Exam 70-761 topic 1 question 37 discussion

I think the correct answer should be store procedure because it satisfies the 2 requirements: 1) The users can't change the table directly 2) It also updates the related tables With a view, you can only update one table per time.

There is no correct answer listed. To maintain a set of denormalized tables, normally we need an instead of trigger on primary table to interpret every insert/update/delete operations on this table and make sure it will map to get reflected to related redundant columns in other tables. I.e. an after trigger on table is necessary and should be an answer/solution to this question. Kindly note: For the solution mentioned in the stackoverflow link, an indexed view is just used to bridge the gap for querying when there are denormalized tables but it won't help when there are insert/update/delete happens on primary table. Unless we go step further and create an instead of trigger on that indexed view and encapsulate every details of underlying table from Application logic for not only select but also I/U/D operations. Even in that case the answer should be View and Trigger other than just a view which is not a complete solution.

Stored procedure if you want to update multiple tables while obscuring the database schema to the user. Only option since triggers aren't listed.

basic work for store procedure/trigger...

I think view is correct, because you can modify the insert into multiple base tables through an instead of trigger.