Re: Idempotence and "Replication Insensitivity" are equivalent ?

<pamelafluente_at_libero.it> wrote:
> What I think I have heard let's not consider COUNTDISTINCT an aggregate
> function just because it does not fit in our fashinating math
> contruction
> would make laugh any dbms user ...

You *really* need to finally decide whether you care about database theory or not. If so, then it certainly matters how COUNT DISTINCT is modeled. If not, then why do you care whether COUNT DISTINCT is modeled as an aggregate function, or as the composition of a relational projection and an aggregate function?

Obviously, no one here is saying that databases shouldn't implement COUNT DISTINCT, or any other feature you like. But you know that. You just find it helpful to try to be persuasive by misrepresenting people. I am now coming to realize that. Sorry it took me so long; I'm relatively new here. Now I see why everyone else is being so rude to you.

There is, of course, a huge benefit to reusing a uniform way of addressing the hard problems of relational transformations. It is quote counterproductive to blindly deciding that COUNT DISTINCT doesn't involve a projection, and thus lose sight of the fact that it carries around exactly the same characteristics, challenges, etc. as a projection, *composed* with an aggregate.

-- 
Chris Smith