[Advanced-java] Fast Persistence of Unique Numbers

[jeff.singer@xxxxxxxxxx (Jeff Singer)]

Hi All

Thanks for all the replies.

>Can you just use an existing UUID or GUID utility instead of re-invenint
>the wheel?

Sorry, I wasn't clearer on my requirements. But we are limited to 9 digit
numeric numbers. As I understand it UUID and GUID are alpha-numeric.

>Depending on why you want to persist the number, you may be able to avoid
it. For instance Date.getTime >gives you a number that is unique, untill you
call it faster than the clock granularity.

With this I was limited by the fact that the number had to be unique for
longer than a day. So there is a slight chance that the numbers could
overlap? Also the Date.getTime returns a long, which breaks our 9 digit
requirement.

I am taking the approach that for a 9 digit numeric I need persistence, and
will use the method described by John Wilson. It also has the advance, as
mentioned, of being atomic.

Thanks Again
Jeff Singer

> -----Original Message-----
> From: advanced-java-admin@xxxxxxxxxxxxxxxxxxxxxx
> [mailto:advanced-java-admin@xxxxxxxxxxxxxxxxxxxxxx]On Behalf Of
> tpanton@xxxxxxxxxxxxx
> Sent: 21 January 2002 01:37
> To: advanced-java@xxxxxxxxxxxxxxxx
> Subject: RE:[Advanced-java] Fast Persistence of Unique Numbers
>
>
> Depending on why you want to persist the number, you may be able
> to avoid it. For instance Date.getTime gives you a number that is
> unique, untill you call it faster than the clock granularity.
> (If I recall on solaris you never get the same number twice).
> URL: http://www.westpoint.ltd.uk/ - internet recon.
>
> _______________________________________________
> Advanced-java mailing list
> Advanced-java@xxxxxxxxxxxxxxxxxxxxxx
> http://lists.xcf.berkeley.edu/mailman/listinfo/advanced-java