Data & Analytics
Airbyte schema review can reopen fields you excluded
Protect field-selection rules when a source schema changes and Airbyte offers a review.
November 11, 2026
A source table gains a harmless column. During Airbyte’s Review changes flow, fields previously excluded to keep personal data out of Snowflake appear selected again in a self-managed 2.4.0 issue. Saving that review could restart transfer of data the team deliberately excluded.
Preserve the approved field set
Before approving a schema change, export or record the connection’s selected fields by stream. Compare the proposed selection against the approved set, highlighting both newly discovered columns and previously excluded columns. Airbyte’s field-selection documentation says newly added fields may be selected by default; that is different from reselecting an existing field the team had deselected.
Make the review a two-person check for streams carrying sensitive data. The privacy owner decides whether a new column is allowed, while the operator confirms that old exclusions remain intact. A field-selection checkbox controls what is written to the destination; Airbyte notes that a source connector may still read deselected fields. Do not mistake destination exclusion for source-side minimization.
Check the first sync after approval
The schema-change guide says manual approval keeps the previously saved schema until changes are approved. After saving, inspect the connection config and destination schema, then run a small controlled sync. Check that excluded columns and values did not reappear. If they did, stop the connection and follow the organization’s data-handling process before resuming.
Adovo’s view
A schema change is also a policy change when field selection carries a privacy promise. Approval should compare the whole contract, not only the new column. The safest workflow makes every unexpected expansion visible before data moves.
