When doing a synchronization script following a compare, it is not possible to take into account that the data being inserted might already exist.
A "synchronization type" (or the like) option for the synchronization script could take this into account, dictating the behaviour of insert statements.
The type options could be:
1. Insert without checking
2. Insert only new rows
(more could be thought up)
This would be checked on the comparison key.
Example of where this would be useful:
If it is possible to do localization of labels at different locations (each with a dedicated database), these localizations could be "harvested" to a central repository, and all localizations be re-deployed back to all installations, using the same synchronization script.
In this case the installation that was the origin of a localized label will have the data already, and checking for existence will be necessary, otherwise the insert will fail.
(.. or am I missing an existing option here