Single or Multiple Fact : Single or Multiple Dimension

Hi Team,

I have a unique problem while designing, i need your valuable feedback.

Definition:
The source is from internet for a pharma dataset. The source contains detail(s) about the sale information of a , product & country & time & manufacture & Supplier.

Problem:
1.The source has product , which can be of 2 types, a pharma product OR an electronic product. This is just a string in the source.
2.The attributes of both are completely different and cannot be linked.

Suggestion required :

1. Shall I have separate the source transaction table into pharma and electronic and link using relevant dimension? Even though rest of the transaction data from source has same information/granularity? OR
2. Shall I have one single product dimension combining both type of product(s), and use this to link to the same transaction table?

Any other solution?

Common practice is to model product as a sub-type cluster. You wind up with 3 tables. First is a generic product dimension with common attributes. Users that need to look at all products would use this dimension. There would then be two type specific product dimensions that contain attributes unique to the type (pharma or electronic). These would have the same surrogate primary key as the corresponding product in the common dimension. Queries for specific types would use the appropriate type dimension as well as the common dimension if needed.