I have a merge replication which I was unfortunate enough to start modifying, in a way that required the subscriptions to get re-initialized. It makes very good sense and all, since I was touching some filters. But in my case, I changed something and then changed it back, so I know that the result of the new snapshot and new initialization will be identical to what I had.
However, the snapshot is re-created and the subscription is marked for re-initialization. And now the snapshot fails to get re-applied because there are data in subscriber's tables and the .bcp files in the snapshot have some data to load (same, actually, as we know). I can empty the .bcp file and then it seems to move on to the next table, so that's my strategy for now. I just wonder if anybody sees any pitfalls here. I suspect they are many and deep, just that my lack of experience doesn't let me see them.