Re: What predicates the following relation represents

From: --CELKO-- <joe.celko_at_northface.edu>
Date: 5 Apr 2004 16:54:18 -0700
Message-ID: <a264e7ea.0404051554.61d96d95_at_posting.google.com>


>> Suppose I have a relation containing the tuple ('nuts', 25). If I
then do an INSERT of the tuple ('nuts', 25) in SQL I get an key violation error (assuming the primary key is defined). In a truly relational DBMS, should it silently accept this, but just do nothing? I'm just repeating a true statement more than once, so it's not really an error. <<

Sounds like we are getting back to the "cat food problem"; it was in an article in DBMS or DATABASE PROGRAMMING & DESIGN years ago. Dave Beech or someone in X3H2 wrote a defense of duplicate rows in reply to a piece by Dr. Codd on the flaws in SQL. I will look up the reference, if anyone cares. The idea was that a cash register slip would show one line for each can of cat food scanned. Since they are not differiated, they can be aggregated into one row with a quantity on the databse side.

The best response in that chain of posting was a rat litter. Ratlets are born in a lab and are tracked as a litter. As they get older, they are distinguished by colored rings made with permantent markers on their tails. Experiments can then be made on litter mates, so you have a genetically similar population.

When did they become individuals instead of units within a collective?  Think of the Borg character "Seven of Nine" on StarTrek.

While I agree that a repetiton is not an error, it is a redundancy and I'd like to be told with a "Yeah, yeah; I know that!" warning. Received on Tue Apr 06 2004 - 01:54:18 CEST

Original text of this message