Re: Lucid statement of the MV vs RM position?
Date: Mon, 08 May 2006 09:02:19 +0200
Message-ID: <e3mqdp$tkg$1_at_orkan.itea.ntnu.no>
Bob Badour wrote:
> Marshall Spight wrote:
>> Can you be more specific about the differences between the
>> two?
>
> In the end, I think you and I focused on function while Jon focused on
> form.
You don't get from relation (1) to relation (2) in my post to Marshall just by using SUM. You have to use SUMMARIZE as well.
For that matter, I don't think neither you nor Marshall really have focused on anything but saying "Yes, it is!". You don't address my arguments at all.
> Functionally, GROUP is an aggregate among other things.
(An *aggregate*? Or an aggregate *operator*? Do you mean to distinguish between those terms?)
No. Functionally (I'm not sure what precisely that means, but...) it can be defined *using* an aggregate. (But you don't have to define it that way).
Is "SUMMARIZE R BY A ADD (SUM(X) AS AGG_X)" an aggregate (operator?)? Is it the aggregate operator SUM?
Is "SUMMARIZE R BY A ADD (SUM(X) AS AGG_X, AVG(Y) AS AGG_Y)" an aggregate? Which aggregate operator is it?
Is "SUMMARIZE R BY A ADD (UNION(RELATION{TUPLE{X}}) AS AGG_X)" an aggregate?
> For instance, it is also a value selector.
I'm not sure what you mean by this term.
> Syntactically, Date and Darwen chose to handle it differently than other
> aggregates. Perhaps to avoid confusion with the UNION aggregate,
But isn't D&D's UNION aggregate precisely what you claim GROUP is? Or do you actually claim that an aggregate operator is *both* the operator itself (defined by identity and repeated operation) *and* a SUMMARIZE invocation using the aggregate operator as a summary (and presumably no other summaries, cf. my SUM/AVG example above)?
> or
> perhaps to provide a symmetric balance to UNGROUP. Perhaps for a lot of
> reasons or perhaps for no particular reason.
>
> One can easily look at GROUP as a convenient shorthand for a combined
> type conversion and aggregate.
No, you need a "surrounding" SUMMARIZE as well.
-- JonReceived on Mon May 08 2006 - 09:02:19 CEST