Re: How to Implement Versioned Rows in PostgreSQL?
От | Alan Gutierrez |
---|---|
Тема | Re: How to Implement Versioned Rows in PostgreSQL? |
Дата | |
Msg-id | avcg70$2vb4$1@news.hub.org обсуждение исходный текст |
Ответ на | Re: How to Implement Versioned Rows in PostgreSQL? (Tom Lane <tgl@sss.pgh.pa.us>) |
Список | pgsql-general |
Tom Lane wrote: > Alan Gutierrez <ajglist@izzy.net> writes: > >>* Where can I read about alternative implemenations for history/versioning? > > > Have you looked at contrib/spi/timetravel.* ? > > regards, tom lane Thank you. The timetravel.* uses dates on each table. The nice thing about the transaction table scheme I've concocted, I think, is that I get a nice round sequential number. Dates seem so fiddly. I also get to see what changed during a transaction. I can also store who changed it with the transaction record. From looking at timetravel.*, at least I know I am not too ambitious. It seems acceptable to have a view on the tables filtering out records. I've not had experience with *large* databases yet. My big concern is that this versioning is not going to scale. That the extra query for each row will be a real dog. But, hey, the application needs versioning, gotta pay to play, right? Maybe I can hack timetravel.* to do my bidding? If anyone wants to caution (or reassure) this newbie, I would be very grateful. Thank you. Alan Gutierrez
В списке pgsql-general по дате отправления: