Re: Normalization and Derived Information
From: Kenneth Downs <firstinit.lastname_at_lastnameplusfam.net>
Date: Sun, 10 Oct 2004 10:41:45 -0400
Message-ID: <8nhbkc.5d7.ln_at_mercury.downsfam.net>
>
> There are some excellent reasons for storing calculated values; especially
> in accounting and other money handling applications. In fact, one might
> even go so far as to state it is advisable to maintain duplicate data in
> order to effectuate financial error control.
>
Date: Sun, 10 Oct 2004 10:41:45 -0400
Message-ID: <8nhbkc.5d7.ln_at_mercury.downsfam.net>
Bill H wrote:
> Kenneth:
>
> "Kenneth Downs" <firstinit.lastname_at_lastnameplusfam.net> wrote in message
> news:mtf9kc.so1.ln_at_mercury.downsfam.net...
>
> [snipped]
>
>> Although I still believe that pre-stored calculated values are faster >> than calc-at-read, I'm going to drop that line for now and try to pursue >> your suggestion of the history table, I want to make sure I fully >> understand what you are suggesting.
>
> There are some excellent reasons for storing calculated values; especially
> in accounting and other money handling applications. In fact, one might
> even go so far as to state it is advisable to maintain duplicate data in
> order to effectuate financial error control.
>
Yes, my final fallback is the audit trail. Seeing each column in a chain of calculations gives the auditors the warm-fuzzies, they hate it when you tell them, "Oh, the gross profit calc is OK, trust me." If I can make the customer happy and guarantee correctness, I'll do it and leave the quetion of theory to another day. In this particular response to Mr. Celko however I am pretending that I never want to materialize anything, and trying to pursue that through to its final endpoint so that I can discover anything I've missed by materializing all these years.
-- Kenneth Downs Use first initial plus last name at last name plus literal "fam.net" to email meReceived on Sun Oct 10 2004 - 16:41:45 CEST