Many to Many Modelling - Qlik v/s Power BI -Power BI shows correct aggregation but misses detail row

[Manguirish]

I have a many to many model in Power BI based on following sample tables

Ord Table

KeyOrdPO	Ord	PO	Product	Qty
1|1	1	1	Q1	100
1|1	1	1	Q2	500
1|2	1	2	Q1	50

Cost Table

KeyOrdPO	ChargeType	Cost
1|1	Shipping	100
1|1	Packing	200

The model here is a many-to-many bidirectional Model

When I do a table visual in power BI, I get the following result

Even if i create a Bridge Table with KeyOrdPO, and model this the resultant visual is exactly same

However, when the exact same scenario is replicated in QlikSense, the model looks like this

The Qlik engine by default creates a bi-directional many to many model on common keycoloumn which is indexed in the memory - the so called tenet of Associative Model of Qlik

The resultant visual looks like this which clearly shows the allocation of Qty-50 belongs to Q1 and doesnt have any corrosponding ChargeType. However, in power BI this row is missing in the visual even though the aggregation is correct. This presents a confusing scenario

So Question - How do i get this right in power BI.

[MarkLaf]

You can try to add in rows to Cost for any missing KeyOrdPO from Ord. Here is a Power Query solution, where Cost_Original and Ord are the table you start with in your OP:

let
    Source = Cost_Original,
    Missing = List.Difference( List.Distinct( Ord[KeyOrdPO] ), List.Distinct( Source[KeyOrdPO] ) ),
    MissingToRows = List.Transform( Missing, each { _, "-", null } ),
    MissingToTable = Table.FromRows( MissingToRows, Value.Type( Source ) ),
    Cost_Combined = Table.Combine( { Source, MissingToTable } )
in
    Cost_Combined

This gives you:

Cost

KeyOrdPO	ChargeType	Cost
1|1	Shipping	100
1|1	Packing	200
1|2	-	null

If you load this as your Cost table, I believe your original M2M model (no bridge) should work as expected:

[lbendlin]

You are absolutely correct - the associative model is the USP (UNIQUE selling point) of Qlik.  Power BI on the other hand is based on the Star Schema mentality.

If you need many to many relationships, stay with Qlik.

In Power BI, enable "show items with no data" and read about "Auto Exist".