Case 2 happens to have radial symmetry around the centroid of the octagonal section. So, it is ‘cheating’ but you could do a polar array around that centroid, Not a ideal approach, can’t be used for attractors, etc, but it will work for simple cases. Good luck