Gary K
08/06/2026, 7:47 AMkapa.ai
08/06/2026, 7:48 AMDisable Final Tables option is automatically converted to the Legacy raw tables option. So if you upgrade without any preparation, your connection will continue writing raw tables only ā it won't automatically switch to direct-load final tables. [Snowflake Migration Guide]
How to switch to Direct-Load (final tables) Since the two modes (raw-only vs. final tables) are no longer supported in a single connection under 4.0.0, you have two options:
Option A: Switch the existing connection to Direct-Load 1. Before upgrading, note that your current connection has Disable Final Tables enabled.
2. After upgrading to 4.0.0, the connector will have Legacy raw tables enabled automatically.
3. To switch to direct-load, disable the Legacy raw tables option in the destination connector config.
4. Be aware this means the connector will now write to final tables only ā raw tables will no longer be written.
5. You will likely need to run a refresh/reset on the connection so data is re-synced into the new final table format.
Option B: Run both in parallel (if you need both raw and final tables) If you need to keep the raw tables and also get final tables, you must create two separate destination connectors:
⢠One with Legacy raw tables enabled (raw tables only)
⢠One with Legacy raw tables disabled (final tables via direct-load)
Then run two connections in parallel. [Snowflake Migration Guide] [Direct-Load Tables]
Additional permission requirement Before upgrading to 4.0.0, make sure your Snowflake user has ALTER TABLE permissions, as the new connector requires this for schema evolution. [Snowflake Migration Guide]
Summary š 0 button š 0 button ā¹ļø Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help., with interactive elementsGary K
08/06/2026, 7:52 AMkapa.ai
08/06/2026, 7:52 AMLegacy raw tables on an already-upgraded 4.0.0 connector.
What the knowledge sources do tell us is relevant context:
⢠Under Direct-Load, the connector writes typed data directly to final tables ā there is no T+D process that reads from raw tables to populate final tables. [Direct-Load Tables]
⢠The old raw tables are not automatically deleted after an upgrade ā they are left untouched. [Upgrading to Destinations V2]
Given these two facts, it is unlikely that simply disabling Legacy raw tables would automatically convert your existing raw table history into final tables. The direct-load paradigm doesn't use raw tables as a source to build final tables.
What I'd recommend The safest approach would be to:
1. Disable Legacy raw tables on the destination connector.
2. Run a full refresh/reset on the connection, which will re-sync all data from MySQL directly into the new final tables in the correct Direct-Load format.
This ensures your final tables are fully populated, but it does mean re-syncing your historical data from the source rather than converting from the existing raw tables.
I'd strongly suggest confirming this with Airbyte's support or community Slack before proceeding, to avoid any unintended data loss. š 0 button š 0 button ā¹ļø Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.Gary K
08/06/2026, 7:55 AMkapa.ai
08/06/2026, 7:55 AMDisable Final Tables is enabled at upgrade time ā Legacy raw tables is automatically enabled (raw-only mode). [Snowflake Migration Guide]
⢠The upgrade process to 3.0.0 (Destinations V2) involved copying raw table data into the new format ā but there's no equivalent description of such a migration path for the 4.0.0 Direct-Load upgrade. [Upgrading to Destinations V2]
So while re-enabling final tables before upgrading might result in a cleaner upgrade path, the knowledge sources don't explicitly describe this scenario or guarantee that your raw table history would be carried over into the final tables.
I'd strongly recommend reaching out to Airbyte directly via the community Slack or support before attempting this, as the wrong sequence of steps could risk data loss or an inconsistent table state. š 0 button š 0 button ā¹ļø Tip: Mention @kapa.ai in the thread for follow-ups or DM the app for 1-to-1 help.