New Table vs. New Schema

Question

Suppose I have a schema with many related tables: users, cities, items, purchases, etc. I now want a table in my database that contains solely event logging data for my internal support. The rows of information in the logging table are self-contained, not at all relational, and unrelated to my other tables. Is it better to create a new table in my existing schema, or to create an entirely new schema? Is one method preferred over the other? Is there a cost associated with one over the other?

Answer 1

In my opinion it all depends on the size of your database. If you are managing dozens of tables with millions of rows of data then you will probably have an easier time isolating and managing these logging tables into their own schema/database. If you are just managing a small app then don't worry, put everything into one database/schema. If your database is large or your anticipate your database becoming large then break them out. Once they are broken into separate entities you can easily manage the communication between the multiple databases/schemas using all kinds of great available tools.

Answer 2

In my opinion, if the data is unrelated, it belongs in a different schema. There's likely to be a very small overhead associated with creating a new schema, as opposed to having everything in a single schema, but I wouldn't have thought it was worth worrying about.