Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> c.d.o.misc -> Re: Inserting a new PK into an existing table

Re: Inserting a new PK into an existing table

From: DA Morgan <>
Date: Thu, 31 Aug 2006 17:50:47 -0700
Message-ID: <>

joel garry wrote:
> DA Morgan wrote:

>> joel garry wrote:
>>> Agree with natual/surrogate post by Daniel - except, SSN is not
>>> guaranteed unique!  (The problem really being insufficient analysis of
>>> natural keys).
>>> jg
>>> --
>>> is bogus.
>> SSN is in most situations. It certainly meets the needs of insurance
>> companies. But you might want to do a two-field PK with SSN and DOB.
>> I like to use it because (A) I've never seen a duplicate and (B) if I
>> saw one I'd like it to generate an error so I could call security.

>>From  "We

> anticipate that the three-card per year limit will impact fewer than
> 10,000 individuals in any given year. For example, of the nearly 12.4
> million replacement SSN cards we issued in 2004, the number of
> individuals who requested more than three replacement cards was 3,818."
> Gee, now why would thousands of people need more than 3 replacement
> cards in a year...?

Thus my point in calling security.

> Of course, there are also situations (generally, government) where
> people may refuse to give their SSN. See
> (and search for the word "index"
> for an odd fact about whether SSN is indexed or not).

Except not as a condition of employment, or banking, or buying a house, etc. etc. etc. In the US if you choose not to identify yourself you'd best be living in a cave and living on grass and berries.

> Things may change:
> One-to-many:
> (I can't seem to find the duplicate ssn links among all the
> anti-gummint craziness in the minimal time allotted. But of course if
> two people refuse ssn's on religious grounds, you have to deal with
> that somehow).
> jg

I'd be interested in knowing from others here that reside in other countries whether they have a government equally capable of bungling with something that is theoretically a personal identifier. Any comments?

Daniel A. Morgan
University of Washington
(replace x with u to respond)
Puget Sound Oracle Users Group
Received on Thu Aug 31 2006 - 19:50:47 CDT

Original text of this message