Question

I want to update a list in the database which has one to many relationship.

For example, There are over 100 subjects and a student can select whatever he wants. Let’s say next time that student edit his selection.

What is the way you would update the database for this case? What I am used to doing is delete all the past selections and add all the newly selected subjects. This is pretty easy logic but I am just wondering whether there is a better way to do this in terms of performance. In my case ‘delete’ is a single database call and inserting all is a loop in a single database connection.

I understand we can add a logic to identify the deleted records and newly added records. Is it realy worth to do that? What you would do for similar cases?

Thanks!

Answer 1

For brevity and simplicity, there is nothing really wrong with a delete-reinsert approach, barring foreign key hooks from child tables or triggers firing for each (re)insert.

Another way to do it could be to pass the new set as a table-valued parameter to a SP to Sql Server, which would then use a duo of queries to make the change

delete many_table
where studentid=@studentid and someid not in (select someid from @tableparam p)

insert many_table(studentid, someid)
select @studentid, p.someid
from @tableparam p left join many_table t on p.someid=t.someid and t.studentid=@studentid
where t.someid is null

(there is a more elegant MERGE syntax for SQL Server 2008, which is yet another option, further testing your TSQL skills)