Re: GUID in postgres

"Kaiserdigital":

> Why would you want to use two columns when you use could one? Two
> columns in each table would clutter your scheme as well as your
> procedures.

It's a relational integrity issue.  A GUID, by definition, contains two
pieces of information:
1. The local primary key
2. The server unique ID
As such, RDBMS design principles (the Second Normal Form, I believe,
correct me if I'm remebering wrong) mandates that they be kept in two
columns.

This is a pet peeve of mine, as DB vendors and beginner DBA's today seem
to be in a rush to embrase "non-atomic" fields willy-nilly, abandoning
20 years of accumulated RDBMS wisdom.  


> It doesn't matter if you change the network card. The GUIDs will
> still
> be unique. The MAC does not make the GUID unique on any given server.
> The rest of the GUID generation function accomplishes this task. All
> that is accomplished by incorporating the MAC into the GUID is
> uniqueness between machines.

Makes sense if you have an open-ended network of machines so that a
simple numbering sequence won't work.  With 3 servers, "1", "2", "3"
work just as well, and don't have the 12-byte overhead of a MAC address.

-Josh


______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