http://kaiv.wordpress.com/2008/03/02/defining-constants-in-plpgsql-with-a-twist/ don't rely on luck, count on it Fri, 23 Oct 2009 10:27:49 +0000 http://wordpress.com/ hourly 1 http://kaiv.wordpress.com/2008/03/02/defining-constants-in-plpgsql-with-a-twist/#comment-491 Anonymous Mon, 23 Jun 2008 20:08:10 +0000 http://kaiv.wordpress.com/?p=29#comment-491 If i just create the type-s i have to drop/recreate them every time i have a new classificator added or maybe i didn't get your point. If i just create the type-s i have to drop/recreate them every time i have a new classificator added or maybe i didn’t get your point.

]]> http://kaiv.wordpress.com/2008/03/02/defining-constants-in-plpgsql-with-a-twist/#comment-489 Andrew Grytsenko Mon, 23 Jun 2008 16:42:34 +0000 http://kaiv.wordpress.com/?p=29#comment-489 Why don't use CREATE TYPE compfoo AS (f1 int, f2 text); ? Why don’t use CREATE TYPE compfoo AS (f1 int, f2 text); ?

]]> http://kaiv.wordpress.com/2008/03/02/defining-constants-in-plpgsql-with-a-twist/#comment-479 Frekb Mon, 24 Mar 2008 20:44:00 +0000 http://kaiv.wordpress.com/?p=29#comment-479 omg.. good work, guy omg.. good work, guy

]]> http://kaiv.wordpress.com/2008/03/02/defining-constants-in-plpgsql-with-a-twist/#comment-473 kaiv Tue, 04 Mar 2008 09:02:20 +0000 http://kaiv.wordpress.com/?p=29#comment-473 You are absolutely correct, we had a discussion about the weakpoints of such an approach and i agree i have a very bad example here. It's not practical to use the variables like this for locating rows based on an index. First it cannot use partial indexes and even if the index is not partial it will not be able to plan as effectively as when you have the status defined as a constant. You are absolutely correct, we had a discussion about the weakpoints of such an approach and i agree i have a very bad example here. It’s not practical to use the variables like this for locating rows based on an index. First it cannot use partial indexes and even if the index is not partial it will not be able to plan as effectively as when you have the status defined as a constant.

]]> http://kaiv.wordpress.com/2008/03/02/defining-constants-in-plpgsql-with-a-twist/#comment-472 Siim Mon, 03 Mar 2008 20:31:46 +0000 http://kaiv.wordpress.com/?p=29#comment-472 What we also found was that pl/pgsql has bug using variable names in query - it cannot use partial indexes then. What we also found was that pl/pgsql has bug using variable names in query – it cannot use partial indexes then.

]]>