Characteristics of Employee table:
Key values in root node of B-tree after new index results in promote/split all the way to the root.
This was ambiguous, since there are a several ways of interpreting B-tree splitting.A simple view would be to say 21 overfills the (19,20,22) node, so we promote 20. This causes (7,13,19) to overfill, so we promoted 13. Thus 13 ends up in the root node (13,25).
An alternative view (not really valid for this question, but reasonable) is that we "temporarily" insert the new key into the node and then work out the "middle" node. If the node becomes (19,20,21,22) then we could choose either 20 or 21. If we choose 21 and push it to (7,13,19,21), then we're likely to choose 19 (for consistency with the choice below), so 19 ends up in the root node (19,25).
I accepted both as correct.
Multi-attribute hashing query. How many pages are accessed to answer a query?
Storage requirements for a bitmap index