Hello! As far as I understand basic normalization has been deprecated. I need to normalize the json fields from the shopify connector. What is the best way to do that? Create a separate job outside of airbyte, or does airbyte have some internal mechanism that can be expanded to faciliatete this?

You are correct: basic normalization in Airbyte has been deprecated in favor of the Typing & Deduping feature. If you need to normalize JSON fields from the Shopify connector, here are your main options based on the current documentation: 1. Typing & Deduping: This is the recommended replacement for basic normalization. It is an internal Airbyte mechanism designed to handle data typing and deduplication, and it should be used for most normalization needs going forward. You can find more about this feature in the [Typing & Deduping documentation](https://docs.airbyte.com/platform/using-airbyte/core-concepts/typing-deduping). 2. Custom dbt Transformations: If you require more advanced or custom normalization logic (for example, to avoid unnesting repeated records into separate tables, as is common with Shopify data), you can use a custom dbt project. Airbyte allows you to configure a custom dbt transformation that runs after the sync. This approach gives you full control over how your JSON fields are normalized and how tables are structured. Note that the output location for custom dbt transformations is determined by your dbt project, not by Airbyte's namespace settings. For more details, see the [Airbyte documentation on custom dbt transformations](https://docs.airbyte.com/operator-guides/transformation-and-normalization/transformations-with-dbt). 3. External Processing: If neither Typing & Deduping nor custom dbt transformations meet your needs, you can always process the raw data outside of Airbyte after it lands in your destination. This is a valid approach, especially if you already have existing ETL/ELT pipelines or prefer to use other tools. Summary: Airbyte does provide internal mechanisms (Typing & Deduping, and custom dbt transformations) that can be expanded to facilitate normalization of JSON fields. Creating a separate job outside of Airbyte is not required unless your use case cannot be handled by these features. For Shopify and similar connectors, using a custom dbt project is a common solution for advanced normalization needs, such as controlling how nested JSON is handled and avoiding table proliferation. See the [forum discussion on Shopify normalization](https://discuss.airbyte.io/t/638) and the [normalization tutorial](https://docs.airbyte.com/operator-guides/transformation-and-normalization/) for more guidance. > Note: If you use custom dbt transformations, you are responsible for managing the dbt project and its output schemas, as Airbyte's namespace settings do not affect custom dbt outputs. > [Namespaces and normalization](https://docs.airbyte.com/platform/using-airbyte/core-concepts/namespaces#using-namespaces-with-basic-normalization)