Data Structure Question

[cbruhn42]

Below I have outlined my series of steps in a process where we track non-conformities or issues with the product at each step.  The non-conformities are logged in a single table and that table has an attribute to assign it to Step 1, 2 or 3.  There is a separate table for Step 4 non-conformities as well.

When I try to link the non-conformity table to step1, 2 and 3 I get an error saying that can't be done because I'm introducing ambiguity between the process step tables.

Do I need to duplicate my non-conformity table and then filter each new table down by the step?  So I'd end up with a non-conformity table for each step in the process.  Or, is there a way to link all these tables together that I just can't figure out.

Thank you in advance.

[cbruhn42]

I'm struggling with my design here because at each step the output has a unique lot number.  Then the next step consumes the previous lot number into a new lot number.  We log all of the lot numbers in a single field for non-conformities, but the table that contains the data for steps 1, 2 and 3 has 3 different lot number columns one for each step.

[Mikelytics]

Hi @cbruhn42 ,

When I understand it correctly the error is produced because you have two issues (see picture red and blue) since you have two active ways (active relationships) which lead to the same table.

A solution could be to make for each conflice (red and blue) one of the relationship inactive and then to work in Measures with the function USERELATIONSHIP().

But in general I think the source of problem is a lack of applying str schema logic in the data model. This is why I recommend to think about the overall design and to follow star schema best practices:

Understand star schema and the importance for Power BI - Power BI | Microsoft Learn

Best regards

Michael

-----------------------------------------------------

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly. Appreciate your thumbs up!

@ me in replies or I'll lose your thread.