UTOUG Training Days 2019

Have you ever looked at a recipe book and noticed it includes what kind of pan to use?  Following the recipe works fine (provided you've got the 3-quart dutch-oven or whatever at hand), but the recipe often omits why one does each step.  But if one is aware of why things work the way they do, they can skip the recipe and choose how to proceed.  Things aren't so different in our SQL recipes--The docs describe what to expect, but not always why to expect it.  The goal of this session is to dig into the why and the how with aggregation, in an effort to make query-writing with aggregates a more conscious endeavor, with fewer surprises.

To explore, we'll build-out some reporting for comic books and the battle between operating systems, and will test these mechanics by creating some of our own recipes (aggregators) along the way.  We'll compare aggregate vs analytic functions' relative costs from a mechanical perspective, and look at what the sql engine is doing along the way.  We'll look at the mechanical impact of windowing, and exlore order of operations, mechanics of parallel aggregation, impact of multidimensional aggregation, index use and cost implications along the way.