On 23/11/2009 4:15 AM, Scott Marlowe wrote: > On Sun, Nov 22, 2009 at 12:50 PM, Thom Brown <thombrown@xxxxxxxxx> wrote: >> Hi, >> This should be simple, but for some reason I'm not quite sure what the >> solution is. I want to be able to update the value of a column for rows >> that have been updated. More specifically, if a row is updated, I want it's >> modified_date column to be populated with the current time stamp. I've >> looked at triggers and rules, and it looks like I'd need to create a >> function just to achieve this which seems incredibly clumsy and unnecessary. >> Could someone enlighten me? > > Well, you DO have to create a function, but it's not all that clumsy > really. Also it's quite flexible so you can do lots of complex stuff > and hide it away in a trigger function. I do think this comes up often enough that a built-in trigger "update named column with result of expression on insert" trigger might be desirable. Especially if implemented in C to avoid the need for PL/PgSQL and to reduce the CPU cost a smidge. Hmm. CC'iing -hackers; there was a discussion earlier on it being desirable to have more "[EASY]" TODO items, and this might be a good one for the job. So might "CREATE LANGUAGE ... IF NOT EXISTS". Maybe even "CREATE ROLE ... IF NOT EXISTS" and "CREATE USER ... IF NOT EXISTS" - I know I'd find them really handy. -- Craig Ringer -- Sent via pgsql-general mailing list (pgsql-general@xxxxxxxxxxxxxx) To make changes to your subscription: http://www.postgresql.org/mailpref/pgsql-general
