With this 'defect' I am unclear without a detailed read what the correct behavior is, so I won't be surprised if the resolutions of most of the parts of this defect are "invalid". However, I believe that the missing solutions is a real defect
- it may be that they should be reported as a scope error for the non-aggregate ?label variable, but the current behavior I suspect is incorrect.
I have a complex query, the part in question boils down to:
which gives just case 1 as the answer set, where I was expecting 1, 3, 2 with 3 before 2 but position of 1 undefined.
The work-around I am using is:
which gives the answer set 1, 3, 2 (i.e. meeting my needs, because 3 comes before 2)