|
|
Posted by Gert-Jan Strik on 08/18/07 13:37
Well, maybe someone did a large insert (of thousands of rows) and
cancelled the insert during operation.
Anyway, IMO an Identity column is a very poor idea if you really need a
sequence without gaps. Identity columns are simply not meant to do that.
If that is what you want, then I would remove the Identity property and
determine the sequence number myself (using a reliable method).
If the Identity column is simply used to generate a unique meaningless
surrogate key, then I wouldn't worry about gaps, and would attempt to
analyze it either...
Gert-Jan
teddysnips@hotmail.com wrote:
>
> And here's another thing that's gives me pause for thought (for other
> information see the thread entitled "ODBC Timeout problems but very
> hard to pin down")
>
> In APP 1 (Access front end, SQL Server 2000 backend) there is an
> operation to generate a discrepancy report. These are numbered
> sequentially, and this uses an Identity column with a seed and
> increment of 1. These reports are added approximately once a week.
> This week the value of the Identity column has jumped from 5,399 to
> 8,420. I need hardly add that no-one has added just over 3,000
> reports in the interim.
>
> Can this be related to the bizarre behaviour as reported in the other
> thread?
>
> Thanks
>
> Edward
Navigation:
[Reply to this message]
|