How do you process a million records into your database a day? Or bulk load several Terabytes of historic data? SQL is "the fastest", PL/SQL gives you far more control. I've been designing and working with VLDBS for twenty years and I've made enough mistakes to learn a few things. In this presentation I will show, with examples from real projects, ways of using SQL and PL/SQL to bulk process data, critical considerations and real-world lessons on how to manage the loading of data.