A common challenge for database modellers is deciding which data type is the best fit for a particular column. It is a problem which involves consideration of both the properties and the scale of the data that will be stored, and in no case is this more evident than when handling numeric values because of the large variety of alternatives that most relational databases provide for their storage.

When designing a database, early decisions can have a huge impact on the performance and storage requirements. These decisions can be difficult to change later, as most subsequent work will depend on the physical model. This article highlights some common design decisions, flaws, and misconceptions. Creating a Primary Key: Uniqueidentifer or Integer When defining a surrogate primary key for a table, two options are the most common: Integer and UniqueIdentifier (aka.