CS348 Midterm: Exam 1 solution.pdf
Document Summary
Suppose that a relational database contains the following large relation: The database has a clustered b+-tree index on the composite key releaseid,tracknum. For full credit, explain brie y (in one or two sentences) why the query would require less i/o if the clustered index were used. select * from track where releaseid > "xxx". The secondary index cannot be used to execute this query. Your example should also require fewer i/o operations to answer using the clustered index than using a full table scan, i. e. , than not using either index. For full credit, explain brie y (in one or two sentences) why the query would require less i/o if the unclustered index were used. select (distinct) tracknum from track. This query can be answered using only the unclustered index - no access to the data would be required. It could also be answered using the clustered index, but that index will be larger.