If I'm not mistaken, the best practice, insofar as Pivot Tables goes, for implementing logical OR conditions and other advanced nested filtering is by adding a column (that performs the true/false test) to the data that feeds the pivot and then using that column as the filter in the pivot. However, what about when the Pivot is a live connection (like "Analyze in Excel") to a cube or Power BI dataset/model? In that scenario a user cannot just add a column to the underlying data, correct? So how would one go about using OR logic and advanced/nested filtering in such a case?
In this case yes, you can't create calculated column/field on client side. But you may create MDX Calculated Measure or MDX Calculated Fields, that is under OLAP tools. On my file connected to Power BI dataset sample calculated measure looks like