3 ms·
Is denormalization (e.g. to star/snowflake schemas) still common or advisable? I know the conventional wisdom is that it can improve performance, but how much s
by tfehring 5y ago
Is denormalization (e.g. to star/snowflake schemas) still common or advisable? I know the conventional wisdom is that it can improve performance, but how much scale does it actually get you in practice (a factor of 2? an order of magnitude?), and can it still be the right call for public cloud based workloads where it's trivial to scale vertically and relatively easy to scale horizontally?
- eloff 5y agoIt very much depends. You have to try and measure it. Sometimes it makes things slower even. In general it's one tool in the toolbox and never the first one reached for.