It's no big deal unless the cloned data has diverged a lot from the snapshot it came from, whereas the snapshot is taking up a lot of needed space. Unfortunately the only fix is to destroy the cloned dataset - so you need to make a copy of it (zfs send/receive) to keep it - and if you are low on space, then it can be a catch-22.
This can all be done from the UI by leveraging replication, but the shell is arguably easier and faster.