Databricks' beta Shopify connector sees only 60 days of orders until Shopify approves a scope

The connector takes over the layer most Shopify data teams build by hand: Admin API calls, pagination logic and retry handling ^[8]. The layers after it stay where they are. Bronze tables land in Unity Catalog, and the team still transforms them into the silver and gold layers that BI tools, Genie or ML workloads read ^[8]. A dev.to walkthrough built on the Databricks and Shopify documentation reports that the limitations page is explicit on this point: the connector ingests raw data without transformations and expects modeling to happen downstream ^[18][6].

The set of raw tables is fixed. The pipeline documentation lists 39 source tables in a default source schema. A pipeline can declare them one at a time or take the whole schema in one declaration ^[4]. Custom metafields, Shopify reports and objects outside those 39 tables are not covered ^[7]. Pipelines can be authored in the UI, through the API or as Declarative Automation Bundles, with orchestration through Databricks Workflows and SCD Type 2 history tracking ^[5].

The Shopify requirements are short. Any plan works, because the connector reads through the Shopify Admin API and every plan has it ^[13]. Someone creates an app in the Shopify Dev Dashboard, in the same Shopify organization as the store. The connector then authenticates with the OAuth client credentials grant ^[14].

The Databricks side is ordinary Unity Catalog grants. Whoever creates the connection needs CREATE CONNECTION on the metastore, or USE CONNECTION to reuse an existing one. They also need USE CATALOG on the target catalog, plus USE SCHEMA and CREATE TABLE on the target schema, or CREATE SCHEMA on the catalog ^[12]. The walkthrough notes that a different person commonly owns each side of the setup ^[20]. In most organisations that means two tickets with two owners before the first sync runs.

I think this is the right default for analytics on a store whose data fits the 39 tables. My context for that is a team whose extractor exists only to feed reporting, so the modeling work was always going to be theirs ^[6][8]. The walkthrough reaches the same verdict, calling the connector a strong default for analytics on Shopify data, with beta limitations to plan around before production use ^[19].

That guidance has a date on it. The walkthrough reflects the documentation as of late September 2026 and tells readers to check the linked docs for changes before building, because the connector is in Beta ^[18].