Question

I have a very large std::vector of std::vectors containing a fixed number of unsigned integers.

All vectors of uints are sorted ascending.

My current way to eliminate duplicate vectors is

unsigned int i = 0;
while ( i < new_combs.size() )
{
  unsigned int j = i + 1;
  while ( j < new_combs.size() )
  {
     unsigned int k = 0;
     while ( k < new_combs.at(i).size() && new_combs.at(i).at(k) == new_combs.at(j).at(k) )
        ++k;
     if ( k == new_combs.at(j).size() )
        new_combs.erase(new_combs.begin() + j);
     else
        ++j;
  }
  ++i;
}

here, new_combs is a vector containing vectors as mentioned above.

Is there a more efficient way to eliminate duplicates if the vector of vectors is unsorted?

Answer 1

A shorter way would be using <algorithm>:

std::sort(new_combs.begin(), new_combs.end());
new_combs.erase(std::unique(new_combs.begin(), new_combs.end()), new_combs.end());

Unless you specifically need a std::vector, you can use std::set to avoid duplicates.