We are using merge replication with pull subscriptions on SQL Express 2005 and a publisher on SQL 2012. We recently upgraded to a merge join filter system. The merge join filters use a column that is also part of a compound foreign key. On the publisher
this foreign key employs cascade update/delete with 'not for replication'. However, the snapshot does not create that foreign key. The same table includes other foreign keys which are propagated to the subscribers. Is there an issue with using a single column
of a compound foreign key as a merge join filter? Is there an issue with using the cascade option with a merge join filter? We are baffled and have found little, if any, related information on the web or in books online. Perhaps we're searching in the wrong
area? Does anyone have any insight into why the FK is not being created as part of the snapshot? Thanks!
↧