I think the data values make difference here. Even I don't understand this function much but I found few data cases work few not. For example, I created this sample data which shows some values.
You may like to view this video to understand further on sample data preparation -
chi2-test.qvf 189.5 K
I agree with you about the data values quality.
After "cleaning" data, the function accept to calculate something.
"Cleaning" data means :
- No null or 0 values
- Respect of my couple dimensions. For each value for Dim1, I need 2 Dim2.
This condition means :
Dim1 Dim2 Val A 1 34 A 2 74 B 1 12 B 2 65 C 1 23 C 2 41 ... ... ...
If I don't have the two lines for a Dim1 value, the calculation fails.
I don't really understand too and don't really know if the result is finally right