Data Normalization

Learned from SE464.

Data Normalization

A normalized relational database has no duplication of data.

• References (foreign keys) point to shared data
• In effect, many users are related to each other by all being linked to that industry
• To optimally partition the rows into shards, we could solve a balanced graph partitioning problem.

SQL uses normalized data. NoSQL uses denormalized data.

#serendipity I try to make my obsidian normalized, but sometimes, it’s just easier to have it denormalized. Places like CS247 or SE464 where I just have all the notes in a centralized location, but also make separate pages for later.