Re: Transactions and insertion ordering
От | Martijn van Oosterhout |
---|---|
Тема | Re: Transactions and insertion ordering |
Дата | |
Msg-id | 20040610101549.GA10766@svana.org обсуждение исходный текст |
Ответ на | Transactions and insertion ordering (James Pharaoh <james@pharaohsystems.com>) |
Ответы |
Re: Transactions and insertion ordering
|
Список | pgsql-general |
nextval() should return value in the order they were called, rather than commit time. I hope you're not relying on the unordered results of a table scan remaining stable. Tables have no intrinsic "order", only one inposed by an external sequence. Hope this helps, On Thu, Jun 10, 2004 at 09:28:50AM +0100, James Pharaoh wrote: > Hi, > > I'm trying to work out how to make sure things are read from a table in > a consistent order. The table represents a queue of items and also the > history of those items. > > Even with "serializable" transaction isolation I can begin two > transactions, insert a record in each, commit the second transaction > first. This second record is now visible and can be read from the queue. > But when I commit the first this one appears before the second one. This > could then be read from the queue second but when I rescan the table to > view history it looks like it was read first. > > Are there any ways to make this work a little more intuitively? > Basically I guess I want to be able to model a queue effectively. > > James > > > ---------------------------(end of broadcast)--------------------------- > TIP 4: Don't 'kill -9' the postmaster -- Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/ > Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a > tool for doing 5% of the work and then sitting around waiting for someone > else to do the other 95% so you can sue them.
Вложения
В списке pgsql-general по дате отправления: