Re: [GENERAL] stored procedure revisited
От | Yin-So Chen |
---|---|
Тема | Re: [GENERAL] stored procedure revisited |
Дата | |
Msg-id | 380623F2.B36C102B@uswest.net обсуждение исходный текст |
Ответ на | Re: [GENERAL] stored procedure revisited (Peter Mount <peter@retep.org.uk>) |
Список | pgsql-general |
Jim Cromie wrote: [snip Virtues of SP - agreed :) And I have more opinions on the drawbacks of SP, as expected... ] > > Drawbacks of SP: > > 1) Secondary BL mechanism - Referential Integrity is generally regarded as > better. Its declarative, so is easier to use in the query optimizer. Triggers can certainly be _procedural_. > > 1) Sub-Optimal location for Business Logic > > Business Logic kinda goes with Business Applications; Apps are the context > and cause for BL, and probably the most natural place to define it, > particularly since the App tends to be more OO than RDBMSs.. Having BL in the BA level means there must be a BA for things to work. Migration path isn't critical for businesses; production databases are seldomly moved. So this is more of an issue for the application vendors than the database administrators (they want to sell the product, of course :P). > > 2 concrete contexts from Perl world. > > DBI->prepare_cached($sql-cmd): method implies that it is stored for speed. This statement only works _as_advertised_ w/ databases that have SP. Note DBD::Pg states that Postgresql does not have a prepare concept (it's there for compatibility, w/ the complete query sent every time). That's why SP rocks :) Regards, yin-so chen
В списке pgsql-general по дате отправления: