Re: 4 the FAQ: Are Commercial DBMS Truly Relational?

From: Paul <paul_at_test.com>
Date: Sun, 10 Oct 2004 11:10:54 +0100
Message-ID: <41690aad$0$59441$ed2619ec_at_ptn-nntp-reader03.plus.net>


Christopher Browne wrote:
> No, the problem is much simpler than that; it doesn't require a join
> to observe it.
>
> If you cut columns off of the result set, it is possible for the
> result set to, in fact, not be a "set", but rather a non-unique "bag"
> of tuples.
>
> Thus, if I have a transaction table, I might write the query:
>
> select customer_id from transactions where txn_date between
> 'one date' and 'another date';
>
> That is NOT going to be a "set" or a "relation" if some customer made
> multiple purchases between those dates.

I think it *would* be a set - in a truly relational system, the SELECT would automatically do de-duplication. i.e. every SELECT would be the equivalent of existing SQL's SELECT DISTINCT.

Paul. Received on Sun Oct 10 2004 - 12:10:54 CEST

Original text of this message