It depends. Are rows simply added to the table or are existing rows edited?

If added, how are PKs created? Find a method that will not allow duplicate PKs to be created. How do you make PKs unique across systems. Do you need to know which systems added the row?

If edited, you need to figure out how to handle update conflicts where System A edits a row, then System B edits the same row. System B finishes and saves. System A then finishes and saves overwriting Systems B's changes. Also, do you need to know which system edited the row?

How are deletes handled? Can one system delete a row that the other may need?



>I am working on a system where one part has a .dbc and a lot of tables attached to the dbc. The other part has free tables. Every now and then data from the free tables has to be transferred to tables in the dbc.
>
>Question: seeing as there are two 'apps' using the same tables, is there any buffering, etc that has to be taken into consideration. Each app is on a different server, but on the same network. Or is use x:\xyz again shared all that is needed?
>
>Many thanks!