Chapter 8: Querying Unstructured and Semi-Structured Data

Many of the world's most valuable datasets are loosely structured. They come from application logs, which don't conform to any standards. They come from event data generated by a system that users interact with, such as a web server, which stores how users navigate an organization's website. They can also come from an analyst generating spreadsheets on a company's financial performance. This data is usually stored and shared in a semi-structured format to make it easier for others to consume. Some query engines have evolved to fully support this semi-structured data.

When talking about structured, semi-structured, and unstructured data, there are many different definitions out there. For this book, structured data is stored in a specialized data format where the schema and the data it represents are one to one. The data is serialized to optimize how the data is read, written, and analyzed. An example...