Re: Is this bad database design

• Post for FREE via your newsreader at post.usenet.com ****

pcmodeler wrote:

> The organization I work for is currently implimenting an Oracle based
> product from a 3rd party vendor (it tracks training history). In
> addition to the standard tables (associate information, etc.), there
> are two tables that store a lot of unlike information (employee
> status, country, delivery type, etc.).

> Essentially, in order to get employee status, I have to link from the
> associate table to the lookup table to the lookup table_string table
> in order to get their status. Then, in order to get their country, I
> have to link from the associate table to the lookup1 table (a
> copy/alias) to the lookup_string1 table (a copy/alias). Then, to get
> delivery type, I have to link from the associate table to the lookup2
> table (another copy/alias) to the lookup_string table (another
> copy/alias). This can happen up to 4 or more times. It would seem to
> me that this would cause some horrible performance problem in that
> they are throwing so much unlike information into these tables.

Mark, I had a little trouble following your description of the process. by copy/alias do you multiple passes through the table do you mean something like
"from table1 a, table1 b" where A and B are the same table and a different row is grapped from each pass based on where clause conditions?

There are a lot of applications that use parameter/value tables and the effect on performance will depend on exactly how and how often then tables have to be involved in queries.

The design does not sound great but it really isn't possible to judge based on the information given.

HTH -- Mark D Powell --

-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

• Usenet.com - The #1 Usenet Newsgroup Service on The Planet! *** http://www.usenet.com Unlimited Download - 19 Seperate Servers - 90,000 groups - Uncensored -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=