Boolean dimensions

Hi there!

Quick question when I have a simple yes/no dimension for example is my product on sale or not what is best pratice to model this?

Thanks

Simmo

Make it the smallest numeric datatype that is still summable. I typically add these to the fact table since folks want to count them.

And it you have a lot of them, you can collect them in a junk dimension.

Thanks for the responses. What is best in terms of performance: degenerate dims for say 3 y/n fields in the fact table or a junk dimension?

My fact table will have around 325 million transactions for about 30 million unique orders...

Its probably a wash. Given its only 3 flags, I would leave them on the fact.