Re: Client/Server query strategy

>Codd & Date and all derivatives and antecedents STRONGLY
>argue against keeping derived data in the database. The argument is
twofold:

> 1) The derived data is redundant. That is, if you have data
> elements A, B and C, you can always calculate any
> function that depends on A, B and C when it's
> needed without incurring the storage necessary
> to maintain the function result.

Oh fiddle-faddle. What does this have to do with performance? Pure dogma.

> 2) If the data is not stable or bounded, there is a high
> potential for data integrity problems. Say that you
> calculate the sum of the values in a particular column
> and store the value but rows are still being added/deleted
> /updated to the table. The sum you calculate will cease
> to match the current value of the sum the moment anything changes.

These are OLTP issues. One would not put aggregated data in the same table with rapidly changing detail.

Neil Raden
Decision Support/Data Warehouse Consulting

Envirometrics, Inc.                            805.564.8672
133 E. De La Guerra St.                    805.962.3895 (fax)
Santa Barbara, California 93101         E-Mail:  NRaden_at_aol.com