Relational Database
A relational database is one that is made up of many tables representing different sets of resources. These tables are linked together in one-many or one-one relationships. Tables are made up of many rows. For each token resource there is a corresponding row.
For example, a hospital database might have a table called “doctors” and a table called “patients”. An example table that might link the two might be “appointments”. That is, one row in the doctors table might correspond to many rows within the appointments table. Similarly, one row in the patients table might correspond to many rows in the appointments table. In this way, doctors and patients can be said to have a many-many relationship.
Normalising and denormalising
The process of deciding how to separate data out of a ‘flat’, spreadsheet-like format into separate tables is called normalisation. The process of combining table entries from multiple tables into single tables (usually for performance reasons) is called denormalisation.
Querying
To query a relational database one usually uses SQL, or Structured Query Language. The SQL Wikipedia page details some of the history of SQL, leaving out the interesting part about its roots in propositional logic.
Example relational database management systems (RDBMS)
In comparison to document-based databases
Document based databases only store documents and usually have no formal system of indicating relationships between them. This simpler storage method is a good thing, depending on the application. Document-based database systems cannot be queried with SQL but require a different approach to writing and performing queries, such as “map-reduce”.