An architectural database refactoring is a change which improves the overall manner in which external programs interact with a database so as to improve your database design without changing its semantics.

Refactoring	Example
Add CRUD Methods. Introduce stored procedures (methods) to implement the creation, retrieval, update, and deletion (CRUD) of the data representing a business entity.
Add Mirror Table. Create a mirror table, an exact duplicate of an existing table in one database, in another database.
Add Read Method. Introduce a method, in this case a stored procedure, to implement the retrieval of the data representing zero or more business entities from the database.
Encapsulate Table With View. Wrap access to an existing table with a view.
Introduce Calculation Method. Introduce a new method, typically a stored function, which implements a calculation that uses data stored within the database.
Introduce Index. Introduce a new index of either unique or non-unique type.
Introduce Read Only Table. Create a read only data store based on existing tables in the database.
Migrate Method From Database. Rehost an existing database method (a stored procedure, stored function, or trigger) in the application(s) which currently invoke it.
Migrate Method to Database. Rehost existing application logic in the database.
Replace Method(s) With View. Create a view based on one or more existing database methods (stored procedures, stored functions, or triggers) within the database.
Replace View With Method(s). Replace an existing view with one or more existing methods (stored procedures, stored functions, or triggers) within the database.
Use Official Data Source (directly). Use the official data source for a given entity, instead of the current one which you are using.
Use Official Data Source (via replication). Use the official data source for a given entity, instead of the current one which you are using.