Databricks Data Intelligence Platform · Topic 1 of 1

Platform Components, Catalog Explorer and Marketplace

Databricks Data Analyst: platform components, Unity Catalog objects, managed vs external tables, views, certified tables, metric views and Marketplace.

5 min

In a nutshell

Names are always catalog.schema.object; access needs USE CATALOG, USE SCHEMA and SELECT.
Dropping a managed table removes the data; dropping an external table leaves the files.
Lakeflow Declarative Pipelines build streaming tables and materialized views; Lakeflow Jobs schedule work.

What it is

The Databricks Data Intelligence Platform (formerly called the Lakehouse Platform) keeps data in open Delta Lake tables in cloud storage and runs every workload on top of them: ingestion, BI, AI and applications. One governance layer, Unity Catalog, controls access to all of it.

Catalog Explorer is the UI where you browse that governed data: catalogs, schemas, tables, views, permissions and lineage. Databricks Marketplace is where you obtain data and AI assets from outside providers.

How it works

The core components

Component	What it does
Delta Lake	Open table format: ACID transactions, schema enforcement, time travel
Unity Catalog	Central governance: permissions, discovery, tags, lineage, auditing
Databricks SQL	SQL editor, SQL warehouses, AI/BI dashboards, alerts, Genie Agents
Lakeflow Declarative Pipelines	Declarative ETL with streaming tables and materialized views (formerly Delta Live Tables)
Lakeflow Jobs	Orchestrates and schedules notebooks, queries, pipelines and dashboards
Mosaic AI	Build, serve and monitor ML models and AI agents
Lakebase	Managed Postgres database for OLTP and low latency application serving
Data Intelligence Engine	AI that learns your data's meaning and usage to power natural language search, Genie and automatic optimization

For an analyst, Databricks SQL is the daily workspace. The other components explain where the tables came from and who keeps them fresh.

The Unity Catalog namespace

Every object has a three level name: catalog.schema.object. A catalog usually maps to a business unit or environment, a schema to a project or domain, and the object is a table, view, volume, function or model.

Managed vs external tables

	Managed table	External table
Storage location	Managed storage chosen by Unity Catalog	A path you give with `LOCATION`
Format	Delta Lake	Delta, Parquet, CSV, JSON and others
`DROP TABLE`	Deletes metadata and data	Deletes metadata only; files stay
Predictive optimization	Yes	No

Databricks recommends managed tables by default. Choose external only when other tools must read the files at a fixed path.

Views

Standard view: a saved query, computed every time it is read, so it is always current.
Dynamic view: a view that uses functions such as current_user() or is_account_group_member() to filter rows or mask columns per user.
Materialized view: stores precomputed results and refreshes on a schedule or trigger.
Metric view: a governed semantic definition of measures and dimensions, created with CREATE VIEW ... WITH METRICS LANGUAGE YAML. Users group by any dimension and the engine computes the measure the same way everywhere, in SQL, dashboards and Genie Agents.

Certified tables, access controls and lineage

A certified asset is one a data owner marks as trusted with the system governed tag system.certification_status (values certified or deprecated). A check mark appears next to the name, and you need the ASSIGN permission on that tag plus ownership or APPLY TAG on the object.

The Permissions tab of an object in Catalog Explorer shows and changes grants such as SELECT or USE SCHEMA. The Lineage tab and lineage graph show upstream sources and downstream tables, notebooks, jobs and dashboards, down to column level.

Databricks Marketplace

Marketplace lists datasets, AI models, notebooks, apps and MCP servers from providers. Listings can be free, free samples, or commercial offerings that need provider approval.

Public listings are open to everyone; a private exchange shows listings only to invited consumers. Shared data arrives through Delta Sharing as a read only catalog in your Unity Catalog, so there is no copy to maintain.

How the exam asks about this

"An app needs fast single row reads and writes on current orders." Lakebase, the managed Postgres OLTP database.
"A dropped table's files must stay in storage for another tool." It must be an external table created with LOCATION.
"Revenue must be computed one way across SQL, dashboards and Genie." Define it once in a metric view.

What to remember for the exam

Names are always catalog.schema.object; access needs USE CATALOG, USE SCHEMA and SELECT.
Dropping a managed table removes the data; dropping an external table leaves the files.
Lakeflow Declarative Pipelines build streaming tables and materialized views; Lakeflow Jobs schedule work.
Certification uses the system.certification_status tag: certified or deprecated.
Metric views separate measures from dimensions so a metric is defined once.
Marketplace data arrives through Delta Sharing as a read only catalog.

Watch out

A standard view stores no data. If a question needs precomputed results for a fast dashboard, a standard view is the wrong answer; a materialized view is the right one.

Common mistakes

Treating Lakebase as a warehouse. Lakebase is an OLTP Postgres database for applications. Analytical queries over large tables belong on a SQL warehouse.

Confusing Lakeflow Jobs with pipelines. Lakeflow Declarative Pipelines define how tables are built. Lakeflow Jobs decide when and in what order tasks run.

Thinking Marketplace data is copied in. The provider shares it with Delta Sharing. You query it in place as a read only catalog and see the provider's updates.

Check yourself

What happens to the data files when you drop an external table?
Only the table's metadata is removed from Unity Catalog. The files stay at the external location.
How does a data owner mark a table as trusted for analysts?
They set the system governed tag system.certification_status to certified, which shows a certified check mark next to the table name.
Which component would you choose to define gross margin once for SQL, dashboards and Genie?
A metric view in Unity Catalog, which defines measures separately from dimensions so every tool computes the metric the same way.

Topic 1 of 1 in Databricks Data Intelligence Platform

Test yourself on Databricks Data Intelligence Platform

The full Databricks Data Analyst Associate question bank has 600 questions, with timed practice exams and an explanation for every answer.

Get Full Access Try Sample Exam

This study guide is independent and is not affiliated with, endorsed by, or sponsored by Databricks. Vendor names, certification names and exam codes are trademarks of their respective owners and are used for identification only.