On Apr 26, 2005, at 12:35 PM, Tom Lane wrote:
> One possibility for an alternative mechanism is to allow triggers on
> views --- but I'm not sure exactly how this would work, or if it would
> solve all the problems. At the very least it would answer the "data
> stability" issue, since I suppose the trigger would receive precomputed
> rows that wouldn't change while it executed.
FWIW, SQLite allows triggers on views. They work very much like Pg's
rules (only I think that its INSTEAD OF syntax short-circuits the
execution of other triggers) in that it allows you to use a series of
SQL statements that will be executed--no function definition required.
http://www.sqlite.org/lang_createtrigger.html
Works great!
Cheers,
David