overwrite - with same _id? not possible. as you process another event record with the same _id, then this record will be ignored irrespective of you pass the same timestamp or different timestamp.
overwrite - with different _id? its technically not overwrite. you are ingesting another event, and both event does exist for the profile.
delete? possible, if you drop the dataset. deleting particular batchid will only delete from the data lake not from the profile store.
if you find a way to overwrite the event record or deleting the record without dropping the dataset, let us know. would be much helpful to know. 🙂