There is also the Real-Time to Offline Job that takes data from real-time tables and writes it to offline-tables. It creates the new segments based on time boundaries as well, which I think means that you could configure it to create offline segments that contained more rows than the real-time segments.