I have a piece of code :

void open_mp_number_thread_test()
{
  #pragma omp parallel
 {
  
  omp_set_num_threads(4);
  #pragma omp for
  for(int i=1; i<4; ++i)
  {
  }

  omp_set_num_threads(6);
  #pragma omp for
  for(int i=1; i<6; ++i)
  {
  }
 }
}

Is this a good idea or practical to change the number of threads in the openmp code in order to adapt the maximal iterations in the for-loops ?