Hi,
Can we join the dimensional table diectly in star schema…
latha08 (BOB member since 2008-04-09)
Directly to what?
Directly to the fact? Certainly.
From one dimension to another? Yes, if the circumstances call for it. Doing so will almost certainly end up creating loops in your universe which will have to be resolved, so you should have a good reason for doing so.
A snowflake structure is just fine, but I don’t think that’s what you’re asking about.
Dave Rathbun 
(BOB member since 2002-06-06)
My questions is, Is it possible to join one dimension to other dimension in star schema …
How the joins is good in snow flake schema…
latha08 (BOB member since 2008-04-09)
That will defeat the definition of star schema. It will be a snow flake.
I didn’t understand this.
zack 
(BOB member since 2007-08-02)
Agreed.
You will have a performance impact if you do this, as opposed to querying a pure star.
The occassional snowflake, IMO, is not necessarily a bad thing, but your data model and requirements drive whether they are a good idea or necessary…
Mak 1 
(BOB member since 2005-01-06)
You can join dimension tables together through factless fact tables to discover alternative scenario’s. (e.g. the max available sales result given stock levels at that time vs. actual sales etc).
Joining two dimenions like “Customers” or “Stores” or “Stores” and “Products” means you are creating a relationship between these entities. You need to provide a context to that relationship for it to be meaningful. A factless fact table or an actual fact table can provide that context to ensure that the relationship between the entities is meaningful.
A snowflake can join dimensions to dimensions but that too is within a specific context. For instance a parent - child relationship.
ErikR 
(BOB member since 2007-01-10)
good, consise explanantion Eric… 
.
Mak 1 (BOB member since 2005-01-06)