If something is wrong during snapshotting of SQL server table, and cause restart of snapshotting, duplicated events will be generated in BQ.
The workaround of this issue is that, users need to stop the replicator pipeline and manually delete the target table and then restart the pipeline.
Discussed with Sagar, looks like it’s tight for 6.3.0 timeline.
Bhooshan, need our decision about whether we need to fix it in 6.3.0