Re: Extending my question. Was: The relational model and relational algebra - why did SQL become the industry standard?

From: Mikito Harakiri <mikharakiri_at_ywho.com>
Date: Tue, 18 Feb 2003 09:50:50 -0800
Message-ID: <otu4a.13$h_3.136@news.oracle.com>

"Bob Badour" <bbadour_at_golden.net> wrote in message news:KZh4a.9$k32.1286180_at_mantis.golden.net...
> "Mikito Harakiri" <mikharakiri_at_ywho.com> wrote in message
> news:ckb4a.14$7H6.123_at_news.oracle.com...
> I disagree. What it says actually depends on the full predicate--internal
> and external--and not on whether Item represents a set or a multiset. To
my
> anglophone human eyes, it looks like it says Milk is an Item.

My typo, again. In multiset model we have

Item

---

Milk
Milk
Soda (no pun intended:-)

whereas, in set model

Item Count
---- -----
Milk 2
Soda 1

Now, the first entry in the multiset model tells that there is at least one Milk container, while the first entry in the set model tells us that there are exactly 2 containers. The set model has more precision, here, granted.

> Are you saying that a set-based relational dbms with full support for
> domains would not allow the user to freely transform data among
> representations?

Please, write down a query that transforms

Num

---
2
5

into

Seq# Num
---- ---
1    2
2    2
1    5
2    5
3    5
4    5
5    5

in pure relational syntax. Now, the same, please, without explode/recursion
operator.



> The distinction [between physical and logical] is an important one.

> You do yourself a disservice by losing sight of it.


There are limitations of every model. The question always is if the goals
that one wishes to achieve warrant changing the model.

Received on Tue Feb 18 2003 - 11:50:50 CST