It depends. On a dimensional model, this is bad practice and/or bad database design.
On a transactional system, you could have an order header table and an order detail table. You'd join them but also create an alias of the order header table. You'd then build your order header dimensions from the original order header table and the order header measures from the order header alias table. Have a read up on chasm and fan traps for further information on that, if that's your problem.
In general, if you're joining the two, you're declaring that there MUST be a fact row in both tables to return data. It's something that will lead to far more errors than you would imagine.
What is the actual requirement?
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.