Hey folks, Is it possible to stop the creation of multiple tables for streams created from API data sources that have nested data? Ideally I just have one table with the top level data and I can nest as I need to further down my pipeline. I’m not referring here to stopping normalization altogether.

We are releasing soon a new feature to run your custom dbt transformation: (issue to track its progress is here: https://github.com/airbytehq/airbyte/issues/2959) Here is an updated tutorial linked to that new feature: https://github.com/airbytehq/airbyte/blob/97f8f3847af7c80df257c28b18af1ad433b8397f[…]ls/transformation-and-normalization/transformations-with-dbt.md This describes how to export the normalization project that is being generated, then from there you could edit it to remove the additional exploded tables for nested data and run your new “custom” normalized dbt project as part of airbyte: https://github.com/airbytehq/airbyte/blob/97f8f3847af7c80df257c28b18af1ad433b8397f[…]ransformation-and-normalization/transformations-with-airbyte.md

Got you! So the proposed way to do it going forward is to either allow the user to tinker with normalization or to do a thorough normalisation of all nested data structures encountered? There will never be a assumed level of normalisation?

there will but it might come later for example this issue: https://github.com/airbytehq/airbyte/issues/3522 the issue is not specifically about turning off the nested tables in their separate table… But we do think about having that option (toggling nested) exposed as configuration too

Ah got you