In some environments, with millions of rows of data on the tables, that there are no primary keys. They are using just indexing for reporting applications. Is this most efficient for read only environments?
asked Dec 06, 2017 at 09:15 PM in Default
It's the good old answer of "It depends".
You don't always need a primary key, a clustered index is more important. You only need a PK if your doing things like replication, or foreign key constraints, but they can get around with a unique key.
Does the table have a clustered index? Are the indexes that are defined sufficient to satisfy the queries being run?