On Dec 21, 2011, at 5:39 PM, Richard S. Russell wrote:

> Repeatedly tell them that they're crazy, that no competent database professional would EVER do it this way, and that they can ask a dozen other authorities if they don't believe you.

No competent database professional would use this value as the primary key in their table, perhaps, but that doesn't make the request invalid.

> If they're still pig-headed about it, devise a 2nd, parallel (and maybe best hidden-from-view) system that does it the right way, so you can sort out the inevitable confusions and mix-ups for these undeserving bozos.

No professional database developer with any amount of social or business competence would denigrate clients, particularly on a public mailing list.


On Dec 21, 2011, at 3:41 PM, Peter Buchanan wrote:

> Also, users have (against my advice...) insisted that when a document is cancelled, the number should be released for use again, and it must be used on the very next shipment to that port. So if 0019 was cancelled and 00030 had was the most recent document, then the next document would get 0019, and not 0031.

It appears you will want to script the process of assigning these values. But before doing so, I would inquire as to what the purpose is behind the request. Usually when I encounter such requests to ensure there are no gaps in the sequence, they are based in a desire to know that nothing has fallen through the cracks (or off the truck) either inadvertently or intentionally. Once you have the reasoning behind the request you can more competently suggest a means of satisfying the intent.

Cheers,
-corn


Cornelius Walker
The Proof Group
http://proofgroup.com/