CODEWITHSUNDEEP
c++pointersc++11vectorreference
gsamaras
gsamaras

Reputation: 73376

How to pass vector from a class to a function?

I have a class Point, which -for sake of simplicity- looks like this:

template<class DivisionSpace>
class Point {
 public:
  typedef typename DivisionSpace::FT FT;

  const std::vector<FT>* get_coords() const {
    return &coords;
  }
 private:
  std::vector<FT> coords;
};

Now, I want in main, to pass the vector of this class to a function which expects std::vector<FT>& q. I pass a reference so that I avoid copying. I could use a pointer, but this would mean that I have to change many lines of code (because of the structure of the project).

How can I do it? Is there maybe any "trick" with C++11?

EDIT

Here is the prototype of the function:

void search_nn_prune(std::vector<FT>& q,
                       std::vector<std::pair<float, int> >& res,
                       int max_leaf_check, bool sorted_results = false, int k =
                           1,
                       float epsilon = 0) {

In main I do

  std::vector< Point<Division_space> > q;

  std::vector<std::vector<std::pair<float, int> > > results(Q);
  for(int i = 0; i < Q; ++i) {
    const std::vector<FT>* query = q[i].get_coords();
    kdf.search_nn_prune(query, results[i], max_leaf_check, false, k, epsilon);
  }

and here is the error

error: no matching function for call to ‘Random_kd_forest<Division_Euclidean_space<int> >::search_nn_prune(const std::vector<int, std::allocator<int> >*&, std::vector<std::pair<float, int> >&, int&, bool, int&, float&)’
note: candidates are:
note: void Random_kd_forest<DivisionSpace>::search_nn_prune(std::vector<typename DivisionSpace::FT>&, std::vector<std::pair<float, int> >&, int, bool, int, float) [with DivisionSpace = Division_Euclidean_space<int>, typename DivisionSpace::FT = int]
note:   no known conversion for argument 1 from ‘const std::vector<int, std::allocator<int> >*’ to ‘std::vector<int, std::allocator<int> >&’
note: void Random_kd_forest<DivisionSpace>::search_nn_prune(size_t, std::vector<std::vector<std::pair<float, int> > >&, int, bool, int, float) [with DivisionSpace = Division_Euclidean_space<int>, size_t = unsigned int]
note:   no known conversion for argument 2 from ‘std::vector<std::pair<float, int> >’ to ‘std::vector<std::vector<std::pair<float, int> > >&’

Upvotes: 0

Views: 1495

Answers (2)

6502
6502

Reputation: 114529

You should declare the function accepting a const reference to the vector

void search_nn_prune(const std::vector<FT>& q, ...

and the you can pass

kdf.search_nn_prune(*query, ...

(note the * dereferencing star)

Upvotes: 1

Wyzard
Wyzard

Reputation: 34563

The problem is that your search_nn_prune function wants a vector<FT>& argument, but you're passing it a const vector<FT>& instead. You can't pass a reference to a const object into a function that wants a modifiable one.

If search_nn_prune isn't supposed to modify the vector<FT> it's given, add const to the argument in the function's declaration. If it is supposed to modify the vector, you'll need to decide on how you want to resolve the situation: Point doesn't allow changes to the coordinate vector it returns, but you want to pass it to a function that will change the coordinates.

Functions that take arguments by reference should generally take const references, unless the function is intended to modify the object it's given.

Upvotes: 1

Related Questions