Re: Dreaming About Redesigning SQL

Originally posted by Cmurthi

> andrewst wrote:

> > I may not have the syntax quite right, I never actually do this,
> but it

> > is right in principle

>

> Do you not do this because you disagree with the principle?

>

That's right. I am not attracted to nested tables, as Oracle calls this feature. They increase query complexity but add no value IMO.

>

> >>1) List Persons Name with PhoneNumber "12345"

> >>and

> >>2) List Persons "Mike" PhoneNumber

> > i.e. syntactically different:

> >

> > List <table> <attribute1> with <attribute2> <value>

> > and

> > List <table> <Value> <attribute2>

> > In SQL:

> >

> > select person_name from person_phones where
> phone_number='12345';

> > and

> > select phone_number from person_phones where
> person_name=Mike';

> i.e. in both cases:

> >

> > select <attribute1> from <table> where <attribute2> =
> <vaue>;

> But you could just as easily say in Pick:

>

> List PhoneNumbers Name with Name "Mike"

>

> which makes it the same syntax as 1)

>

Can you? What even if the phone number data is embedded as an MV attribute within the Persons file/table? Or do you mean you can do that if you also have PhoneNuumbers file/table with an MV attribute for people? In which case we are back to redundancy issues - explicitly storing the same data in 2 different ways.

--
Posted via http://dbforums.com