Re: Data model design problem

You could look at multidimensional database design (as for a datawarehouse) for good performance.
Your customer table could be a fact table with different dimensions that you mention (although this fact would be factless as it does not measure anything other than change)
Thanks
Aakash

"S Krishnan" <santakris_at_hotmail.com> wrote in message news:3b1eb79a_at_news.microsoft.com...
> Hi:
>
> Am designing a data model for an application that needs to track info on
> customers. However, the nature of information
> that needs to be tracked keeps on changing with time. For instance, today
 we
> don't need to find what cars people drive
> but this may become important tomorrow. All the information that would be
> collected can be classified into some categories
> such as demographic info, professional info, etc. However the specific
> fields within these categories can change. These changes will be done
> through the front-end or scripts.
>
> The primary use of this database is to query for customers with specific
> attributes, say, 30-50 year olds who work as designers, for instance.
> Typical queries are much more complex and performance is a significant
> factor in the design.
>
> One solution is to have a meta-data for the info fields and keep the
> customer data with foreign keys into the meta-data tables.
> However, this would impact performance.
>
> Is there a better solution ?
>
> tia,
> skrishnan
>
>
Received on Sun Jul 22 2001 - 01:30:48 CEST