Is a Sanity Dataset like a database or a table?

Great question! A dataset is more like a database than a table.

In Sanity's architecture, a dataset is a named, isolated collection of all your JSON documents and assets within a project. Think of it as a complete, self-contained database environment where all your different content types (which you might think of as "tables") live together.

Here's the breakdown:

• Project = Your overall Sanity account/application
• Dataset = A complete database (like production, development, or staging)
• Document types (defined in your schema) = More like tables within that database

Each dataset is completely isolated from others - you can't directly query data from one dataset while connected to another. This isolation makes datasets perfect for creating separate environments. For example, you might have:

• A production dataset with your live content
• A development dataset where developers can test schema changes without affecting live content
• A staging dataset for validating changes before going live

The cool part is that datasets are schemaless, meaning different datasets can have completely different content structures if needed. You can configure your Studio to work with different datasets using environment variables, so running sanity start locally might use your development dataset while sanity deploy uses production.

Free plans include 2 datasets, and paid plans offer more depending on your needs. If you're on an Enterprise plan, you also get access to Advanced Dataset Management features like cloud cloning and hot swapping, which make managing multiple datasets even easier.