Hi, Just a question which I haven't been able to answer by myself, for the Postgres source connector, in the Incremental + Deduped History mode, the deduplication is performed at the normalization step, right? So if I choose not to have the basic normalization step, I would have to implement the deduplication logic on my side. If so, is there a recommended way to do this? For example using ROW_NUMBER or DISTINCT ON ?

Yes, you need to implement the dedup process yourself. The best way probably is grouping by the primary key from your source.

Ok thanks! 🙂