Re: views and rules
От | Chad Thompson |
---|---|
Тема | Re: views and rules |
Дата | |
Msg-id | 021401c22904$5949ec20$32021aac@chad обсуждение исходный текст |
Ответ на | views and rules ("Chad Thompson" <chad@weblinkservices.com>) |
Список | pgsql-novice |
Perfect! Thanks Chad ----- Original Message ----- From: "Masaru Sugawara" <rk73@sea.plala.or.jp> To: "Chad Thompson" <chad@weblinkservices.com> Cc: <pgsql-novice@postgresql.org> Sent: Thursday, July 11, 2002 10:57 AM Subject: Re: [NOVICE] views and rules > On Tue, 9 Jul 2002 12:01:56 -0600 > "Chad Thompson" <chad@weblinkservices.com> wrote: > > > --this is what he submits to me > > insert into call_results (phonenum, callnum, calldate, start_time) > > values ('5552552555', 4524, '6/24/2002', '17:00:32') > > > > To get around this i have tried to create the following. > > > > CREATE TABLE "call_results_fixed" ( > > "id" int8 DEFAULT nextval('"call_results_fixed_id_seq"'::text) NOT NULL, > > "callnum" int4, > > "calldate" timestamp, > > "phonenum" varchar(15), > > "start_time" timestamp > > CONSTRAINT "call_results_fixed_pkey" PRIMARY KEY ("id") > > ) WITH OIDS; > > > > CREATE VIEW "call_results" AS SELECT * from call_results_fixed > > > It appears that the types of the columns defined as a view are different > from those of the values inserted into it. I think it is probably a useful way > to explicitly cast like the following. > > CREATE VIEW call_results AS > SELECT phonenum, callnum, calldate, start_time::text > FROM call_results_fixed; > > > > > > CREATE RULE ins_call_results AS ON INSERT TO call_results > > DO INSTEAD INSERT INTO call_results_fixed (calldate, callnum, phonenum, start_time) > > VALUES ("timestamp"(new.calldate), new.callnum, new.phonenum, "timestamp"(((to_char(timestamptz(new.calldate), 'MM-DD-YYYY'::text) || ' '::text) || text(new.start_time)))); > > > > This would effectivly concatinate the date and time and insert them into the call_results_fixed table. (or so it would seem) > > > > I get an error when i set this up, however that says > > > > "Bad timestamp external representation '17:00:32' > > > > In other words, it seems that it is testing the column type without taking into account the rule. > > > > > > Regards, > Masaru Sugawara > > >
В списке pgsql-novice по дате отправления: