It's been a while since I 've been here so I apologize if this is in the wrong section.
This is a general design question about a relational structure to support a cube. I've simplified it.
Let's say I got a relatively large fact table that can be "filtered" by a variety of "flags" (bit columns) (among other dimensions like dates and Ids). ie: IsInvoiced_Ind, IsOrdered_Ind, etc. There are millions of rows.
What is the best-practice design for the corresponding dimension table?
Is a view on the fact table including the primary key and only all bit indicators appropriate (it seems to take forever to process), or should there be a seperate dimension table or view for each indicator?
I'm kind-of stuck. I can't get this large amount of data to process, and the time is related to this particular large fact table. I'm associating each combination of PK and indicator in the DSV.