Re: Antw: Re: Revised proposal for QUERY DTD



Le jeu, 13 jui 2000, vous avez écrit :
> You wrote:
> > ...
> > > - when you distinguish func and aggregate, then the attribute group in func/field is needless ?
> >
> > The distinction between functions and aggregates is there to the GROUP BY. But
> > aren't some cases where the group attribute could be necessary?
> > ...
> 
> You can use GROUP BY only in conjunction with aggregat functions. 
> Then you have to group by all attributes in the source list, that are not aggregated.
> 
> So if we have both the group attribute and the distinction between functions and aggregates it's redundant. 
> But as it simplifies the generation of SQL statements, I will accept it.

Ok, then let's remove the group by attribute, no need to be redundant in the
XML file.

> 
> Another suggestion: We could put this distinction as an attribute to the func element:
> <!ATTLIST func
>                    aggregate (yes|no) "no"
> ......
> 

Yes, that seems to be a better idea, go with it.

Thanks,

Vivien




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]