RichB (6/9/2010)
Thomas-282729 (6/8/2010)
Again, what you are suggesting is a poorly written query where the developer uses the wrong alias (A.Id instead of B.Id). How does that improve clarity in the scenario where say "EmployeeId" exists in five or ten tables? It will be just as easy to mistakenly write A.EmployeeId instead of B.EmployeeId as it would writing A.Id vs B.Id.Well, in this case the difference between getting a.EmployeeId instead of b.EmployeeId is likely to be less significant than getting
a.ID when you wanted b.ID where a.ID comes from Employee a ([employee]Id, name, stuff) and b.ID comes from ProjectStaff([project]Id, employeeid, stuff)
Also - that bug is almost impossible to find, whereas with properly named columns its pretty straightforward.
How do you figure it is easier to find? Suppose the PK is named EmployeeId. It is just as easy to write A.EmployeeId instead of B.EmployeeId in the join as it would to write A.Id vs B.Id. In fact, I would say that having the PK be named "Id" makes it easier to see which column is the PK and which column is the FK. That would make it easier to find situations where the developer accidentally joined the column to itself (e.g. A.EmployeeId = A.EmployeeId).
Of course, the developer used a better alias, then it would be even easier to see a problem.