PowerPivot Relationships Many to Many

Question

The objective I am trying to achieve is to have 2 slicers in PowerPivot, ClientID and CSQName. When a ClientID is selected only the CSQnames that are related to that ClientID show up ,and vice versa

Relationship diagram link: https://goo.gl/photos/PnCZrnsXXTx3oFGh8

I am having a problem linking a many to many relationship in PowerPivot. A brief background on the application I am trying to build...

I am trying to combine a SQL database (IDM) and Informix SQL database (Cisco Call Data). The IDM database includes the Client Data and TBAS Open Case Data. Each Client has a specific ClientID. The Cisco database includes Call Detail Info and CSQNames(queue names). A many to many relationship exists, for example, a clientid can have multiple CSQname (clientid 3 has CSQ names of "A" and "B"). Also a csqname can have multiple clientids (csqname "Z" includes clientids "99", "98" and "97". Therefore I created an innerjoin table to create the many to many relationship called "Clients_CSQ".

I am trying to use this innerjoin table for both the "TBAS Open Cases" and "Call Detail". When I use this table for my filters, PowerPivot is stating that no relationships exist. Are there any solutions? If this does not make sense please let me know and I will try to specify. I have ready many posts but am unable to grasp how to make the DAX many to many relationship work with the calculate function. If someone can shed some light on the issue I am having it would be greatly appreciated. Thank you.

Answer 1

This really depends upon the data you are looking to report on. When you add two slicers to a PowerPivot table, the available selections in each slicer will be affected by the selection in the other slicer IF and ONLY IF all of the fields in the Values section of the Pivot Table are reliant on the entries in both of the slicer fields.

In your case, it is possible to make this work (as an example) by creating 3 measures:

[Call Total]=SUM('TBAS Open Cases'[Case duration])
[Number of Calls]=COUNTA('Call Detail'[appname])
[Calls by Duration]=SUMX('Clients_CSQ',DIVIDE([Call Total],[Number of Calls]))

Place the last of these 3 measures in a pivot table with the slicers set to use 'Clients_IDM'[ic_client_id] and 'CSQ Name'[csqname] and "Hey Presto!"

The first two measures are straightforward enough. The third one is cycling through each entry in the only table that these two slicer fields have in common (Clients_CSQ) and performing a calculation using the data from your FACT tables. I have no idea if the [Calls by Duration] measure that I've come up with makes any sense with your data set, but hopefully the example will help you reach the solution you want. Again depending on what data you want to show it doesn't really matter if this measure returns junk, the important thing is that it's pulling your two data sets together.

Remember that as soon as you add any raw field from either of the fact tables to this 'unifying pivot table', the inter-relationship between the slicers will break. !!!BUT!!! there is nothing to stop you from linking the csqname slicer to another pivot on the same sheet which contains fields from your Call Detail table and likewise linking the ic_client_id slicer to a pivot that contains TBAS Open Cases data. In fact, the 'unifying pivot table' could be on a different sheet from your slicers, so you only see the two sets of data that you are interested in.

And ignore that warning about no relationships existing!