Re: MySQLs' "FOREIGN KEYS¨
От | Rajesh Kumar Mallah |
---|---|
Тема | Re: MySQLs' "FOREIGN KEYS¨ |
Дата | |
Msg-id | 200311101946.39867.mallah@trade-india.com обсуждение исходный текст |
Ответ на | Re: MySQLs' "FOREIGN KEYS¨ (Jeff <threshar@torgo.978.org>) |
Ответы |
Re: MySQLs' "FOREIGN KEYS¨
|
Список | pgsql-advocacy |
On Monday 10 Nov 2003 6:34 pm, Jeff wrote: > On Mon, 10 Nov 2003 08:42:13 +0530 > > Rajesh Kumar Mallah <mallah@trade-india.com> wrote: > > Oops! really sorry , the tables were not of InnoDB type > > that could have supported FKEYS. *Will post again* once > > i get to put FKEYS properly on InnoDB tables. > > > > Still in that case also the beast should tell loudly about > > whats going on behind. > > Wow. You know, I'd really expect it to at least go > "Hey, you're using MyISAM tables, which don't actually do foriegn keys.. > so.. heh. Hope you don't actually need them!" > > instead of being led to believe (like what happened to you) that you > actually had FK's and started building an app that relies on that fact. The Mandrake RPMS' of mysql 4-0-14 does not create InnoDB tables by default. In fact it does not even provide a /etc/my.cnf to tweak around. I am not sure if their RPMS' (in mysqls' website) have default settings of table handler to be InnoDB.In case default table handler is MyISAM then its REALLY bad thing to hide and remain seilent abt. Ok one thing that i think is good and MySQL enforces is INDEXES on both the keys otherwise it gives ERROR 1005. I think there was a old thread on "Indexes on Foreign keys". Now regarding the original problem, MySQL does not have that defect of treating 0 as NULL , the confusion was becoz the tables were MyISAM that did not support fkeys. Again the error message that is given on FKEY violation is not as good as ours as in 7.4 , it does not indicate which key is violated where ours does ! Regards Mallah.
В списке pgsql-advocacy по дате отправления: