Question

In a SQL Server 2005 database, I have lots of tables like this Products table

ProductID (PK)
ProductCategoryID (IX)
Description
Price
ExpiryDate
BreakableYN

…where there is a primary key, a foreign key and then a bunch of other fields. Another characteristic of this type of table is that lots of queries only use the 2 ID fields (ProductID, ProductCategoryID), e.g. Employees JOIN EmployeeProductJoin JOIN Products JOIN ProductCategories JOIN ProductDepartments.

If ProductID and ProductCategoryID are already indexed, is it worth adding another index for ProductID, ProductCategoryID?

I know it seems that I’m asking if adding a covering index will help, but what I’m really asking is whether a covering index will help if the fields in that covering index are already indexed individually.

These are definition tables that are not huge, so I’m not worried about adding extra time to INSERTs etc.

Answer 1

Is the primary key clustered? If it is, then adding a new index will accomplish nothing, because the ProductCategoryID index will already contain the ProductID values, so it effectively “covers” both columns.