Thread: OID
Could someone explain what this is and what it's used for? Going through the iDocs, I see that it's short for Object Identifier. But I never saw a clear explanation of what it is/used for. thnx, Chris
On Mon, 20 Jan 2003 08:28:22 -0600, "Chris Boget" <chris@wild.net> wrote: >Could someone explain what this is A database wide 32 bit identifier, automatically assigned to each new tuple, unless you created the table WITHOUT OIDS or (as of 7.3) you created the table by CREATE TABLE AS ... (see recent discussion on -hackers: "default to WITHOUT OIDS"). OID is not guaranteed to be unique, because it can roll over, and its use as a PK for user tables is discouraged. > and what it's used for? It's heavily used for identifying internal objects. For user tables it is only overhead, unless you make explicit use of it. Servus Manfred
> > and what it's used for? > It's heavily used for identifying internal objects. For user tables > it is only overhead, unless you make explicit use of it. How do you make use of it? What benefits does using OID provide? Chris
On Mon, 20 Jan 2003 09:53:38 -0600, "Chris Boget" <chris@wild.net> wrote: >How do you make use of it? My opinon: Don't use it. >What benefits does using OID provide? None that I know of. Servus Manfred
Manfred Koizar <mkoi-pg@aon.at> writes: > On Mon, 20 Jan 2003 09:53:38 -0600, "Chris Boget" <chris@wild.net> > wrote: >> How do you make use of it? > My opinon: Don't use it. >> What benefits does using OID provide? > None that I know of. See the discussion on this page: http://www.ca.postgresql.org/users-lounge/docs/7.3/postgres/datatype-oid.html which basically says "OIDs are best used only for references to system tables". regards, tom lane
Hello Chris, Am 08:28 2003-01-20 -0600 hat Chris Boget geschrieben: > >Could someone explain what this is and what it's used >for? Going through the iDocs, I see that it's short for >Object Identifier. But I never saw a clear explanation of >what it is/used for. Read: O'Reilly (Practical PostgreSQL) Page 87 ;-)) ---------------------------------------------------------------------- On table management concern can be how to distinguish between two rows whose collumn values are identical. A very useful PostgreSQL feature is that every row has its own 'object identifier' number, or OID, which is unique within that table. In other words, no two rows within the same table should ever have the same OID. This means that even if a table were designed in such a way that two rows might be identical, there is still a programmatic way to discern between them: through the OID. ---------------------------------------------------------------------- I hope this helps Michelle
Hello Tom, Am 15:52 2003-01-20 -0500 hat Tom Lane geschrieben: >See the discussion on this page: > >http://www.ca.postgresql.org/users-lounge/docs/7.3/postgres/datatype-oi d.html > >which basically says "OIDs are best used only for references to system >tables". I use the OID's on my 80 GByte Datebase very heavy... Without it I am realy f...ed !!! Michelle