Re: db design for measurement data?

1. at what rate do you expect each observation to change?
2. do you expect some to remain at the same value while others fluctuate more frequently?
3. if so, which ones?
4. when queries are run, are all variables per unit time requested, or is it more often to ask for only one or two of the values per unit time?
5. if only one or two variables per unit time are queried, which ones and with what frequency?

with a better understanding of how the data is used, a better model can be provided.

"Ronald Schirmer" <roschi_at_hrz.tu-chemnitz.de> wrote in message news:3B091420.21E82C87_at_hrz.tu-chemnitz.de...
> Hi Van,
>
> thanks for your reply.
>
> > It depends on what you want to do with the data. Do you need every row
> > available or will summaries suffice?
>
> I will need every row. There will be no chance to shorten or summarize
> the measurement results.
>
> > Does the data have to be available immediately or is some lag permitted?
>
> Depends on what you consider "some lag" :-) . When a query (e.g. via web
> frontend) is made to the data base, the results should be retrieved
> within a few minutes.
>
> > Can you use a datawarehouselike star schema?
>
> I've already heared about "datawarehouselike" things - but I've got no
> clue what that is? Could you, please, explain or do you no a useful URL
> dealing with that topic?
>
> > How long do you have to keep the data available online?
>
> This is going to be a long term project, so there's no
> end-of-store-date.
>
> > Can you partition your tables (in Oracle)?
>
> I'm not sure yet, which dbms to use. I wanted to decide after the db
> design.
>
> Thanks for every hint.
>
> Ronald.
Received on Sun Jul 22 2001 - 01:27:29 CEST