You are given the following information:

The table Dog has attributes id, breed, name, and age; all are string fields of the same length.

The id attribute is a candidate key.

The relation contains 10,000 pages and each page has 100 tuples.

The indices indicated below are in memory.

You can assume that the size of RID pointers in the index is ignorable.

Consider the following query: SELECT D.breed, D.id FROM Dog D WHERE D.breed=Poodle; Assume that only 5% of Dog tuples meet the selection condition.

(a) Suppose that a clustered B+ tree index on breed is the only index available. Describe the possible query plans (if any) under Alternatives 1 and 2 and their costs. Compare the two. Does using indexing help? (b) Suppose that an unclustered B+ tree index on breed is the only index available. Describe the possible query plans (if any) under Alternatives 1 and 2 and their costs. Compare the two. Does using indexing help? (c) Suppose that a clustered B+ tree index on (D.id, D.breed) is the only index available. Describe the best query plan and its cost.