The performance issues are part of why I would prefer to have those statements be a single entry. I had gotten in the habit of immediately flushing large bulk inserts as soon as executed.
It doesn’t happen often, but from time to time I need to sync data manually from our PROD to our QA environment, and that can be literally thousands of rows each across several tables when it happens. generating and executing the insert or update statements works well enough, but having to select and delete that many single entries from the history is a pain. And there is no reason to keep them.
To answer your earlier reply, no the filter doesn’t particularly help, since the few categories available are so broad. SQL, DML, PL/SQL and other.
Still, thanks for the info.