Re: Real world issue:- OT recreational interval

pamelafluente_at_libero.it wrote:
> kvnkrkptrck_at_gmail.com ha scritto:
>
> > If it's not fear, then what is it? Pride? Lack of intelligence?
>
>
>
> I repeat this for the last time
> ==========================
>
> (1) Consider n=2, (n is the number of operands),
> what is count distinct of { 3 , 3 } ?
>
> (2) Answer 1. Agree ? If yes proceed.
>
>
> So. Does Count Distinct satisfy this definition:
>
> (3) A binary function f ( x , y ) is called idempotent
> if for all x ,
> f ( x , x ) = x ?
>
>
> (4) count distinct of { 3 , 3 } = 1, which is different from 3.
> It doesn't.

Wuh? Count_distinct that you were discussing takes a mult_set as input. What's that got to do with the binary operation you have listed above? You certainly can't build the aggregate version of count_distinct out of it. Consider min:

e.g.: amin( {a,b,c,d} ) = bmin(a, bmin(b, bmin(c,d)))

and note the difference between the aggregate min (amin) which takes a multiset, and the binary form (bmin) which has two inputs - they are totally different beasts, and I believe you are confusing them.

>
>
> (5) So, for n=2 Count Distinct is NOT idempotent

Hence this is an incorrect conclusion, and further steps are therefore incorrect too.

>
>
> (6) Now we know that Count Distinct is "Replication Insensitive"
>
> Therefore,
> (7) "replication Insensitive" => "binary idempotent"
> does not hold.
>
>
> (8) Therefore. The notion of idempotence and "Replication
> insensitivity"
> are not the same thing, for n=2.
>
> (9) We have just seen that for n=2.
> This is sufficient to say that equivalence does not hold.
>
>
> Tell me the precise statement which is wrong
> and prove that it is wrong. Show me your intelligence.
>
>
> -P
Received on Tue Sep 19 2006 - 03:21:32 CEST