Looping through a vector of vector

Should we looping like declaring another vector:

Code:

---

#include <vector>
#include <iostream>

using namespace std;

int main() {
    vector< vector<int> > ph;

    vector<int> p;
    p.push_back(1);
    p.push_back(2);
    p.push_back(3);

    vector<int> q;
    q.push_back(10);
    q.push_back(20);
    q.push_back(30);
    q.push_back(40);

    ph.push_back(p);
    ph.push_back(q);

    for (vector< vector<int> >::size_type u = 0; u < ph.size(); u++) {
        for (vector<int>::size_type v = 0; v < ph[u].size(); v++) {
            cout << ph[u][v] << " ";
        }
        cout << endl;
    }
}

---

or is it ok just simply with an unsigned int:

Code:

---

for (unsigned u = 0; u < ph.size(); ++u) {
  for (unsigned v = 0; v < ph[u].size(); ++v) {
    ...
  }
}

---

Practically speaking, either is fine. If you want to be super correct, use the first.

Hell, I sometimes use plain old int, but you need to be aware of the size of the vector and know that that's a safe thing to do.

In case you're wondering why I would ever do that, it's because of patterns like this:

Code:

---

int bestIdx = -1;
for( int i = 0; i < (int)vec.size(); ++i )
{
    if( MeetsSomeCondition( vec[ i ] ) && ( bestIdx < 0 || BetterThan( vec[ i ], vec[ bestIdx ] ) )
        bestIdx = i;
}

---

In other words, I want to use -1 as a placeholder for "don't have a result yet"

Thank you very much guys! :)

It was working fine with an int but i told myself better be safe than sorry. :)