Dimensional keys in both parent and child fact tables

I have some measures which cannot be broken down to the child level so I have created two fact tables.

My question is should the parent dimensional keys be stored in the child fact table as well as the parent fact table? Or would a join between the parent and child fact tables suffice?

For example, my parent-child relationship is theatre sessions and operations.

A session has dimensions such as session type, session scheduling type, session specialty, session consultant. SessionKey is the degenerate dimension.

An operation has dimensions such as operation type, operation consultant, operation specialty, operation anaesthetic type. OperationKey and SessionKey are degenerate dimensions.

Which of the following should the child fact table look like?

Child Only Dimensions	Child & Parent Dimensions
OperationKey	OperationKey
SessionKey	SessionKey
OperationType	OperationType
OperationConsultant	OperationConsultant
OperationSpecialty	OperationSpecialty
OperationAnaestheticType	OperationAnaestheticType
Measures	SessionType
	SessionScheduledType
	SessionSpecialty
	SessionConsultant
	Measures

The parent fact table as well as having it's own measures would have also have aggregated child measures.

The concept of 'parent' and 'child' facts does not exist in a dimensional model. Each star schema (fact table) is a stand-alone entity. Any fact table contains all dimensions that properly represent the grain and context of the measures. The operation fact should have the sessions' dimensions as well.

ngalemmo wrote:The operation fact should have the sessions' dimensions as well.

That's what I thought. Thanks for your response.