Re: Is the use of VARCHAR(256) as Primary Keys preferred in Oracle?

I like to think of tables relating to themselves entirely separately than the data they contain.

I urge the use of surrogate keys, and if you must have a unique varchar2 value of 256 bytes, then do so with an unique index.

I would bet that as mentioned with other posters, that the 256 length is in error.

Jeff

"Russ Clancy" <rclancy_at_rotech.com> wrote in message news:c86e5ca8.0211161258.70aef47e_at_posting.google.com...
> Hello all,
> First off, the disclaimer to explain my ignorance of the Oracle
> Server. I am not an Oracle DBA. I am not an Oracle Developer. I am
> just the poor bastard in the office that knows enough about database
> design to be asked the database questions. Since I plead complete and
> utter ignorance to the nuts and bolts of Oracle, I ask this group for
> its help.
> A vendor we use has provided us with a schema of an Oracle database
> that their application utilizes. Aside from some basic normalization
> issues in the database design, I have a question about Oracle's
> efficiency in regards to using VARCHAR(256) as primary and foreign
> keys. The tables in question are without any type of surrogate keys.
> Would someone please explain (in relative layman's terms, I am a
> DBA, but not an Oracle DBA) the process Oracle goes through to
> optimize and utilize these values. The vendor swears up and down that
> this new design will increase performance. I am skeptical about
> that(Go figure, a DBA being skeptical about vendor developers'
> assertions). Do the inner workings of Oracle advocate the VARCHAR(PK)
> design and in what type of environment has to be in place for this
> scenario to work as advertised in terms of fragmentation, memory
> residence and allocation, process priority, etc. I will appreciate
> all of your responses and thank you in advance for your input.
Received on Mon Nov 18 2002 - 08:02:06 CST