Re: Idempotence and "Replication Insensitivity" are equivalent ?

 <pamelafluente_at_libero.it> wrote:
>
> Chris Smith ha scritto:
>
> > For example, AVG is an aggregate function because it is defined by
> > SUM(P) / COUNT(P), where SUM and COUNT can be defined as primitive
> > aggregate functions... but there is no function g that generates AVG
> > directly.
>
> consider also these "aggregate" functions:
>
> - range of values (difference between max and min)

This is MAX - MIN, so it is an aggregate function.

> - median
> - Any quantile
> - standard deviation
> - variance

These are not definable as aggregate functions.

> - regression coefficient
> - Pearson's skewness

I don't know what these are.

-- 
Chris Smith