On Thursday 15 January 2004 13:13, pginfo wrote:
> Hi,
>
> I am using pg 7.4.1 and have created a trigger over table with 3 M rows.
> If I start masive update on this table, pg executes this trigger on
> every row and dramaticaly slows the system.
> Exists in pg any way to define the trigger execution only if I have
> changes on some fields?
Not at the moment (and I don't know of any plans for it).
> For example I am able to declare this in oracle.
> My trigger is writen in pgSQL.
Hmm - I can only think of two things you can try:
1. check for the change first thing you do and exit if not there
2. do the same, but write the trigger function in 'C'
--
Richard Huxton
Archonet Ltd