This is in preparation of the spatial clustering of search results which
is more comprehensible if the dataset bounding boxes are fully contained within
the query bounding box instead of just intersecting it.