Обсуждение: fake vs real CREATE OR REPLACE VIEW
Dear Dave, + create "view_a" querying "view_b", + alter "view_b", = "view_a "is broken because "view_b" changed OIDs. This to say we really need CREATE OR REPLACE VIEW in PostgreSQL. Do you remember the email of the developper who proposed to add this feature to PostgreSQL? Could you contact him off the list to ask if this is still on his agenda? Regards, Jean-Michel
> -----Original Message----- > From: Jean-Michel POURE [mailto:jm.poure@freesurf.fr] > Sent: 26 February 2002 08:09 > To: pgadmin-hackers@postgresql.org > Subject: [pgadmin-hackers] fake vs real CREATE OR REPLACE VIEW > > > Dear Dave, > > + create "view_a" querying "view_b", > + alter "view_b", > = "view_a "is broken because "view_b" changed OIDs. Yes. I thought I suggested adding a warning to that effect at one time. > This to say we really need CREATE OR REPLACE VIEW in > PostgreSQL. Do you > remember the email of the developper who proposed to add this > feature to > PostgreSQL? Could you contact him off the list to ask if this > is still on his > agenda? It was Gavin Sherry I think, but I'm not going to hassle him about it, he specifically said he wasn't going to do it immediately. How's the ALTER VIEW patch coming along btw? Regards, Dave.
Le Mardi 26 Février 2002 09:40, Dave Page a écrit : > Yes. I thought I suggested adding a warning to that effect at one time. You are right. We could query using existing views with ILIKE %view_bar% and issue a warning if some view matches, but this wron't fix the broblem. We need CREATE OR REPLACE VIEW... CREATE OR REPLACE TRIGGER is not really needed... > How's the ALTER VIEW patch coming along btw? Do you mean issuing a warning? Cheers, Jean-Michel
> -----Original Message----- > From: Jean-Michel POURE [mailto:jm.poure@freesurf.fr] > Sent: 26 February 2002 09:28 > To: pgadmin-hackers@postgresql.org > Subject: Re: [pgadmin-hackers] fake vs real CREATE OR REPLACE VIEW > > > Le Mardi 26 Février 2002 09:40, Dave Page a écrit : > > Yes. I thought I suggested adding a warning to that effect at one > > time. > You are right. We could query using existing views with ILIKE > %view_bar% and > issue a warning if some view matches, but this wron't fix the > broblem. We > need CREATE OR REPLACE VIEW... CREATE OR REPLACE TRIGGER is > not really > needed... > > > How's the ALTER VIEW patch coming along btw? > Do you mean issuing a warning? No, you were looking at a list of issues I had. I haven't had a chance to check the code yet - you may have already committed it. /D