The pivot tables (Cubes) have come a long way and if I understand correctly, they are still under development.
I have encountered some scenarios where I had the need for a more dynamic grid. In some scenarios we used views as a complicated work around, where the headers and amount of columns are fixed, and in the first row we create our own headers and we pivot the data in the query. Of course a sub optimal solution.
In other scenarios we used the grid + grouping as an alternative, so data is shown close to each other
As a practical example:
In a proof of concept, a customer wants to compare the texts of different contracts per subject and per supplier.
The ideal situation would be to have the suppliers on one axis and the subject on the other
the specific text of each subject / supplier would be visible in a cell.
In a pivot table in Excel and in Power BI it is perfectly possible to use a string for the data. It simple demonstrates the text or counts the amount of entries.
I noticed that the cube of Thinkwise can only handle numeric fields.
If the cube could handle strings as well this could be very valuable to create matrixes like the one I mentioned.
Do you think this is technically viable? Or do you see other possibilities that I'm missing?



