Optimize your code by reducing disk access

When writing an algorithm that requires informations from a database you should always prefer collecting the required data first in order to feed your code.

Instead of:

newCities.forEach(city -> {
   City cityById = cityDao.findById(city.getId());
   ...
});

prefer:

List<City> cities = cityDao.find(filter);
newCities.forEach(city -> {
   City cityById = cities
                         .stream()
                         .filter(c -> c.getId().equals(city.getId()))
                         .findFirst()
                         .orElseThrow(() -> new RuntimeException("City not found"));
    ...
});

It does not mean you should always load data from your database and then filter it in Java ! Collect only the required data needed by your algorithm when making the database query.