So I am working on query plans and understanding the I/O cost model for Clustered and Unclustered (non-clustered) B+-tree indices for database relations. The scenario is I want to do a lookup for a certain attribute for a relation. Each index is on a non-key of this relation. I have then been told that the costs are: .Clustered - Height(I) + #matching leaves .Unclustered - Height(I) + #matching records The thing that is confusing me is that I don't understand how this lookup even works. How is that we can actually find the correct tuples when the index is not on the correct attribute we are looking for. Would it not be the case that we have to examine every page of memory or is there a way we can exploit the B+-tree indices.



Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.