Re: Lots of Idiotic Silly Braces?

From: Brian Selzer <brian_at_selzer-software.com>
Date: Mon, 23 Jul 2007 17:00:19 GMT
Message-ID: <D85pi.10322\$eY.6989_at_newssvr13.news.prodigy.net>

"paul c" <toledobythesea_at_oohay.ac> wrote in message news:1j4pi.2\$_d2.0_at_pd7urf3no...
> Brian Selzer wrote:

```>> "paul c" <toledobythesea_at_oohay.ac> wrote in message
>> news:JzMoi.137922\$1i1.132340_at_pd7urf3no...
>>
>>>Brian Selzer wrote:
>>>...
>>>
>>>>I think I see where you're going with this: correct me if I'm wrong, but
>>>>aren't you saying that P and P2 are independent?  ...
>>>
>>>Maybe I would have been more on the point to say about your
>>>
>>>r = {{S={{A=3, B=4}, {A=3, B=5}}}, {S={A=3, B=4}}}
>>>
>>>that the two tuples of r, say s1 and s2, are independent relvars.
>>>
>>
>>
>> Are you trying to say that each tuple in a relation value is independent
>> of all other tuples in the same relation?
>>
>
```

> I was trying to say that "primitive" operations on r, such as "delete" or
> "insert", don't operate on the tuples that appear in s1 and s2, they
> operate on the (relation) values of s1 and s2, which are the tuples that
> appear in r.
```>
```

> By the way, UNGROUP seems to have the effect of dispensing with the
> relation values of s1 and s2 and replacing them with the tuples that
> appear in s1 and s2. I presume this means that s1 and s2 must be of the
> same relation type, ie., same heading, otherwise UNGROUP would need to
> allow a result of more than one relation value, each value having
> different headings. Offhand, I don't know if there is a way to construct
> a relation that would have such relation values as tuples but if there is
> a way, then I think trying to UNGROUP it would be a problem!
>

That's true. It is precisely the fact that each relation value for an rva has the same heading that caused me to question whether an rva can be a key or the only key on a base relation. This is due to the fact that a relation schema with a particular heading can only have one value in a given database value. Now it is quite possible to have two different relation schemata with the same heading, thereby having two independent relation values with the same heading in a given database value, but relation schemata are named, whereas rva's are not (unless of course the rva is a dependent attribute). What I question is whether a relation value can be self-determining in the context of the other relation values (in other tuples) in the same relation. The lack of duality between UNGROUP and GROUP means that a given value for a relation schema that has an rva as the only key may determine a different relation value for the same schema in addition to itself. In other words, there can be more than one relation value for the same schema that means precisely the same thing.

> p

>
> p
Received on Mon Jul 23 2007 - 19:00:19 CEST

Original text of this message