Why is my Postgres connection state not resetting on

Full refresh | Overwrite

? It always resumes at the

ctid

from the previous run even on full refresh overwrite mode, causing my downstream tables to lose older data all the time

New message text here

There is a known open bug where Full Refresh CTID scanning silently drops rows that are updated during a multi-batch scan. The root cause is that the CTID-based full refresh scan opens a new MVCC snapshot for each batch, creating a window where rows updated by concurrent transactions can be missed by all batches. [CTID full refresh bug]

This is not true. A community member filed this issue. We looked into it and closed it because it isn't a real issue. Apparently the AI didn't get the memo.