first, last, value);
}
};
template <LookupScheme Scheme>
struct ladder_base
{
using value_type = price_depth;
using ladder_type = std::vector< value_type >;
void insert(value_type v)
{
auto iter = Scheme::exec(std::begin(data_), std::end(data_), v);
data_.insert(iter, std::move(v));
}
protected:
std::vector< value_type > data_;
};
} // namespace detail
struct bid_ladder : detail::ladder_base< detail::LowHigh >
{
};
struct offer_ladder : detail::ladder_base< detail::HighLow >
{
};
You can see the same approach used in the standard library's sorted containers, such a std::map<>'s Compare parameter.

Cannot deduce return type of lambda with std::map

This is a strange case.
The following code produces error when using User defined types but not when using primitives!
It produces errors with std::map (also QMap) but not if replaced with std::pair (also QPair)!
Why?
I'm using MSVC 2013.
class Class {};
typedef double Type; // This works fine
//typedef Class Type; // This produces error
template<typename T, typename ValueFunctor>
std::map<int, typename std::result_of<ValueFunctor(T)>::type >
testFun(ValueFunctor valueFunctor)
{
std::map<int, typename std::result_of<ValueFunctor(T)>::type > map;
return map;
}
void test()
{
std::map<int,Type> output = testFun<Type>(
// The error can be removed by explicitly specifying return type
[](const Type &pair)/*->T2*/{
return pair;
}
);
}
What's happening here is that in MSVC 2013 and earlier, the std::result_of template is returning a const Class instead of just Class. That causes the value type of map to be non-copyable.
You can see exactly the same problem by changing typedef double Type to typedef const double Type.
If you still have to use a non C++11 compliant library, you can specify the return type of the lambda to be Type but that would involve a copy.
Alternatively, remove the constness but changing:
typename std::result_of<ValueFunctor(T)>::type
to
typename std::remove_const<typename std::result_of<ValueFunctor(T)>::type >::type
template<typename T, typename ValueFunctor>
std::map<int, typename std::remove_const<typename std::result_of<ValueFunctor(T)>::type >::type>
testFun(ValueFunctor valueFunctor)
{
std::map<int, typename std::remove_const<typename std::result_of<ValueFunctor(T)>::type>::type > map;
return map;
}

How can I use C++11 variadic templates to define a vector-of-tuples backed by a tuple-of-vectors?

Suppose I have a bunch of vectors:
vector<int> v1;
vector<double> v2;
vector<int> v3;
all of the same length. Now, for every index i, I would like to be able to treat (v1[i], v2[i], v3[i]) as a tuple, and maybe pass it around. In fact, I want to have a a vector-of-tuples rather than a tuple-of-vectors, using which I can do the above. (In C terms, I might say an array-of-structs rather than a struct-of-arrays). I do not want to effect any data reordering (think: really long vectors), i.e. the new vector is backed by the individual vectors I pass in. Let's .
Now, I want the class I write (call it ToVBackedVoT for lack of a better name) to support any arbitrary choice of vectors to back it (not just 3, not int, double and int, not every just scalars). I want the vector-of-tuples to be mutable, and for no copies to be made on construction/assignments.
If I understand correctly, variadic templates and the new std::tuple type in C++11 are the means for doing this (assuming I don't want untyped void* arrays and such). However, I only barely know them and have never worked with them. Can you help me sketch out how such a class will look like? Or how, given
template <typename ... Ts>
I can express something like "the list of template arguments being the replacement of each typename in the original template arguments with a vector of elements of this type"?
Note: I think I might also want to later be able to adjoin additional vectors to the backing vectors, making an instance of ToVBackedVoT<int, double, int> into, say, an instance of ToVBackedVoT<int, double, int, unsigned int>. So, bear that in mind when answering. This is not critically important though.
One idea is to keep the storage in the "struct of array" style in form of vectors for good performance if only a subset of the fields are used for a particular task. Then, for each kind of task requiring a different set of fields, you can write a lightweight wrapper around some of those vectors, giving you a nice random access iterator interface similar to what std::vector supports.
Concerning the syntax of variadic templates, this is how a wrapper class (without any iterators yet) could look like:
template<class ...Ts> // Element types
class WrapMultiVector
{
// references to vectors in a TUPLE
std::tuple<std::vector<Ts>&...> m_vectors;
public:
// references to vectors in multiple arguments
WrapMultiVector(std::vector<Ts> & ...vectors)
: m_vectors(vectors...) // construct tuple from multiple args.
{}
};
To construct such a templated class, it's often preferred to have a template type deducting helper function available (similar to those make_{pair|tuple|...} functions in std):
template<class ...Ts> // Element types
WrapMultiVector<Ts...> makeWrapper(std::vector<Ts> & ...vectors) {
return WrapMultiVector<Ts...>(vectors...);
}
You already see different types of "unpacking" the type list.
Adding iterators suitable to your application (you requested in particular random access iterators) is not so easy. A start could be forward only iterators, which you might extend to random access iterators.
The following iterator class is capable of being constructed using a tuple of element iterators, being incremented and being dereferenced to obtain a tuple of element references (important for read-write access).
class iterator {
std::tuple<typename std::vector<Ts>::iterator...> m_elemIterators;
public:
iterator(std::tuple<typename std::vector<Ts>::iterator...> elemIterators)
: m_elemIterators(elemIterators)
{}
bool operator==(const iterator &o) const {
return std::get<0>(m_elemIterators) == std::get<0>(o.m_elemIterators);
}
bool operator!=(const iterator &o) const {
return std::get<0>(m_elemIterators) != std::get<0>(o.m_elemIterators);
}
iterator& operator ++() {
tupleIncrement(m_elemIterators);
return *this;
}
iterator operator ++(int) {
iterator old = *this;
tupleIncrement(m_elemIterators);
return old;
}
std::tuple<Ts&...> operator*() {
return getElements(IndexList());
}
private:
template<size_t ...Is>
std::tuple<Ts&...> getElements(index_list<Is...>) {
return std::tie(*std::get<Is>(m_elemIterators)...);
}
};
For demonstration purposes, two different patterns are in this code which "iterate" over a tuple in order to apply some operation or construct a new tuple with some epxression to be called per element. I used both in order to demonstrate alternatives; you can also use the second method only.
tupleIncrement: You can use a helper function which uses meta programming to index a single entry and advance the index by one, then calling a recursive function, until the index is at the end of the tuple (then there is a special case implementation which is triggered using SFINAE). The function is defined outside of the class and not above; here is its code:
template<std::size_t I = 0, typename ...Ts>
inline typename std::enable_if<I == sizeof...(Ts), void>::type
tupleIncrement(std::tuple<Ts...> &tup)
{ }
template<std::size_t I = 0, typename ...Ts>
inline typename std::enable_if<I < sizeof...(Ts), void>::type
tupleIncrement(std::tuple<Ts...> &tup)
{
++std::get<I>(tup);
tupleIncrement<I + 1, Ts...>(tup);
}
This method can't be used to assign a tuple of references in the case of operator* because such a tuple has to be initialized with references immediately, which is not possible with this method. So we need something else for operator*:
getElements: This version uses an index list (https://stackoverflow.com/a/15036110/592323) which gets expanded too and then you can use std::get with the index list to expand full expressions. The IndexList when calling the function instantiates an appropriate index list which is only required for template type deduction in order to get those Is.... The type can be defined in the wrapper class:
// list of indices
typedef decltype(index_range<0, sizeof...(Ts)>()) IndexList;
More complete code with a little example can be found here: http://ideone.com/O3CPTq
Open problems are:
If the vectors have different sizes, the code fails. Better would be to check all "end" iterators for equality; if one iterator is "at end", we're also "at end"; but this would require some logic more than operator== and operator!= unless it's ok to "fake" it in; meaning that operator!= could return false as soon as any operator is unequal.
The solution is not const-correct, e.g. there is no const_iterator.
Appending, inserting etc. is not possible. The wrapper class could add some insert or and / or push_back function in order to make it work similar to std::vector. If your goal is that it's syntactically compatible to a vector of tuples, reimplement all those relevant functions from std::vector.
Not enough tests ;)
An alternative to all the variadic template juggling is to use the boost::zip_iterator for this purpose. For example (untested):
std::vector<int> ia;
std::vector<double> d;
std::vector<int> ib;
std::for_each(
boost::make_zip_iterator(
boost::make_tuple(ia.begin(), d.begin(), ib.begin())
),
boost::make_zip_iterator(
boost::make_tuple(ia.end(), d.end(), ib.end())
),
handle_each()
);
Where your handler, looks like:
struct handle_each :
public std::unary_function<const boost::tuple<const int&, const double&, const int&>&, void>
{
void operator()(const boost::tuple<const int&, const double&, const int&>& t) const
{
// Now you have a tuple of the three values across the vector...
}
};
As you can see, it's pretty trivial to expand this to support an arbitrary set of vectors..
From asker's clarification on how this would be used (code that takes a tuple), I'm going to propose this instead.
//give the i'th element of each vector
template<typename... Ts>
inline tuple<Ts&...> ith(size_t i, vector<Ts>&... vs){
return std::tie(vs[i]...);
}
There's a proposal to allow parameter packs to be saved as members of classes (N3728). Using that, here's some untested and untestable code.
template<typename... Types>
class View{
private:
vector<Types>&... inner;
public:
typedef tuple<Types&...> reference;
View(vector<Types>&... t): inner(t...) {}
//return smallest size
size_t size() const{
//not sure if ... works with initializer lists
return min({inner.size()...});
}
reference operator[](size_t i){
return std::tie(inner[i]...);
}
};
And iteration:
public:
iterator begin(){
return iterator(inner.begin()...);
}
iterator end(){
return iterator(inner.end()...);
}
//for .begin() and .end(), so that ranged-based for can be used
class iterator{
vector<Types>::iterator... ps;
iterator(vector<Types>::iterator... its):ps(its){}
friend View;
public:
//pre:
iterator operator++(){
//not sure if this is allowed.
++ps...;
//use this if not:
// template<typename...Types> void dummy(Types... args){} //global
// dummy(++ps...);
return *this;
}
iterator& operator--();
//post:
iterator operator++(int);
iterator operator--(int);
//dereference:
reference operator*()const{
return std::tie(*ps...);
}
//random access:
iterator operator+(size_t i) const;
iterator operator-(size_t i) const;
//need to be able to check end
bool operator==(iterator other) const{
return std::make_tuple(ps...) == std::make_tuple(other.ps...);
}
bool operator!=(iterator other) const{
return std::make_tuple(ps...) != std::make_tuple(other.ps...);
}
};
You may use something like:
#if 1 // Not available in C++11, so write our own
// class used to be able to use std::get<Is>(tuple)...
template<int... Is>
struct index_sequence { };
// generator of index_sequence<Is>
template<int N, int... Is>
struct make_index_sequence : make_index_sequence<N - 1, N - 1, Is...> { };
template<int... Is>
struct make_index_sequence<0, Is...> : index_sequence<Is...> { };
#endif
// The 'converting' class
// Note that it doesn't check that vector size are equal...
template<typename ...Ts>
class ToVBackedVoT
{
public:
explicit ToVBackedVoT(std::vector<Ts>&... vectors) : data(vectors...) {}
std::tuple<const Ts&...> operator [] (unsigned int index) const
{
return at(index, make_index_sequence<sizeof...(Ts)>());
}
std::tuple<Ts&...> operator [] (unsigned int index)
{
return at(index, make_index_sequence<sizeof...(Ts)>());
}
private:
template <int... Is>
std::tuple<const Ts&...> at(unsigned int index, index_sequence<Is...>) const
{
return std::tie(std::get<Is>(data)[index]...);
}
template <int... Is>
std::tuple<Ts&...> at(unsigned int index, index_sequence<Is...>)
{
return std::tie(std::get<Is>(data)[index]...);
}
private:
std::tuple<std::vector<Ts>&...> data;
};
And to iterate, create an 'IndexIterator' like the one in https://stackoverflow.com/a/20272955/2684539
To adjoin additional vectors, you have to create an other ToVBackedVoT as std::tuple_cat does for std::tuple
Conversion to a std::tuple of vectors (vector::iterators):
#include <iostream>
#include <vector>
// identity
// ========
struct identity
{
template <typename T>
struct apply {
typedef T type;
};
};
// concat_operation
// ================
template <typename Operator, typename ...> struct concat_operation;
template <
typename Operator,
typename ...Types,
typename T>
struct concat_operation<Operator, std::tuple<Types...>, T>
{
private:
typedef typename Operator::template apply<T>::type concat_type;
public:
typedef std::tuple<Types..., concat_type> type;
};
template <
typename Operator,
typename ...Types,
typename T,
typename ...U>
struct concat_operation<Operator, std::tuple<Types...>, T, U...>
{
private:
typedef typename Operator::template apply<T>::type concat_type;
public:
typedef typename concat_operation<
Operator,
std::tuple<Types..., concat_type>,
U...>
::type type;
};
template <
typename Operator,
typename T,
typename ...U>
struct concat_operation<Operator, T, U...>
{
private:
typedef typename Operator::template apply<T>::type concat_type;
public:
typedef typename concat_operation<
Operator,
std::tuple<concat_type>,
U...>
::type type;
};
// ToVectors (ToVBackedVoT)
// =========
template <typename ...T>
struct ToVectors
{
private:
struct to_vector {
template <typename V>
struct apply {
typedef typename std::vector<V> type;
};
};
public:
typedef typename concat_operation<to_vector, T...>::type type;
};
// ToIterators
// ===========
template <typename ...T>
struct ToIterators;
template <typename ...T>
struct ToIterators<std::tuple<T...>>
{
private:
struct to_iterator {
template <typename V>
struct apply {
typedef typename V::iterator type;
};
};
public:
typedef typename concat_operation<to_iterator, T...>::type type;
};
int main() {
typedef ToVectors<int, double, float>::type Vectors;
typedef ToVectors<Vectors, int, char, bool>::type MoreVectors;
typedef ToIterators<Vectors>::type Iterators;
// LOG_TYPE(Vectors);
// std::tuple<
// std::vector<int, std::allocator<int> >,
// std::vector<double, std::allocator<double> >,
// std::vector<float, std::allocator<float> > >
// LOG_TYPE(Iterators);
// std::tuple<
// __gnu_cxx::__normal_iterator<int*, std::vector<int, std::allocator<int> > >,
// __gnu_cxx::__normal_iterator<double*, std::vector<double, std::allocator<double> > >,
// __gnu_cxx::__normal_iterator<float*, std::vector<float, std::allocator<float> > > >
}
As an alternative similar to boost::zip_iterator I wrote a zip function with a very simple interface:
vector<int> v1;
vector<double> v2;
vector<int> v3;
auto vec_of_tuples = zip(v1, v2, v3);
For example, iterate over these tuples:
for (auto tuple : zip(v1, v2, v3)) {
int x1; double x2; int x3;
std::tie(x1, x2, x3) = tuple;
//...
}
Here, zip() takes any number of ranges of any type. It returns an adaptor which can be seen as a lazily evaluated range over a tuple of elements originating from the wrapped ranges.
The adaptor is part of my Haskell-style functional library "fn" and implemented using variadic templates.
Currently it doesn't support modification of the original ranges' values via the adaptor because of the design of the library (it's intended to be used with non-mutable ranges like in functional programming).
A brief explanation on how this is done is: zip(...) returns an adaptor object which implements begin() and end(), returning an iterator object. The iterator holds a tuple of iterators to the wrapped ranges. Incrementing the iterator increments all wrapped iterators (which is implemented using an index list and unpacking an incrementing expression into a series of expressions: ++std::get<I>(iterators)...). Dereferencing the iterator will decrement all wrapped iterators and pass it to std::make_tuple (which is also implemented as unpacking the expression *std::get<I>(iterators)...).
P.S. Its implementation is based on a lot of ideas coming from answers to this question.

Adding element to back of STL container

I'm looking for a general way of adding an element to the back of an STL container. I would like the code to support as many types of STL container as possible. The following piece of code demonstrates my problem:
#include <vector>
#include <string>
using namespace std;
template<typename T>
class S {
T built;
typename T::iterator built_it;
public:
S() : built{}, built_it{built.end()} {}
void add_to(typename T::value_type e) {
built.emplace(built_it, e);
++built_it;
}
const T& get() {
return built;
}
};
int main()
{
S<std::vector<int>> e;
S<std::string> f;
e.add_to(3); // works
f.add_to('c'); // doesn't
}
The problem here is subtle. This code works great for vectors, because std::vector implements the emplace function. But std::string does not! Is there a more general way to perform the same operation?
The most generic way (not necessarily the most efficient way) is:
c.insert( c.end(), value );
where, of course, value needs to be suitable for the container c (you may use decltype(c)::value_type). In case of an associative container, e.g. map, it is a std::pair.
This works for all standard containers except for std::forward_list. For some containers the element is then added at the end, for some the c.end() is just a hint that might be ignored.
As a follow up to the comments, here's the advanced stuff ;)
When you want to insert a known number of elements into a given container c (of type C) and you want to be at least somewhat efficient, you should detect wether the container type supports reserve() and call it before inserting elements.
The following method detects reserve() correctly (the link explains how):
template< typename C, typename = void >
struct has_reserve
: std::false_type
{};
template< typename C >
struct has_reserve< C, std::enable_if_t<
std::is_same<
decltype( std::declval<C>().reserve( std::declval<typename C::size_type>() ) ),
void
>::value
> >
: std::true_type
{};
Now you can use it with std::enable_if_t to optionally reserve space. An example could look like this:
template< typename C >
std::enable_if_t< !has_reserve< C >::value >
optional_reserve( C&, std::size_t ) {}
template< typename C >
std::enable_if_t< has_reserve< C >::value >
optional_reserve( C& c, std::size_t n )
{
c.reserve( c.size() + n );
}
template< typename C, typename T, std::size_t N >
void add_array( C& c, const std::array< T, N >& a )
{
optional_reserve( c, N );
for( const auto& e : a ) {
c.insert( c.end(), typename C::value_type( e ) ); // see remark below
}
}
add_array can now be called with all standard containers (except std::forward_list) and it will call reserve() for std::vector and the unordered associative containers.
As the above does not need explicit specialization or overloading for specific container types, it also works for non-standard containers as long as their interfaces are designed reasonably similar to the standard containers' interfaces. (In fact I had several such "home-made" containers in the past and the above Just-Works™)
A remark about the conversion in the above code: The reason for converting the Ts to C::value_type is just to show that this would be the correct place if it is needed. In the above example it might look superfluous, but in my real-world code I call a special conversion traits class to convert the es (which are encoded strings) into the correct value type for any container.
Most often, people use traits.
Many boost libraries have solved this same problem, so you might be able to reuse existing traits.
A simple demonstration: Live on Coliru
#include <vector>
#include <set>
#include <string>
namespace traits
{
template <typename Container, typename Enable = void>
struct add_at_end;
template <typename... TAs>
struct add_at_end<std::vector<TAs...> >
{
using Container = std::vector<TAs...>;
template <typename... CtorArgs>
static void apply(Container& container, CtorArgs&&... args) {
container.emplace_back(std::forward<CtorArgs>(args)...);
}
};
template <typename... TAs>
struct add_at_end<std::set<TAs...> >
{
using Container = std::set<TAs...>;
template <typename... CtorArgs>
static void apply(Container& container, CtorArgs&&... args) {
container.insert(container.end(), { std::forward<CtorArgs>(args)...});
}
};
template <typename... TAs>
struct add_at_end<std::basic_string<TAs...> >
{
using Container = std::basic_string<TAs...>;
template <typename... CtorArgs>
static void apply(Container& container, CtorArgs&&... args) {
container.insert(container.end(), { std::forward<CtorArgs>(args)...});
}
};
}
template <typename Container, typename... CtorArgs>
void add_to(Container& container, CtorArgs&&... args) {
traits::add_at_end<Container>::apply(container, std::forward<CtorArgs>(args)...);
}
int main()
{
using X = std::pair<int, std::string>;
std::vector<X> v;
std::set<X> s;
std::wstring wstr;
std::string str;
add_to(v, 12, "hello");
add_to(s, 42, "world");
add_to(wstr, L'!');
add_to(str, '?');
}
Basically, what you do, is have a free-standing utility function add_to that uses a trait class traits::add_at_end that can be specialized (in this case for any vector<...>, set<...>, or basic_string<...> template instances.
In practice, you would share the implementation for similar containers (e.g. deque and vector) by inheriting the common implementation.
push_back is supported by std::string, std::vector, and std::list. With this, your class template is simply:
template<typename T>
class S {
T built;
public:
S() : built{} {}
void add_to(typename T::value_type e) {
built.push_back(e);
}
const T& get() {
return built;
}
};

How to determine the iterator type for container?

What do I have to write for YYY, ZZZ in order to set iterator_type to the iterator type associated with T ? It should work in Visual Studio C++ 2010, if possible (but general standard solution is ok as well).
template<class T>
struct iterator_for {
typedef YYY<T>::ZZZ type;
}
Hence I want:
iterator_for<double[3]>::type is double *
iterator_for<std::string>::type is std::string::iterator
iterator_for<char[12]>::type is char *
etc.
I have a templated wrapper class Wrapper<T> storing something iterable (i.e. a container or string or an array) and I want to define a function returning an iterator pointing into the wrapped object. For that purpose, I need to be able to speak about the iterator type corresponding to T. For an array, the corresponding iterator would be a pointer, and for a string that is whatever string defines to be its iterator type.
My solution is:
typedef decltype(std::begin(std::declval<T&>())) type;
The iterator type is the type returned by std::begin when called on an instance of T. std::declval<T&> is declared to return a T&, on which we can call std::begin.
This differs from JohnB's answer in that we pass a reference to the container type to std::begin. This is needed because:
std::declval<T>() returns an rvalue reference (T&&).
std::begin takes its argument by non-const reference.
A non-const reference can't bind to a temporary.
std::declval<T&> however returns an lvalue refernce because of reference collapsing.
If you just want to separate containers from pointers, you can try this
template<class T>
struct iterator_for
{
typedef typename T::iterator type;
};
template<class T>
struct iterator_for<T*>
{
typedef T* type;
};
template<class T, std::size_t N>
struct iterator_for<T (&)[N]>
{
typedef T* type;
};
Ok, one possibility probably is (in C++11, but does not work in VS 2010):
typedef typename std::remove_reference<
decltype (
begin ( std::declval<T> () )
)
>::type
type;
The Boost library already has this:
#include <boost/range.hpp>
#include <iostream>
#include <string>
#include <vector>
template <typename T> void print_list(const T& container) {
typedef typename boost::range_iterator<const T>::type iter;
for (iter i = boost::begin(container); i != boost::end(container); ++i)
std::cout << *i << ";";
std::cout << "\n";
}
int main() {
double array[] = {1.0,2.0,3.0};
std::string str = "Hello";
std::vector<int> vec(3, 10);
print_list(array); // prints 1;2;3;
print_list(str); // prints H;e;l;l;o;
print_list(vec); // prints 10;10;10;
}