Re: Referential constraints in version 8
От | Tom Lane |
---|---|
Тема | Re: Referential constraints in version 8 |
Дата | |
Msg-id | 2590.1113572299@sss.pgh.pa.us обсуждение исходный текст |
Ответ на | Referential constraints in version 8 (Bob Smith <bsmith@h-e.com>) |
Список | pgsql-admin |
Bob Smith <bsmith@h-e.com> writes: > We have just finished upgrading Postgres from 7.2 to 8, and I must live > right or something because there was only one glitch. When the dump > from 7.2 was restored into 8, some foreign key references which should > have been initially deferred had become non-deferrable. I had to > re-define those references by dropping the corresponding triggers and > using ALTER TABLE to put them back as foreign key constraints, which > seems to have fixed the problem. However, those references which I > re-defined now show up explicitly in the table descriptions as foreign > key constraints, and the corresponding triggers are not listed. This > is great since it makes the table descriptions much more intelligible, > but my concern is that all the other references which I didn't > re-define still show as triggers and not as foreign key constraints. > Is this just a cosmetic issue with psql's table description, or is > there actually a functional difference? Should I re-define all the > other foreign key constraints to be safe? Yeah, you should --- if it shows as a trigger then the system hasn't fully grokked it. The contrib/adddepend script may help you. regards, tom lane
В списке pgsql-admin по дате отправления: