Re: Data partitioning

We do something similar. What we do is create tables for the different
companies (with an underscore as first char). Then we create a meta-lookup
table that maps the customer_id to the correct table name.

There are some real big problems with this.
1. Tier 1 looks ugly, very ugly.
2. It takes some of the functionality that should be in the RDBMS, and
shoves it out the programmers, who make mistakes.
3. One more lookup, before you even put together your query.
4. If you have relations between products and other tables :( (look at #2)

We are moving away from this model to something more like what you are
doing. You clients may not like it, but they'll hate a really buggy system
more.

Christian Brink
CTO
ONSITE! Technology
www.onsitetech.com
503.233.1418
cb(at)onsitetech(dot)com

Taking e-Business and Internet Technology To The Extreme!

> -----Original Message-----
> From: pgsql-admin-owner(at)postgresql(dot)org
> [mailto:pgsql-admin-owner(at)postgresql(dot)org]On Behalf Of Ravindra Wankar
> Sent: Wednesday, December 19, 2001 11:02 PM
> To: pgsql-admin(at)postgresql(dot)org
> Subject: [ADMIN] Data partitioning
>
>
>
> We offer a web based application to companies. By keeping a company_id
> in the schema we differentiate the data amongst companies. e.g the user
> table has a company_id field to distinguish users between companies.
>
> However, most companies are feeling "insecure" about their data not
> being stored separately from others. Also from a maintenance perspective
> it seems it might be better. e.g restoring the data of an individual
> company will be impossible.
>
> Is there a way to handle this? Preferably using a single database?
>
> Thanks,
> -- Ravi.
>
>
> ---------------------------(end of broadcast)---------------------------
> TIP 2: you can get off all lists at once with the unregister command
> (send "unregister YourEmailAddressHere" to majordomo(at)postgresql(dot)org)