Re: Idempotence and "Replication Insensitivity" are equivalent ?
Date: 20 Sep 2006 01:54:13 -0700
Message-ID: <1158742453.894062.301640_at_m73g2000cwd.googlegroups.com>
pamelafluente_at_libero.it ha scritto:
>
> computational complexity / market request
Ah, and ...
and clearly the above is unnecessarily restrictive. :))
ALL decent tools allow the user to define its own CUSTOM aggregate functions.
If one user has a tiny db of a few records and he want to do some complicate computation, even quadratic or factorial, it's totally up to him. Let him discover how long it takes!!
The last time I coded something similar, I just gave the user to call
its own code
on a group of records. If I didn't, I would be out of the market.
To me an aggregative function is this:
"Any function that takes n (>0) objects and returns 1 object."
where "object" is to be intended as any type including "System.Object" (datetime, double, single, string, integer, etc.)
In general we could even allow the return to be a "collection" (which could be further processed), even though this seems not well represented by the term "aggregate".
ok I said that :) ready for flames :))
-P Received on Wed Sep 20 2006 - 10:54:13 CEST