On Wed, 25 Apr 2001 06:21:38 -0700 (PDT), Lawrence Lustig wrote:
>Having seen the various reponses to this question it's
>not clear if you're trying to solve the problem of
>getting unique IDs for a table (in which case AUTONUM
>is the way to go) or you are trying to solve the
>problem of unique IDs across multiple instances of the
>same database on different machines.
If there are fewer than 10 databases, you can set different autonumber
formulae for each database in this pattern:
--database 1:
AUTONUM idcolumn IN tablename USING 1001, 10
--database 2:
AUTONUM idcolumn IN tablename USING 1002, 10
etc.
Then the right-most digit identifies the source of the data.
If you have a potential for more than 10 databases, but fewer than 100,
use 100 as your interval:
--database 1:
AUTONUM idcolumn IN tablename USING 1001, 100
--database 2:
AUTONUM idcolumn IN tablename USING 1002, 100
You could also do this with the starting number, using different ranges,
but it's harder to deal with growth you didn't expect when your ranges
start to overlap.
Bill
