Relational Database Anti-Patterns

In this article, I will review few of database design anti-patterns and some solutions that avoid those problems. Nowadays, rare application does not have database on the backend, and while relational databases exist for a long time, it is still the most used type of database. Relational databases have often replaced other types of databases because they are easier to understand and use. On the other side, because of this easiness in the overall concept many anti-patterns are overlooked and used in many systems. Introduction If you know the patterns, it is very good, if you know the anti-patterns it is even better. Anti-patterns exist everywhere, in every aspect of our lives and it is crucial to know them and use techniques and solutions to avoid them. Relational Databases exist for very long time, but still, many people use not optimized techniques to design and develop databases. It is very common for people to stick with the way they used to do it and not to question themselves: “is it the really best way to do it?” Optimization of developing techniques helps to avoid lifetime long optimization of the code. There are many well known database anti-pattern related to database design, development and use in the application. This article will concentrate on few database design anti-patterns. Anti-patterns 1. Assigning value to nothing (NULL) Null should not have any value in database logic; this approach is bad for anything, so it should not be used in database either. 2. Attribute with multi-value (list of values) This anti-pattern is one of the main consequences of non-normalized database design. 3. Table is not a list (super-wide tables) Long table is good, wide table is bad. Anti-pattern 1:  Name Assigning value to nothing (NULL)  Problem / Bad solution Often databases designed with intend to use NULL for something with special meaning. While in reality, NULL means nothing then nothing cannot be something. Typical twotable database relationship is shown in Example 1.