People commonly take any large collection of information to be a database, regardless of how the information is actually organized. Most IT professionals, however, define the term in a more elaborate wording.

When we speak of databases, it’s not so much the amount of data that matters. It’s more about the structure of the data at hand, and how this structure or schema facilitates quick lookups, inserts, updates, searching and data mining in general.

The database engine

For a typical database system, the term «structure» applies to the way our data is organized in tables, rows and columns. How the binary data is actually laid out in the physical data file(s) is another story, and fortunately of no concern to us.

We interact with the data through a database engine, or a DBMS (Database Managment System), that takes care of the file-level stuff. We tell the engine to execute our queries, and the engine reports the results back to us.

Relations

The most common type of DBMS’ are so-called relational database systems, where different information fragments connect to each other in terms of relations. In the vehicle registry, you may have a person related to a car. In the cadastre, you may relate geographical areas to one or more owners.

Relations can go across different tables, and even across different databases. Such coupling of various data sources is an important method in database journalism.