Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Surrogate Key vs Production Key

Re: Surrogate Key vs Production Key

From: Noons <wizofoz2k_at_yahoo.com.au>
Date: 13 Oct 2004 04:36:19 -0700
Message-ID: <73e20c6c.0410130336.188c14fa@posting.google.com>


ats <damorgan_at_x.washington.edu> wrote in message news:<1097551094.370084_at_yasure>...

>
> Solve this with a surrogate key if you can. I can't!

Me neither. Did I ever say I didn't need a unique key? I do recall writing about happilly trading off the overhead of the surrogate key and unique index for the flexibility. The overhead is minimal for the gain and certainly a good trade-off with modern hardware.

Let's see again.

natural keys, "parent" table:
one (or more) natural key columns.
one (possibly compound) unique index to enforce PK.

surrogate keys, "parent" table:
one extra column, for sk.
one unique index to enforce SK as PK.
one (or more) natural key columns.
one (possibly compound) unique index to enforce NK uniqueness.

Precisely as I said: one extra column, one extra index. Where exactly do you see one missing uniqueness integrity? Received on Wed Oct 13 2004 - 06:36:19 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US