Modeling Semi-Structured Data

So far, this book has focused on modeling structured data, the kind used in relational databases since the early 70s. However, with the rise of the internet, a different style of data became prevalent: semi-structured. Semi-structured data, such as website traffic and social media feeds, contain some organizational structure but do not conform to the formal structure of a relational database.

New file formats also emerged to support this new type of data, starting with the advent of Extensible Markup Language (XML) in the early 2000s, followed by JavaScript Object Notation (JSON), and, with the rise of distributed computing, formats such as Avro, ORC, and Parquet. These formats offered a lightweight and flexible way to structure data, making them ideal for web-based and mobile app data.

The popularity of semi-structured data can be attributed to its flexibility, adaptability, and ability to handle data sources that do not fit neatly into traditional...