[bhl]

Regarding columnar databases, is there an optimal data structure in between storing data as rows and storing data as columns that would make it fast to sort and filter by either dimension? Some mix of B-trees or interval trees?