It might help if you tell us a little more about what you’re trying to do. There are lots of ways to operate on your data to only show a subset of it. Do you know the composition of the subsets before you plot them, or do you determine that while they are being visualized, for example?
“Ghost” cells are often used for finite-element simulations, where blocks are broken up and handled by different threads/processes. The process needs to know a little bit about boundary cells in order to properly compute an answer. It isn’t so much so that they can be shown or not shown.