My web application will have various types of users: a root user, sys-admins, customers, contractors, etc. I’d like to have a table “User” that will have the columns “Username”, “Password”, and “Role” with role being one of the aforementioned roles. I will, for example, also have a table called “Customer” for storing customer-specific attributes.
Here’s my question. Since all usernames will be unique, I could use the username as the primary key for the User table. Would it make more sense, though, to create a “User ID” column and use it rather than the Username column as the PK? There are many other tables that will use this User PK as a foreign key and I would think that, from a performance standpoint, it will be faster to compare user IDs rather than username text strings.
Thank you for your response.
I would create the userid because while usernames might be unique, they are not generally unchangeable and I would prefer not to have to change thousands or millions of records because HLGEM wanted to become HLGEM_1. Further joins on integers are faster.