In the case of DAX variables (VAR) in Power BI Measures, the following best practices can be followed to resolve the issues:
Know the Scope of Variables: DAX variables are assigned a value only once, and at that time, the variables are defined, and therefore, the value is retained within their scope forever. If you are calling a variable and expecting it to change with every row, like in the case of SUMX or FILTER, then the results are more likely to be wrong. In such illustrating cases, try to think about the logic again and look at how that variable is positioned.
Use Variables for Complex Calculations: In a bivariate or multivariate measure, reduce the calculations to a few named variables. This not only reduces the bulk of the measure but can also help point out where the trouble is. For example, suppose that your measure draws many calculations. Give each calculation its own variable. Each variable should be tested separately, giving expected values.
Avoid Over-Nesting: Excessive nesting with VAR can lead to difficulties in debugging and performance issues. It is advisable to maintain the variables at the top level and refrain from creating nested dependencies because the complexity increases with every additional nested VAR.
Debugging Women's Measures: Debug Hybrid formulas - Use Proxy Measures. Remove proxy measures if the results are as expected. Other variables will be held constant, and effects concerning the specific use variable will be measured. In situations where such a variable is not considered - expression will not be evaluated.
Row and Filter context: Variable context is understood very well in DAX. Remember to apply the right context for that calculation since results may vary with respect to row and filter context. To help eliminate which filter is a problem, try to take it out or put it in a section to see how each variable acts when on its own.
In this manner, controlling and debugging DAX variables becomes easier, resulting in more precise measures.
