I am developing a small-scale client-server database application. The server creates a handler thread for each client request and uses connection pooling.
How do I perform update transactions (single rows updated by row key)? If I start a transaction and submit a SELECT ... FOR UPDATE followed -- some time later -- by an UPDATE statement and a commit, I'll be splitting my transaction in two. Can I do this? And if these two requests arrive via different db connections will the DBMS still accept the UPDATE and unlock the row?