Re: FW: Surrogate keys (Was: enums) - Mailing list pgsql-hackers

From Josh Berkus
Subject Re: FW: Surrogate keys (Was: enums)
Date
Msg-id 200601181913.11349.josh@agliodbs.com
Whole thread Raw
In response to Re: FW: Surrogate keys (Was: enums)  ("Dann Corbit" <DCorbit@connx.com>)
List pgsql-hackers
Dann,

> The primary key should be immutable, meaning that its value should not be
> changed during the course of normal operations of the database.

Why?   I don't find this statement to be self-evident.   Why would we have ON 
UPDATE CASCADE if keys didn't change sometimes?

> At any rate, the use of natural keys is a mistake made by people who have
> never had to deal with very large database systems.

Oh, I guess I'm dumb then.  The biggest database system I ever had to deal 
with was merely 5 TB ...

Anyway, my opinion on this, in detail, will be on the ITToolBox blog.  You can 
argue with me there.

-- 
Josh Berkus
Aglio Database Solutions
San Francisco


pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: pgxs/windows
Next
From: uwcssa
Date:
Subject: suppress output for benchmarking