Matthew Kerridge

A relational database, one of several prevalent database types or models, serves as a good example for describing database concepts. The associated database software can be referred to as a relational database management system or RDBMS. Popular implementations include Microsoft SQL Server, MySQL, and Oracle.

When operating upon data through an RDBMS, there are two sides to the operation of which one should be aware. The server side handles the physical storage of the data. The client side interfaces with the end user. The user process may take on many forms, from a report job running on a mainframe to an interactive application on a wireless hand-held device. Because only the server side does the actual storing and retrieving of data, and only the client side talks to the customer, the two work in tandem to perform all operations.

Consider an example of a restaurant. Customers interface with the wait staff, who function as the client side. When a meal is ordered, the waiter delivers the order to the kitchen, which is equivalent to the database server. Once the chefs have prepared the meal, the waiter delivers it to the diners' table.

Just as there is a menu governing what may be ordered in a restaurant, specific actions may be performed on an RDBMS. The American National Standards Institute (ANSI) has defined a language of commands for interacting with an RDBMS called Standard Query Language or SQL. Although individual implementations of SQL may vary to some degree, they all allow the same basic operations.

The database server can handle multiple databases. For example, a state government could have separate databases for the state's parks, driver registration, and department of revenue, all being controlled by the same RDBMS server.

Within a database, data is stored in tables. Each table contains fields called columns. In the licensed drivers database, relevant columns for the drivers table would include first, middle, and last name, class of license, date of birth, and date of license expiration. Columns are defined as different types of data; the name fields would hold character data, while a specific type exists for holding the dates. Each licensed driver would be stored in a single entry or row in the table. Tables also contain a field called the primary key that ensures that each row in the table is unique; for the drivers table, the primary key would be the license number. An index can be added to one or more columns in a table to speed access to data.

Data resides in multiple tables within a database. The street, city, state and zip code for each driver could be stored in an address table, again including the license number as a primary key. It would then be possible to bring the name and address information together by linking the license number from the drivers table with the same column in the address table; this operation is called a join.

SQL statements include write operations to insert a new row, and update or delete an existing row, plus select statements for reading data from tables. Operations can either be performed on a single row, e.g. by specifying a license number, or on a set of rows, say all drivers who reside in a certain city.

Although specifics vary for other types of database software, many of the concepts remain the same.