Question

I’m aware of IDENTITY fields but I have a feeling that I couldn’t use one to solve my problem.

Let’s say I have multiple clients. Each client has multiple orders. Each client needs to have their orders numbered sequentially, specific to them.

Example table structure:

Orders: OrderID | ClientID | ClientOrderID | etc... 

Some example rows for this table would be:

OrderID | ClientID | ClientOrderID | etc... 1 | 1 | 1 | ... 2 | 1 | 2 | ... 3 | 2 | 1 | ... 4 | 3 | 1 | ... 5 | 1 | 3 | ... 6 | 2 | 2 | ... 

I know the naive way would be to take the MAX ClientOrderID for any client and use that value for INSERTs but that would be subject to concurrency issues. I was considering using a transaction but I’m not quite sure what the broadest isolation scope that can be used for this. I’ll be using LINQ to SQL but I have feeling that isn’t relevant.

Answer 1

Somebody correct me if I’m wrong, but as long as your MAX() call is in the same step as your insert, you won’t have a problem with concurrency.

So, you could not do

select @newOrderID=max(ClientOrderID) + 1 from orders where clientid=@myClientID;  insert into ( ClientID, ClientOrderID, ...) values( @myClientID, @newOrderID, ...); 

But you can do

insert into ( ClientID, ClientOrderID, ...) select @myClientID, max(ClientOrderID) + 1, ... from orders where clientid=@myClientID; 

I’m assuming OrderID is an identity column.

Again, if I’m incorrect on this, please let me know. Preferably with a URL