Jon,
> Which could be replaced by a single table
>
> Table UPDATE_HISTORY
> REFERRED_OBJECT OID,
> LAST_UPDATED_BY OID,
> LAST_UPDATED_DATE Date
>
> which applies to all tables. You could also have a notes table for
> everything -
>
> table NOTES
> REFERRED_OBJECT OID,
> Note Text
Actually, I'm doing this with one of my applications. Rather than using
the (problematic) OID, I simply established an independant sequence
('universal_sq') and used that as the primary key for all of my
important data tables. So long as the total records in these tables
stays < 2.4 billion, I'm doing fine. Nothing in PostgreSQL prevents you
from using a single independent sequence as the key for multiple tables.
If you are concerned about having > 2.4 billion recs, then perhaps it's
time to hack an INT8 sequence functionality. I think that adding INT8
sequences to the PostgreSQL database would be a *lot* easier than
modifying OID functionality. In fact, if it matters to you, why not pay
for it to get done?
-Josh Berkus
P.S. A lot of these concerns affect only developers with high-traffic
web applications and similar. For example, in my small business
software, it will take <> 1 million days to exhaust the TXN register.
Not something I need to worry about.
Is there some good way that we can "vote with our pocketbooks" for
various development issues in the PostgreSQL to-do list, short of hiring
a C programmer ourselves? I, for one, am desperately eager for real
stored procedures, and could get my clients to contribute toward the
development, but not more than 4 figures ...
______AGLIO DATABASE SOLUTIONS___________________________ Josh Berkus Complete
informationtechnology josh@agliodbs.com and data management solutions (415) 565-7293 for law firms, small
businesses fax 621-2533 and non-profit organizations. San Francisco
