PostgresSQL drop table -> drop serial
От | Jason Hihn |
---|---|
Тема | PostgresSQL drop table -> drop serial |
Дата | |
Msg-id | Pine.SGI.4.21L.01.0008141454350.2650817-100000@irix1.gl.umbc.edu обсуждение исходный текст |
Ответы |
Re: PostgresSQL drop table -> drop serial
|
Список | pgsql-general |
/* please forgive me if you've seen this before. I've tried sending this several times, but I have not seen it show up on the list yet. I was having some subscribing trouble */ Hello. I was reading through the postgre docs, and saw that it was listed to 'fix' the non-autodrop of serials. If that is correct, then I'd like to pursuade you to not do that, or at least be able to disable it. See, I have a database schema such that 2 (or more) tables use the same serial. I won't get into what I was smoking when I did that, but it extremely useful. I am concerned that if auto-dropping was enabled, then if I ever had to drop a table for whatever reason, that I would lose my serial in another table. And even if both tables were dropped, I'd still like to have that serial there incase those tables are brought back. I know it is the complaint of many MySQL users that serials should be dropped automatically, but hey, these people are using mysql in the first place which calls the value of their opinions into question =) Any real DB person would not mind such a 'quirk' of a DBMS - and I am exploiting that feature (I don't know which is worse =) . So, if it is to be implemented please leave it as an easily configurable option. And please forgive me if this opion has already been rasied. (New to the list) Thank you for your time and wonderful DBMS. PS. The serial is being used between one table and another table which is a weak entity set of the 'master.' When there is a tuple having a weak entity id == master id, then that record is considered a special tuple, and is handled accordingly. I still won't say what I was smoking when I came up with that though =). (I realize this limits the number of weak entities+ masters to 2^31, which is fine for me.)
В списке pgsql-general по дате отправления: