Re: Logical equivalence of simple and complex types under the relational model?

"Gene Wirchenko" <genew_at_mail.ocis.net> schrieb im Newsbeitrag news:3kmpq01638b6om0tirukm2eqfka6s1gjte_at_4ax.com...
> "Rene de Visser" <Rene_de_Visser_at_hotmail.com> wrote:
>
> [snip]
>
>>Simple in the sense of RM means that the value is atomic with respect to
>>RM.
>>i.e. there are no operations within RM that can access parts of the value.
>>e.g. the number 2 is atomic in RM if there are no operations that let you
>>look into the structure of a '2'.
>>
>>All other types are complex.
>
> If the DBMS has a function that returns the integer part of a
> float, this means that float is a complex type?
>
I was wondering about such things myself. What if you have a function that returns that the nth bit of an integer? Or allows you to do prime decompositions of a number?

I am not sure that the split between complex and simple is well defined at all.
Especially if the DBMS allows user defined functions that can operate on the phyiscal level...

I have further doubts that it even makes a difference.

I have been doing some programming in a relational data base language (AP5), and have been considering whether
using complex types can help me get any additional reuse. So far I haven't found any way to use complex types to any (great) advantage. There are some differences of course. You can do direct function polymorphism if you use values from different classes, but you can also just create a relation (categorization, function), which achieves the same thing.

Rene. Received on Tue Nov 30 2004 - 21:48:26 CET