The Purpose and Features of a Relational Database
By John Lister
A relational database is one that offers extremely complex and sophisticated queries and searches thanks to two factors: tables and cross-referencing. It stores data as tables rather than plain lists, making it easier to filter individual elements of each record. It also allows cross-referencing between different sets of data.
A basic database stores all the details in a single file, made up of a string of records. For example, a police database might have a list of records of individual crimes, each with details such as the name of the offender, the type of crime and the date and location. A traditional database would simply be an electronic equivalent of a collection of forms or cards and in technical terms would be similar to a plain list in a word processor document. If you wanted to find a particular pattern such as all offenses committed by a particular person, you would need to search through each mention of the name and then copy all the details from each crime record.
Tables in Relational Databases
A relational database puts information into tables rather than a straight list. In such a table, each column represents a particular type of information (a field) and each row lists one record. In the crime example, you could have columns for the type of offender, the type of crime, the date and the location, with each crime on a separate row. This structure means the computer can carry out more sophisticated searches. For example, it can put together a list of all the rows featuring a particular offender's name, which creates a detailed history of that offender's crimes. Most databases allow more sophistication and can search for records that meet multiple criteria. For example, a database could put together a list of records that involve a particular crime and a particular data to show if certain offenses are seasonal, which could help police chiefs assign officers more effectively.
Cross-Referencing in Relational Databases
In practice, relational databases can be even more sophisticated than simply using separate tables. Many will be made up of different tables that can be cross-referenced. For example, as well as a crime table, you could have a table for medical treatment, one for school records and one for tax revenues. The term "relational" comes from the way the database is set up to recognize common factors between each database. In the example, the database would know that a particular city mentioned in each table refers to the same place. This can allow extremely complex queries, for example to find out whether cities that have poor school performance are more likely to have high levels of a particular crime.
While computer databases were originally designed to be standalone software, relational databases are now a key part of many websites. For example, an online retailer could have a database that brings together tables covering individual purchases, reviews and product offers. This technology means websites can quickly create highly personalized pages: for example, a book website could look at a customer's past purchases, cross-reference to find people who have bought the same book, find other books those people have bought, then pick out books from this list that are on special offer and suggest them to the customer.
A professional writer since 1998 with a Bachelor of Arts in journalism, John Lister ran the press department for the Plain English Campaign until 2005. He then worked as a freelance writer with credits including national newspapers, magazines and online work. He specializes in technology and communications.