The best way to divide a vector into two smaller tables?

advertisements

What I'm trying to do:

I am trying to split a vector into two separate arrays. The current int vector contains an element per line in a text file. The text file is a list of random integers.

How I'm planning to do it:

My current idea is to create two regular int arrays, then iterate over the entire vector and copy n/2 elements to each of the arrays.

What I would like to know:

What is the most elegant way of accomplishing my task? I have a feeling that I can do this without iterating over the vector multiple times.

Code:

#include <vector>
#include <fstream>
#include <iterator>
#include <iostream>
using namespace std;

vector<int> ifstream_lines(ifstream& fs)
{
  vector<int> out;
  int temp;
  while(fs >> temp)
  {
    out.push_back(temp);
  }
  return out;
}

vector<int> MergeSort(vector<int>& lines)
{
  int split = lines.size() / 2;
  int arrayA[split];
  int arrayB[split];
}

int main(void)
{
  ifstream fs("textfile.txt");
  vector<int> lines;
  lines = ifstream_lines(fs);

  return 0;
}

Thank you :)


Use iterators.

std::vector<int> lines;
// fill
std::size_t const half_size = lines.size() / 2;
std::vector<int> split_lo(lines.begin(), lines.begin() + half_size);
std::vector<int> split_hi(lines.begin() + half_size, lines.end());

Since iterator ranges represent half open ranges [begin, end), you don't need to add 1 to the second begin iterator: lines.begin() + half_size isn't copied to the first vector.


Note that things like

int split = lines.size() / 2;
int arrayA[split];
int arrayB[split];

Are not standard C++ (and as such not portable). These are so-called variable-length arrays (VLAs for short) and are a C99 thing. Some compilers have them as an extension while compiling C++ code (GCC, Clang). Always compile with -pedantic to get a warning. These VLAs act funky for non-POD types and aren't generally useful, since you can't even return them.