>Is there a way to return which record(s) is locked when an attempt to lock the table header fails?
>I know I could scan and check for record locks, but this would take too long. I guess what I want to know is, How does a process know that a record is locked without checking that individual record? Maybe I just need some insight into the inner workings of record and table locking.

I should say again that RI cascade/update triggers really SCAN and RLOCK() inside transaction. It gives some indication that this is the only reliable way in multiuser environment.