On Fri, Apr 07, 2000 at 01:18:47PM -0700, Andrew Schmeder wrote:
...
> So far, I have learned that every row, table, etc in postgres can be associated
> to a globally unique object id (oid), correct? I tried to explore the system
> tables, pg_, etc but I could not find a table which actually just contained the
> oids, so I assume that this is hidden somewhere on a layer below the
> database...
Maybe this is useful for your quest for oids:
regression=# \d test
Table "test"
Attribute | Type | Modifier
-----------+-----------+----------
d | timestamp |
regression=# select * from test;
d
-----------
-infinity
infinity
(2 rows)
regression=# select oid,* from test;
oid | d
--------+-----------
148169 | -infinity
148170 | infinity
(2 rows)
So, if you use "oid" specifically, it appears.
> I want to know what the best method to implement this system will
> be... I am thinking triggers / store procedures? but where do I store them...
> on each table that may have a dependency?
>
> any advice would be appreciated...
I don't know much about these things - my guess would be triggers in the 10
tables which on {update,insert,delete} deleted your cached page, but hopefully
someone more knowledgeable will answer this bit.
Cheers,
Patrick