An optimization that made Pivot significantly faster on GROUP BY queries with a Top-K operator above them, and a walkthrough of how analytical databases run GROUP BY in the first place.