Shrunken Dimension: Model and loading

Hello:

We have the following model:
Fact_Order->Dim_Item

Dim_Item contains the attributes Item,Subcategory and Category

Now we have to load two new facts:
Fact_Budget: Category level
Fact_Invoice: SubCategory Level

What we have to do? Create two new dimensions Dim_Category and Dim_Subcategory from Dim_Item? We have to repeat the values of category on Dim_Subcategory for example?

Thanks

Invoicing is typically at the line level, which is usually item. If you are aggregating the source at the sub category level, I would suggest you don't.

But if that is what you have, then yes, you would create category and sub category dimensions. I would populate category info in the sub category dimension.

Thank you ngalemmo.

Our invoicing System don´t record transactions at Item level. We always load the maximum grain available (thank you "The Data Warehouse Toolkit")

Another question, our BI tool recommends (God knows why) a snowflake schema (in fact modeling a star schema is more complicated in the tool).

Have any sense to you, loading the dimensions in snowfalke style, (i mean, Dim_Item only contains Item and surrogate key to Dim_Subcategory that contais only Subcategory info and surrogate key to Dim_Category...), and then create materialized views for resolving the "star dimensions". I mean, the joins with the fact always be through the materialized views (only a join) due to query rewriting.

Is that still a Star Schema? Its just an idea, i don´t know if it make sense.

Thanks again.

if you go with create two Dimensions from the d_item.. than your d_item will be role playing . I think you create a view for category and subcategory and use that in your respective fact tables.

and that would be still start schema.