if a field in source SFDC is dropped, would airbyte connecting to it and syncing into a target Snowflake table also drop the field from the target? the connection sync mode is Incremental | Append + Deduped and Detect and propagate schema changes is set to propagate field changes only.

No, Airbyte will not drop the field from the target Snowflake table in this scenario. According to Airbyte's schema change management behavior, when a column/field is removed from the source: > "The old field will be retained in the destination, but stop updating with updated values. If the connection is ever cleared or refreshed, the field and its historical data will be removed entirely." [Schema change management] So with your setup (Incremental | Append + Deduped + Propagate field changes only), here's what will happen: • The dropped SFDC field will remain in the Snowflake destination table with its historical data intact. • New syncs will simply stop populating that column — new or updated rows will have

NULL

(or empty) values for that field going forward. • The field will only be physically removed from the destination if you ever perform a full clear or refresh of the stream. One additional note: the "Propagate field changes only" setting means only column-level changes are propagated — new or removed streams are ignored. A dropped column is treated as a column-level change, so the behavior described above applies. [Schema change management] 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.

@kapa.ai but that is exactly what it did for some reason. wondering why

This is a known issue that has been reported. There is a GitHub issue specifically documenting this unexpected behavior with the Snowflake destination connector. A user reported that when a column was renamed in a Postgres source (using CDC/WAL mode), Airbyte treated it as a deletion of the old column and addition of a new one — and instead of retaining the old column in Snowflake as the documentation states, the old column was completely deleted from Snowflake, causing data loss. The reporter also confirmed that simply deleting a column (without renaming) produced the same unexpected behavior — the column was also deleted from Snowflake. [Snowflake column deletion issue] The reporter noted: > "This behavior contradicts what I understand from the official documentation, which states: 'The old field will be retained in the destination, but stop updating with updated values.'" So while the documented behavior is that dropped source fields should be retained in the destination, there appears to be a bug in the Snowflake destination connector (at least in certain configurations, particularly with CDC mode and "Propagate field changes only") where the column is actually dropped instead. The knowledge sources do not contain information about a confirmed fix for this issue. I'd recommend: 1. Checking the GitHub issue for updates on resolution status. 2. Reviewing your connector versions to see if a newer version of the Snowflake destination connector addresses this. 3. Being cautious about using "Propagate field changes only" with CDC sources and Snowflake until this is confirmed resolved. 👍 0 button 👎 0 button ℹ️ Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.