Time: 10:30 - 11:20
Room: New York
One of the new features in PostgreSQL 10 was the ability to create multi-column statistics, helping the optimizer to understand dependencies between columns, produce more accurate estimates and hopefully better query plans. In this talk I'll explain why we need this capability at all, how it works, which cases it already helps with, and what improvements are in the queue for future PostgreSQL versions.
The following slides have been made available for this session: