The dataset that has the same name as the pool is special, which is why best practice is to not put anything directly in it, but in child datasets.
As far as I know there is no zfs command to directly migrate data from this top level dataset into a child.
Do you have any other child datasets (e.g. MyPool/MoreData )? If not, then you should be able to zfs send and zfs recv into a dataset created by the recv under the MyPool dataset.