Dimension Roles

I'm modeling health care data warehouse, which has patient and physician dimension. They share pretty much the same information: names,gender, dob, address, some identification number etc. And the question is whether to load them into different tables or combine into only one?
Sure, the attributes could change in future, but now they are the same?

Thanks

In a general sense, yes they are the same, but there are significant privacy and legal issues based on the role. Information about a physician in the role of a physician is not as protected as the information about the physician as a patient. Same goes with employees as employees and employees as patients. Managing security becomes much more complicated if you attempt to treat all within the same dimension.

Putting security aside I didn't come out with any reasonable argument that they can not be combined, even if the attributes differ a little bit.
I've already modeled them as separate tables, but felt tempted to create one table and add views on top of it.

So thanks for pointing that out.

I wouldn't combine them into 1 dimension.

Combining them into 1 dimension doesn't really save any space. Dr. Johnny Fever the Physician and Dr. Johnny Fever the pateint will be 2 rows because Physician and patient have slightly different attributes.

On a fact table, you would have to have a dimension key for Physician and one for patient, so no savings there.

There's going to be a lot more patients than physicians and physicians will be queried much more frequently. Physicians will account for what 10% of the rows. You could create views and add indexes to improve performance, but they'll be a point where the work to make the views perform as fast as seperate tables is more than the work to make them seperate dimensions.

What Jeff said. They are different things. People often want to do the same thing for code tables. Bad idea. Additionally, Physician and Patient will be related to different facts. You've abstracted the model with no added value.