Forum Discussion
Arbitrary Shaped Sets
The description we provide in the book applies to MDX queries and to other scenarios like those obtained using the context transition in a filter context managed by SUMMARIZE and/or ADDCOLUMNS.
Power BI uses a function (SUMMARIZECOLUMNS) that introduces another behavior (could it be called a bug on top of another bug?) which generates results that are less clear and predictable. We never described in detail what happened, because the important thing is that you should use KEEPFILTERS around the table function in the iterator. Or iterate over a single column.
Thank you all (Benjamin, Marco, Alexis) for your time. The goal of my original question was not to find out how to get the right result. I know that. The goal was to understand how the evaluation of that query works.
It would be great to find out more details about this behavior.
Same here, I spent 3 days tryging to figure out what is happening on the subsume level but the logic for totals does not match for subtotals.
As far as I understand (which is different from your idea) is that the filter context for 2019 matrix row is 2019, 11 and 2019, 12 therefore Values(year) should iterate only for 2019 (as this is the only year from context coming from VALUES(year), isn't it? Why do you expect to have all months there as filter context for this row is 11 and 12 only ?
Due to context transition (iteration over only 2019) - I would expect to 2019 become a new filter which should chenge the original one for this column (which is the same - 2019) -> therefore the subsume should be calculated in a correct way(2019, 11 and 2019 12).
But the result for the 2019 subsume is year = 2019, 2020 and month = 11,12
(2019/11, 2019/12, 2020/11, 2020,12)