Question

This is a contrived example, but consider a scenario where employees have working locations, and where employees have managers who also have working locations:

create table WorkingLocation (
    ID int not null primary key identity(1,1), 
    Name varchar(50)
)

create table Employee (
    ID int not null primary key identity(1,1), 
    Name varchar(50), 
    WorkingLocationID int null,
    ManagerID int null,
    constraint FK_Employee_WorkingLocation foreign key (WorkingLocationID) references WorkingLocation (ID),
    constraint FK_Employee_Manager foreign key (ManagerID) references Employee (ID)
)

Now consider a business rule that wants the employee’s WorkingLocation, but will settle for his manager’s WorkingLocation if he doesn’t have one. At this point you have two options:

1: Have a query that gets both and let business rules decide which to use:

select 
    e.*,
    emp_location.*,
    mgr_location.*
from Employee e
    left join WorkingLocation emp_location on e.WorkingLocationID = emp_location.ID
    left join Employee mgr on e.ManagerID = mgr.ID
    left join WorkingLocation mgr_location on mgr.WorkingLocationID = mgr_location.ID
where e.ID = @id

2: Make separate calls to the database to retrieve the manager’s details if the employee has no WorkingLocation set.

Which do you prefer and why?

Answer 1

There is another option – specify the rule in the T-SQL query using COALESCE or use the null-coalescing operator ?? in your code (also works in LinqToSQL).

Either of these will only then require one call the database, so it’s +1 for option 1.