Consider the following small modification to the above query: we add an unclustered b tree index on s.d as well as a clustered b tree index on r.b. Without re-computing the new best plan, explain how these changes affect the optimization process for this query?