At the expense of beating a (hopefully) dead horse, there are many
theoretical problems with multi-table update, particularly the type
where one or more tables is actually changed. Among the problems:
* Are triggers fired for update tables referenced but not changed?
* Does a record in a table referenced but not changed get an update
lock (or equivalent)?
* Does a multiple table update block on a record that won't be
changed but write locked by another transaction?
* Does a "non-update" to a referenced but not changed table block
other transactions?
* Is a record that occurs two or more times in a join updates
multiple times?
I don't think the concept of multi-table update has been thought
through. It may have seemed like an easy hack for a non-transaction
ISAM, but folded into rigorously defined transactional frameworks, it's
a mess.
It's also non-standard. I don't generally object to non-standard
extensions (I'm a repeat offender myself), but I draw the line at
extensions that are essentially unsound in conception.
--
Jim Starkey
Sent from Shearwater, off
the coast of New England
_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help : https://help.launchpad.net/ListHelp
