Skip to main content
A data product is a curated, self-contained unit of data designed to serve specific consumer needs. In CDF, data products are a governance layer around data modeling: they wrap data modeling views with ownership, metadata, usage terms, and versioned interfaces that consumers can discover and use.

Schema space and ownership

Each data product owns a schema space. The views that the data product governs live in that space, and the schema space belongs exclusively to the data product. Set schemaSpace when you create the data product. The space must already exist and must not belong to another data product. You cannot change schemaSpace later. The data product carries ownership and access through its schema space and dataProductsAcl. Typical actions are READ, CREATE, UPDATE, and DELETE. Creating a version requires UPDATE on the parent data product, not a separate create action.

Versions and views

A data product is the parent of one or more data product versions. A version is identified by a semantic version (major.minor.patch, for example 1.0.0) and references a collection of views. After you add a view reference, you cannot remove or change it. You can add more view references later, and you can still update metadata such as description, usage terms, and status. Consumers get a documented, versioned interface rather than an ad hoc set of views.

When to use a data product

Use a data product when you publish data modeling views for others to consume. The product gives consumers a stable, versioned interface, usage terms, and a schema space with clear ownership without exposing how you organize source pipelines internally.

Limits

  • A project can have a maximum of 100 data products.
  • The create endpoint accepts one data product per request.
  • You cannot delete a data product while it still has versions. Delete the versions first.
  • Deleting a version also removes the access filters that applied to that version.

Key capabilities

  • Own a schema space for the views the product governs
  • Version views with view references you can add but not remove or change, and updatable metadata
  • Publish usage terms and documentation for consumers
  • Control access with dataProductsAcl
Last modified on September 24, 2026