Adding a measure to table affects other measures

[Lauri]

I'm noticing behavior in a straight table I haven't seen before: When I add a measure, it affects other measures in the table. Here's a screenshot of the table without the new measure:

And here it is with the new measure:

Note how the 2nd, 3rd and 4th columns have different values after the last column is added. (The table has other dimensions and measures not shown here.)

Is this expected behavior? 

I'm on QS Enterprise Feb 2024. 

[marcus_sommer]

I think it's a rather unusual behaviour but I don't want to say that's a bug because depending on the data-model (any synthetic keys or circular loops) and/or the data-associations (missing key-values or NULL - in the data-model as well as within the object) and/or the data-set (NULL and/or multiple values in fields which are queried per only() or firstsortedvalue() in regard to the selection state / object-dimensionality) and/or any calculated dimensions and/or using key-fields within aggregations - there may some unexpected side-effects.

If anything of the above hinted possibilities is there you should consider to solve it - at least temporary and step by step - in another way.  

[Lauri]

One of the other measures in the table (not shown in my screenshots) contained set analysis to ignore several dimensions. After removing that measure, the above strange behavior stopped. 

[Lauri]

Thank you, Marcus. Indeed, the NOTES field has this characteristic: "NULL and/or multiple values." The system generates a null value for NOTES on the same date as when there is a string, so there are multiple rows of data, effectively duplicates. I'll work on removing the unnecessary rows.