Mixed workflow: in-memory and Direct Discovery

[ciampaciampa]

Hello,

I would like to build a QV data model that allows me to switch from the QV in-memory mode to the direct discovery (DD) mode.

For example, I want to load (fully) in memory, in a QV classic way, a Fact Table aggregated at level of  Dim1,Dim2,Measure.

This could be the "normal use case".

In some cases, details from a Dim3 could be required then, I would like to "drill" in a DD mode because the huge volume of data.

Also, I want to show only the Dim3 values  related to the current selection on Dim1,Dim2 before the DD mode switch.

Dim1=City  >>in memory

Dim2=Shop >>in memory

[Dim3]=Product>>in memory by  DD

Measure= Qty Sold (in memory or from DB depending on the level of my navigation request)

Is it possible to manage both the fact tables ,memory+DD, and switching between them?

Does Anyone already tried this?

Hope it's clear.

Thanks a lot  in advance

[ciampaciampa]

It seems to work creating a combined key K1=Dim1+Dim2 to put in both the fact tables.

Using Syntetic Key or Link Table, it seems not working.