Slicing data on a multiple one to many relationships

[Hammerhead]

This is an extremely basic version of the model that I currently have, but the issue is still the same. My current model consists of 20 other tables and millions of records, so please bear this in mind with regards to performance etc. 

In reality, Attribute has over 8 millions records, and Product 250,000. It is not possible to combine these tables.

The simplified model is as follows:

A customer would like to have a report that displays Product, Identifier and Attribute fields in the same table visualisation, but this is obviously not possible due to "Can't determine relationships between the fields". 

I would therefore like to display Identifier.Subgroup, Product.Product in a table, but then sliced by Attribute.Locale. This currently does not work. How can this be achieved?

[johnyip]

If you need to slice Identifier table using Attribute table, you should set the cross filter direction between the two tables as both.

[Hammerhead]

Many thanks, but this causes issues with regards to other connecting tables and previously calculated measures. In addition, I would be concerned regarding both performance and potential ambiguities. I was hoping that there might be a solution that could be created with a measure and applied to Filters for the specific visual.

[johnyip]

In this case, create a measure to capture the selected value of [Locale] from Attribute.

SelectedLocaleIdentifier = CONCATENATEX(ALLSELECTED('Table 2'),[Identifier],", ")

And then, another measure to scan if the SelectedLocaleIdentifier matches with your data.

ScanMatch = IF(MAX('Identifier'[Identifier]) IN SelectedLocaleIdentifier,1,0)

And then put ScanMatch in Filters on this visual as equals to 1.

I just typed these out without testing, this should work if your visuals are a table / matrix and 'Identifier'[Identifier] is included in your visuals.