Re: [INTERFACES] large objects
От | Peter T Mount |
---|---|
Тема | Re: [INTERFACES] large objects |
Дата | |
Msg-id | Pine.LNX.3.96.980727142212.934K-100000@maidast.retep.org.uk обсуждение исходный текст |
Ответ на | Re: [INTERFACES] large objects (Aleksey Demakov <avd@gcom.ru>) |
Список | pgsql-interfaces |
On 27 Jul 1998, Aleksey Demakov wrote: > Peter T Mount <peter@retep.org.uk> writes: > > > > > The object is stored using a table/index pair, and takes up more disk > > space than a single object. > > > > How much more? And what about performance? > > In fact I don't need very large objects. 64k or even 32k will suffice. > Could such not-so-large objects be implemented more effectively? Anything larger than the page size (default is 8k, but can be larger) are a large object. The internals do alow for multiple storage managers, so it is possible to use other methods than the current "Inversion" scheme. We did have a posibility of a different manager, but I've heared nothing of it for some time now (pre 6.3 if memory serves)/ > It would be very nice to have a new postgres type for this. The 8k limit > is very restrictive. What do postgres gurus think? Part of the solution for the BLOB orphaning problem that JDBC & ODBC have requires a unique type 'lo' (required to differenciate between a generic oid, and a large object while vacuuming). However, for most cases, the existing lo interface is the only realistic way of handling them. -- Peter T Mount peter@retep.org.uk or petermount@earthling.net Main Homepage: http://www.retep.org.uk ************ Someday I may rebuild this signature completely ;-) ************ Work Homepage: http://www.maidstone.gov.uk Work EMail: peter@maidstone.gov.uk
В списке pgsql-interfaces по дате отправления: