On one of my first database projects, I developed an absurdly inefficient SQL database for time series data. Oh but it had a beautiful, easily generalizable, fu

On one of my first database projects, I developed an absurdly inefficient SQL database for time series data. Oh but it had a beautiful, easily generalizable, fully normalized schema which I was so proud of... Who cares that it requires a complex join before every insertion? And that multiple tables needed to be updated to keep track of metadata for each insertion? It went into production without any load testing, and it struggled to keep up with even a few dozen time series. I suppose the good news was that the product it was developed for never got many users, so it just hobbled along after applying some caching hacks. And of course I learned an important lesson about considering database access patterns.