Skip to content
Tutorials Center
My learning

Topics

Lesson 5 of 36 ยท Platform and Workspace

A Tour of the Databricks Workspace

Learn how the Databricks workspace is organised: users, folders, notebooks, the catalog, objects and permissions, plus how account, workspace and metastore relate.

The workspace is the web environment where you do your work. It holds notebooks, folders, queries, dashboards, jobs and links to compute and data. Understanding what lives where saves a lot of confusion, especially the difference between things stored in the workspace (code, notebooks) and things stored in the catalog (data).

This lesson explains the layers around the workspace, the main object types and how permissions flow. It ends with two stand-ins so you can practise the logic.

Account, workspace, metastore

Three levels sit above your notebook:

flowchart TD
    A["Databricks account<br/>billing, users, groups"] --> W1["Workspace: dev"]
    A --> W2["Workspace: prod"]
    A --> M["Unity Catalog metastore<br/>one per region"]
    M --> W1
    M --> W2
    M --> C["catalogs โ†’ schemas โ†’ tables, volumes"]

One account can have several workspaces. A metastore in a region can be shared by workspaces so they all see the same catalogs.

  • The account is the top level: billing, identity (users and groups) and administration across workspaces.
  • A workspace is a deployment where people work. Teams often use separate workspaces for development and production.
  • A metastore (Unity Catalog) holds the data catalog. Attached to a workspace, it makes catalogs and tables visible there.

In Free Edition you have a single workspace and the account details are managed for you.

Workspace objects

The workspace file browser shows these object types:

Object What it is
Folder Organises other objects; has permissions
Notebook Cells of code and text
File A plain file such as .py, .sql or .md
Query A saved SQL query
Dashboard Charts built from queries
Alert A condition on a query that notifies you
Job A scheduled run of tasks
Pipeline A managed data pipeline definition
Experiment, Model MLflow objects

Your personal space is under Workspace โ†’ Users โ†’ your email. Shared work goes in Workspace โ†’ Shared or a team folder. Putting everything in your user folder makes handover painful when people leave, so use shared or Git-backed folders for team code.

The catalog is not the workspace

Notebooks live in the workspace. Tables, views and volumes live in the catalog, which has its own three levels:

Text
catalog          e.g. main, samples, prod
  โ”” schema       e.g. sales, bronze, silver
      โ”” table / view / volume / function / model

You refer to a table by its full name main.sales.orders. You can set defaults so you can write only orders:

SQL
USE CATALOG main;
USE SCHEMA sales;
SELECT * FROM orders LIMIT 5;

Permissions in brief

Access is granted to users, groups or service principals:

  • Workspace objects (notebooks, folders) have levels such as Can Read, Can Run, Can Edit and Can Manage.
  • Catalog objects (tables, schemas) use SQL privileges such as SELECT, MODIFY, USE CATALOG, USE SCHEMA and CREATE TABLE.

A privilege on a table is useless unless you can also reach its parents: to read main.sales.orders you need USE CATALOG on main, USE SCHEMA on main.sales and SELECT on the table. Privileges are inherited downward. Granting SELECT on a schema covers all current and future tables in it (read-only Databricks code):

SQL
GRANT USE CATALOG ON CATALOG main TO `analysts`;
GRANT USE SCHEMA ON SCHEMA main.sales TO `analysts`;
GRANT SELECT ON SCHEMA main.sales TO `analysts`;

Stand-in: finding objects in a folder tree

A folder tree is the same structure as nested dictionaries. The stand-in lists everything under a path, which is what the workspace browser does.

Python Stand-in: walk a workspace folder tree

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

The second stand-in models privilege checking, including the rule that you need access to every parent level.

Python Stand-in: privileges need every parent level

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Searching and navigating

The workspace search bar finds notebooks, queries, tables and more. The Catalog Explorer lets you browse schemas, see column types, sample data and lineage (which upstream tables feed this one). Learn it early: reading a table's details is often quicker than asking someone.

Tidy-workspace tips

  • Use naming like team/project/notebook and avoid spaces in names.
  • Keep production code in Git-backed folders, not in personal folders.
  • Delete scratch notebooks and tables, which cost storage and clutter search.
  • Use groups for permissions, not individual users, so access survives staff changes.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Where do tables and volumes live?

    Notebooks are workspace objects; tables, views and volumes belong to the Unity Catalog namespace.
  2. 2. To read main.sales.orders, which privileges are required?

    Access needs every level: USE CATALOG, USE SCHEMA and SELECT on the table, or on a parent that grants it by inheritance.
  3. 3. Why grant permissions to groups instead of individuals?

    Changing group membership updates access in one place instead of editing many grants.
  4. 4. What is a metastore?

    The metastore holds the catalog metadata and can be attached to several workspaces in a region.

Summary

  • Account โ†’ workspaces โ†’ metastore is the hierarchy above your notebooks.
  • Workspace objects hold code; the catalog holds data as catalog.schema.table.
  • Privileges flow down but need every parent level to be usable.
  • Share work through shared, Git-backed folders and group-based permissions.

Next: Compute and Clusters