7

Using C++, I want to create an array which only contains the first n elements of another array. Like so in Scala: val arrayTwo = arrayOne.take(n)

I know I can use a loop and copy the elements one by one, but this is much more complicated than necessary, it takes unnecessary space, and that makes it less readable. Is there a simple, readable function to create a new array from the first n elements of a given previous array? Also I would like to reuse a function from somewhere, rather than writing one on my own, because I don't want to pollute the namespace unnecessarily. Performance doesn't matter as long as it takes O(n).

std::copy_n looked like it, but I can't get it to work because std::back_inserter for some reason doesn't accept my array (I also tried with a pointer instead of array, still not working).

This is my best attempt so far.

#include <iostream>
#include <utility>
#include <algorithm>
#include <vector>
#include <iterator>
#include <stdio.h>
#include <math.h>

using std::pair;

int main() {
  pair<double, double> fabricatedPoints[] = { { 15.3, 12.9 }, { 88.6, 56.0 },
            { 0.4, 18.0 }, { 5.0, 3.13 }, { 2.46, 86.01 } };
  pair<double, double> points[] = {};
  std::copy_n(std::begin(fabricatedPoints), 3, std::back_inserter(points));
}

It can be done either with copy_n, or by other means, I don't mind as long as it is readable. If there exists no readable solution in libraries (not necessarily the standard libraries - it could also be Boost or something, as long as it's a widely used library), then I will accept an answer which provides convincing evidence of no such solution existing.

T.C.
  • 133,968
  • 17
  • 288
  • 421
Velizar Hristov
  • 662
  • 2
  • 10
  • 23

4 Answers4

12

If you were using vectors (and you should, you're using C++), you could just do this:

using std::vector;
vector<pair<double, double>> a{ {15.3, 12.9}, ...};
vector<pair<double, double>> b(a.begin(), a.begin() + 3);

For arrays you will have to make sure to preallocate the array to the right size:

pair<double, double> b[3];
std::copy_n(a, 3, b);
nneonneo
  • 171,345
  • 36
  • 312
  • 383
1

You can't append to normal C-style arrays like points (actually, I would be surprised if the declaration didn't generate compiler errors). Trying to append to a C-style array would write beyond the bounds, leading to undefined behavior (and here too I'm surprised that std::back_inserter would compile when passed a C-style array).

Instead use a std::vector.

Some programmer dude
  • 400,186
  • 35
  • 402
  • 621
  • "I'm surprised that std::back_inserter would compile when passed a C-style array" I think it doesn't: "std::back_inserter for some reason doesn't accept my array". – svick Sep 15 '14 at 01:47
  • I know. However, std::back_inserter demanded an array type; and the code failed before the size of `points` was taken into account, so I omitted that. – Velizar Hristov Sep 15 '14 at 01:49
  • 1
    @svick Oh, missed that, it's much to early (or alternatively way to late) for me. – Some programmer dude Sep 15 '14 at 01:50
1

From C++20 ranges (#include <ranges>)

Simply use take_view

//DATA
std::pair<double, double> fabricatedPoints[] = { { 15.3, 12.9 }, { 88.6, 56.0 },
    { 0.4, 18.0 }, { 5.0, 3.13 }, { 2.46, 86.01 } };

//USE take_view
auto points = std::ranges::take_view(fabricatedPoints, 2);

//TEST THEM
for (auto p : points)
{
     std::cout << p.first << "  " << p.second << std::endl;;
}

Or using view adapters

//YOUR DATA
std::pair<double, double> fabricatedPoints[] = { { 15.3, 12.9 }, { 88.6, 56.0 },
        { 0.4, 18.0 }, { 5.0, 3.13 }, { 2.46, 86.01 } };

//GET FIRST TWO POINTS
auto points = std::views::all(fabricatedPoints) | std::views::take(2);

//TEST THEM
for (auto p : points)
{
    std::cout << p.first << "  " << p.second << std::endl;;
}
Pavan Chandaka
  • 11,671
  • 5
  • 26
  • 34
0

I would use vector for this

vector<int> vec1;
vector<int> vec2;
vector<int> merged;

//insert(where you want to start adding, from which index, to which index)
//in this case we are adding the first to n-th elements from vec1 to the last element of merged 
merged.insert(merged.end(), vec1.begin(), vec1.begin() + n);

//here we are adding the m-th to n-th elements of vec2 to the first element of merged
merged.insert(merged.begin(), vec2.begin() + m, vec2.begin() + n);