Question

I have two arrays or vectors, say:

  int first[] = {0,0,1,1,2,2,3,3,3};
  int second[] = {1,3};

I would like to get rid of the 1s and 3s in the first set, set_difference can only get rid of the first occurrences of these values however this is not what I want to have.

Should I do this with remove_copy by iterating on the second range and each time remove one entry from the first set.

What would be the best way to do this in C++?

Answer 1

Write a specialised set_difference:

template <typename InputIterator1, typename InputIterator2, typename OutputIterator>
OutputIterator set_difference_any( InputIterator1 first1, InputIterator1 last1,
                                   InputIterator2 first2, InputIterator2 last2,
                                   OutputIterator result )
{
  while ( first1 != last1 && first2 != last2 )
    if ( *first1 < *first2 ) {
      *result = *first1;
      ++first1;
      ++result;
    } else if ( *first2 < *first1 ) {
      ++first2;
    } else {
      ++first1;
      //++first2; // this is the difference to set_difference
    }
  return std::copy( first1, last1, result );
}

Then apply it to the problem:

#include "set_difference_any.h"
#include <boost/range.hpp>
#include <iterator>
#include <vector>

std::vector<int> result;
set_difference_any( boost::begin( first ), boost::end( first ),
                    boost::begin( second ), boost::end( second ),
                    std::back_inserter( result ) );

This algorithm is linear (max. last1-first1 + last2-first2 comparisions)