Hi @nainiu440,
With respect to the filter, don’t consider True or False as Boolean conditions, they are only labels. You can achieve the same filtering behaviour with different labels. You could change “True” to “Outlier” and “False” to “In Spec” in the test measure. You would need to change the disconnected table labels to “Outlier” and “In Spec” as well but, when you choose “In Spec” from the slicer only the “In Spec” (AKA “False”)are displayed. When you choose “Outlier” only the “True” is displayed and when the slicer has no selections, both “Outlier” and “In Spec” are displayed.
I’ve given some thought to the potential solution and yes, what you’ve indicated is one solution. However, it won’t be as elegant as you would like it to be. The problem is the change in the context created by Allselected() functions when the table is further filtered further from the project level. I’m thinking a pattern like this may do the trick for the [mean] and [std] measures;
Task Mean = CALCULATE([Mean],
Filter(ALL(‘Master_Table’),
Master_Table[TaskName] = Max(Master_Table[TaskName]))
)
This will fix the task mean and std deviation to the tasks selected and should not change when you filter down to individual projects and tasks.
Attached is a Power BI model I created from your data as an example of how it could work. You will notice I’ve changed the structure a bit… I strongly suggest the use of Power Query to shape your data into Lookup and data tables, it improves performance and compression. Let me know what you think.
Project Model.pbix (224.9 KB)