GROUPBY looses visual filtering

[monstrator]

Hi Experts!

I try to create a IT infrastructure dashboard which shows the software packages that all servers have in common or the software packages that NOT all servers have in common. The goal is easily identify any differences between server setups:

As you can see, I have selected 4 servers which have 2 different OpenSSH versions installed.

In the top table I have aggregated the count of servers per package. 

What I try to achieve now is to filter all packages which do not occur for the same number as number of servers that I have currently selected. This means that those packages are not common for all servers.

I tried to reproduce the behaviour of the top table in a measure table with the following code:

Servers with Package = groupby(Sheet1;Sheet1[Server];Sheet1[Unique Package ID];"Server Count";countx(CURRENTGROUP();[Unique Package ID]))

But now the problem is that I am loosing the filter context from the visuals partly as I can only either link the server OR the unique package ID of the now two tables together in the model:

The result is that the list of packages I want to filter either lists all packages or the data for all servers. I need some help here!

I uploaded the PBIX and the Excel source sample data I was using to: https://monstrator-my.sharepoint.com/:u:/g/personal/denis_monstrator_se/EZOhoY1q6WZBh-U---Tx4uUBbUdZ...

BR, Denis

[vanessafvg]

@monstrator  create a composite key, on both tables, create a new column.  combine the link and package values in one field and then join on that

[monstrator]

Thanks @vanessafvg! 

Unfortunately, this does not solve it. There is already a unique package ID (composite key) . If I join on that, I get the result for all servers. If I join on the server, I get results for all packages. I want to be able to get the grouped count per server and per unique package for the filtersI have set. 

Creating another compaund key including the server name does not help either because then I am not grouping over the unique package anymore.

BR, Denis

[vanessafvg]

@monstrator  please provide an example of what you expecting

[monstrator]

@vanessafvg I mentioned in my initial post that I would like to get the same data that I have in the top table (Count of Server/Unique Package ID) which I achieved by doing a distinct count in the visual as result from my new measure table.

So in short: I want to be able to filter on the data which the visual gives me but inside a measure table 🙂

BR, Denis