The most efficient way of achieving that particular goal is to
consider the issue when designing and normalizing your table schema.
If you're worried about concurrency and conflicts, partition your
tables so that users are not working with the same rows at the same
time. If you have a situation where users "own" certain rows, you can
create a default value to place the login ID in a row and then have
users only be able to edit the rows that they own. Another possible
solution is to create an integer column named ConcurrencyID or some
such, and increment it by 1 on every successful save. Your code can
check that the value currently in the ConcurrencyID is the same as the
value retrieved when the row was loaded, and if not, don't save the
record.
-- Mary
MCW Technologies
http://www.mcwtech.com