---
title: "Best data catalogs for Databricks"
url: "https://atlan.com/know/databricks/best-data-catalogs-for-databricks/"
description: "Best data catalogs for Databricks: Unity Catalog is the native baseline. See how Alation, Atlan, Collibra, DataHub and OpenMetadata connect to it."
author: "Emily Winks"
author_role: "Data Governance Expert"
published: "2026-11-18"
updated: "2026-11-18"
---

## What is the best data catalog for Databricks?

Unity Catalog is the native baseline: Databricks builds it in and turns it on for workspaces created after November 8, 2023, and a Databricks-only estate may need nothing more. When your estate reaches past Databricks, five third-party catalogs each document a Unity Catalog route: Alation through an OCF connector, Atlan by crawling each metastore, Collibra through Edge, DataHub through its Databricks ingestion source and OpenMetadata through a dedicated connector.

Unity Catalog plus five third-party catalogs:

- **Unity Catalog:** the native governance layer built into Databricks
- **Alation:** a Unity Catalog OCF connector with lineage from system tables and DDL
- **Atlan:** a crawler per metastore with column-level lineage and tag and description sync back
- **Collibra:** inbound and outbound sync through Edge, including attribute values to Unity Catalog tags
- **DataHub:** a Unity Catalog source with opt-in metric views, notebooks and Genie spaces
- **OpenMetadata:** a Unity Catalog connector with lineage, data quality and tags

## Key takeaways

- Unity Catalog is on by default for Databricks workspaces created after November 8, 2023.
- Alation, Atlan, Collibra, DataHub and OpenMetadata each document a route into Unity Catalog.
- Several catalogs read the same Unity Catalog lineage system tables, so compare what each adds on top.
- A Databricks-only estate can often start with Unity Catalog alone.

---

Databricks Unity Catalog is the governance layer built into Databricks, on by default for every workspace created after November 8, 2023. Five third-party catalogs document a route into it: Alation, Atlan, Collibra, DataHub and OpenMetadata. Atlan's docs say it crawls Unity Catalog schemas, tables, views, volumes and AI models and builds column-level lineage automatically. For a platform team whose data also lives outside Databricks, the cost of the wrong pick is data that stays hard to find: Databricks says assets not registered in Unity Catalog might not be discoverable through some of its tools.

### Map Your Catalog Coverage Beyond Databricks

Give it every system in your estate and which catalog governs each today. It returns a coverage table and the specific gap ranked by what breaks first. [Read the skill](/skills/catalog-coverage-gap-map.md).

*Paste into a new chat*

```
Use the skill at https://atlan.com/skills/catalog-coverage-gap-map.md to map catalog coverage across our Databricks estate and the platforms outside it. Ask me for whatever it needs.
```

*Run once in a terminal*

```
curl -fsSL --create-dirs \
  -o ~/.agents/skills/catalog-coverage-gap-map/SKILL.md \
  https://atlan.com/skills/catalog-coverage-gap-map.md
```

*For an agent*

```
curl -fsSL https://atlan.com/skills/catalog-coverage-gap-map.md
```


    Human
    Agent


  ClaudeChatGPTGeminiCursorOther
  Copy

Unity Catalog comes first as the native baseline Databricks builds in. For a Databricks-only estate it is often enough. The five third-party catalogs follow in alphabetical order. Every cell comes from the vendor's own documentation for its Unity Catalog route, read on November 17, 2026. Other catalogs also document a Unity Catalog route, Microsoft Purview among them, which the section on when to add a catalog covers.

* Unity Catalog covers eight areas in the Databricks documentation: access control, discovery, lineage, auditing, data classification, data quality monitoring, data sharing and AI governance.
* Databricks discovery tools are optimized for data governed by Unity Catalog.
* External engines read Unity Catalog through its implementation of the Iceberg REST catalog API.
* Catalog federation brings other catalogs in: Databricks connects Hive Metastore, AWS Glue or Snowflake catalogs, read only, for an incremental migration or a long-term hybrid model.

| Item | Detail |
|---|---|
| Native baseline | Databricks Unity Catalog |
| Third-party catalogs covered | 5, listed alphabetically: Alation, Atlan, Collibra, DataHub, OpenMetadata |
| Open interface external tools use | Iceberg REST catalog API, implemented by Unity Catalog |
| Open-source Unity Catalog | Apache 2.0 license, an LF AI & Data project |
| Ordering | Native baseline first, then alphabetical. No scores |
| Last checked | November 17, 2026 |

---

## What counts as a data catalog for a Databricks estate?

A data catalog for Databricks works at one of two layers: the governance layer Databricks builds in, or a third-party catalog that reads from it. The Databricks documentation calls Unity Catalog "the unified governance layer for data and AI built into Databricks." It enforces access control, tracks lineage and logs activity for auditing.

Unity Catalog organizes objects in a three-level namespace, catalog.schema.object. Those objects include tables, views, volumes, functions, models and services, which cover model services and MCP services.

Business meaning sits on top. Metric views, domains, Pages, wikis and certification make up what Databricks calls Unity Catalog semantics, and [Unity Catalog business glossary and domains](https://atlan.com/know/ai-agent/databricks/unity-catalog-business-glossary-domains-which-features-are-ga-preview/) walks through each one. Together they form the modeled knowledge in [Genie Ontology](https://atlan.com/know/ai-agent/databricks/genie-ontology/).

A third-party catalog reads those objects from outside. Databricks lists external engines that work through its Unity REST API or the Iceberg REST catalog API, among them Spark, Trino, DuckDB, Dremio and Starburst. Those are query engines reading tables. Third-party catalogs collect metadata through their own connectors, and each vendor documents its route separately; the profiles below name each one.

The two layers differ in reach. [Data catalog vs context layer](https://atlan.com/know/data-catalog-vs-context-layer/) covers the wider shift, where catalog metadata becomes context an AI agent can use. For a Databricks team, the practical question is which layer holds that context for the systems Unity Catalog does not govern.

---

## Best data catalogs for Databricks: the comparison table

The table lists Unity Catalog first as the native baseline, then five third-party catalogs in alphabetical order. Each row says how the catalog reaches Unity Catalog and what its own documentation names for that route.

| Catalog | Role | How it connects to Unity Catalog | What its docs name |
|---|---|---|---|
| **Databricks Unity Catalog** | Native baseline | Built into Databricks; on by default for workspaces created after November 8, 2023 | Access control, discovery, lineage, auditing, data classification, data quality monitoring, data sharing, AI governance |
| **Alation** | Third-party | Databricks Unity Catalog OCF connector | Lineage from Unity Catalog lineage system tables and DDL statements |
| **Atlan** | Third-party | Crawls Unity Catalog metadata, one connection per metastore | Tables, views, volumes, AI models, column-level lineage, tag import, tag and description sync back to Databricks |
| **Collibra** | Third-party | Collibra Edge | Inbound sync of databases, schemas, tables, views, metric views, columns, volumes, directories, files and notebooks; sampling, profiling and classification; outbound sync of attribute values to Unity Catalog tags |
| **DataHub** | Third-party | The databricks (unity-catalog) ingestion source | Datasets, schema fields, containers, table and column lineage, usage, profiling, ownership, metric views, notebooks, Genie spaces, tags |
| **OpenMetadata** | Third-party | Unity Catalog connector | Metadata ingestion, incremental extraction, data quality, lineage, dbt, tags |

Each row reflects the vendor pages listed in Sources, so a shorter row is not a feature count. Confirm anything you plan to depend on in the vendor's current documentation.

### Write-back into Unity Catalog

On the vendor pages listed in Sources, two catalogs document writing back into Unity Catalog: Collibra syncs attribute values to Unity Catalog tags, and Atlan syncs tag and description edits back. [Unity Catalog vs Collibra](https://atlan.com/know/ai-agent/context-layer/unity-catalog-vs-collibra/) compares what each of those two governs. For the same shortlist format on another platform, see [best data catalogs for Microsoft Fabric](https://atlan.com/best-data-catalogs-for-microsoft-fabric/).

---

## How are these catalogs ordered?

Databricks Unity Catalog comes first because it is the native baseline every Databricks workspace with Unity Catalog already runs. That position reflects how Databricks ships, and it carries no score.

The five third-party catalogs follow in alphabetical order: Alation, Atlan, Collibra, DataHub, OpenMetadata. Nothing is scored. The same order holds in the table, the profiles and the lists.

We read each vendor's own documentation for its Unity Catalog route on November 17, 2026, and each profile describes the pages listed in Sources. An evaluation scorecard belongs to your team, built from your estate. The [data catalog RFP template](https://atlan.com/know/data-catalog/data-catalog-rfp-template-with-evaluation-criteria/) turns a shortlist into questions for vendors, and [context layer evaluation criteria](https://atlan.com/know/ai-agent/context-layer/context-layer-evaluation-criteria/) adds the tests for AI agent use.

---

## Databricks Unity Catalog: the native baseline

Unity Catalog is the governance layer built into Databricks. The Databricks documentation says it is automatically enabled for all workspaces created after November 8, 2023.

For workspaces created earlier, [Unity Catalog vs Hive Metastore](https://atlan.com/know/ai-agent/databricks/unity-catalog-vs-hive-metastore/) covers what changes on migration.

### What Unity Catalog covers

Unity Catalog governs tables, views, volumes, functions, models and services in one catalog.schema.object namespace. The Databricks documentation lists eight capability areas: access control, discovery through Catalog Explorer, lineage, auditing, data classification, data quality monitoring, data sharing and AI governance.

### Discovery, lineage and metric views

Discovery runs through Genie One, the Discover page, Catalog Explorer and keyword and semantic search. Genie One lets people browse shared assets, search by name, ask questions in natural language and filter by domain; [what Genie One is](https://atlan.com/know/ai-agent/databricks/genie-one/) goes deeper. Catalog Explorer can also provide AI-generated comments for tables.

Lineage is captured automatically for queries run on Databricks. It goes down to the column level, across all workspaces attached to the metastore. Queries must use Spark DataFrame or Databricks SQL interfaces, and tables must be registered in a Unity Catalog metastore. Lineage shown in Catalog Explorer is retained indefinitely, and the Databricks docs say all lineage captured after September 1, 2024 is available there.

Databricks calls metric views "the core implementation of Unity Catalog semantics." You define a metric once and query it by any dimension from SQL editors, notebooks, dashboards, Genie Agents, alerts and BI tools such as Power BI, Tableau and Sigma. [How to create Databricks metric views](https://atlan.com/know/ai-agent/semantic-layer/how-to-create-databricks-metric-views/) has the steps, and [Unity Catalog metrics](https://atlan.com/know/ai-agent/databricks/unity-catalog-metrics/) explains the model.

### Semantics, domains and the glossary

Unity Catalog semantics groups five components. Metric views govern KPIs. Domains group assets by business purpose. Pages define business concepts for people and Genie One to reference. Wikis document what a table means. Certification and deprecation mark assets as trusted or outdated.

Databricks says these make up the modeled knowledge in Genie Ontology, which Genie One combines with knowledge learned from assets and usage into "a unified context layer."

### Open interfaces

Unity Catalog implements the Iceberg REST catalog API, so Spark, Flink and Trino can read and write Iceberg tables registered in it. Managed Iceberg tables support read and write; foreign Iceberg tables and Delta tables with Iceberg reads enabled are read only. External access needs the setting enabled on the metastore, the EXTERNAL USE SCHEMA privilege and a token or OAuth. Iceberg clients can also read Unity Catalog table and column tags.

An open-source Unity Catalog also exists, under the Apache 2.0 license as an LF AI & Data project. [Apache Polaris vs Unity Catalog vs Glue vs Snowflake Open Catalog](https://atlan.com/know/iceberg/polaris-vs-unity-catalog-vs-glue-vs-snowflake-open-catalog/) compares the open catalog options.

An estate that lives entirely in Databricks may not need a second catalog at all. The comparisons with [AWS Glue Data Catalog](https://atlan.com/know/ai-agent/aws/unity-catalog-vs-aws-glue-data-catalog/) and [Dataplex](https://atlan.com/know/ai-agent/gcp/unity-catalog-vs-dataplex/) show where its scope sits against other clouds' native catalogs.

---

## Alation: an open connector route to Unity Catalog

Alation reaches Unity Catalog through its Databricks Unity Catalog OCF connector, built on Alation's Open Connector Framework.

**Pros**

* Alation's docs list it as a core connector, included with all Alation platform tiers subject to each tier's connector limits, and fully supported by Alation.
* Lineage combines two inputs: lineage metadata from Unity Catalog lineage system tables and DDL query statements run in Compose.

**Good to know**

* Direct lineage extraction reads system.access.table_lineage and system.access.column_lineage.
* Alation's docs say direct lineage extraction is on by default.

**Capabilities**

* Unity Catalog metadata extraction through the OCF connector.
* Lineage that joins Databricks system-table lineage with DDL captured in Compose.

**Choose Alation if** your organization already runs Alation and wants Databricks lineage alongside the sources it catalogs today. [Alation vs Informatica](https://atlan.com/know/ai-agent/context-layer/alation-vs-informatica/) sets Alation's catalog approach against another long-standing catalog.

**Official sources:** Alation documentation (docs.alation.com): Databricks Unity Catalog OCF Connector overview; Databricks Unity Catalog OCF lineage extraction.

---

## Atlan: cross-platform context for a Databricks estate

Atlan, an AI Context Platform, crawls Unity Catalog into one context layer with the other platforms it connects to.

**Pros**

* Crawls schemas, tables, views, materialized views, volumes, external locations and AI models, and builds column-level lineage automatically.
* Analyzes the same system tables Databricks uses, including the query history table system.query.history and the lineage tables system.access.table_lineage and system.access.column_lineage, then filters lineage to valid relationships.
* Imports Unity Catalog tags, and reverse sync pushes tag changes made in Atlan back to Databricks. Descriptions written in Atlan can land in Databricks as comments.

**Good to know**

* One connection per metastore covers every catalog, schema and table the service principal can see.
* Unity Catalog keeps enforcing access: grants, row filters, column masks and audit logs stay in Databricks, and Atlan reads and enriches the metadata.

**Capabilities**

* Context Engineering Studio deploys context to Genie: it pushes descriptions into Unity Catalog, creates one metric view per table and configures a Genie Space. See [adding enterprise context to a Genie Space](https://atlan.com/know/ai-agent/databricks/add-context-to-databricks-genie-space/) and [Genie Ontology and the Atlan context layer](https://atlan.com/know/ai-agent/databricks/genie-ontology-and-atlan-context-layer/).

**Choose Atlan if** Databricks is one platform among several and you want catalog edits to flow back to Unity Catalog.

**Official sources:** [Atlan Databricks docs](https://docs.atlan.com/apps/connectors/data-warehouses/databricks); [lineage FAQ](https://docs.atlan.com/apps/connectors/data-warehouses/databricks/faq/atlan-vs-databricks-lineage); [governance FAQ](https://docs.atlan.com/apps/connectors/data-warehouses/databricks/faq/unity-catalog-governance); [Deploy to Genie](https://docs.atlan.com/product/capabilities/governance/context-engineering-studio/how-tos/deploy-databricks).

---

## Collibra: Edge-based integration with Unity Catalog

Collibra's documentation states that you can integrate Databricks Unity Catalog with Collibra only through Edge.

**Pros**

* Inbound sync brings in databases, schemas, tables, database views, metric views, columns, volumes, directories, files and notebooks.
* Outbound sync propagates attribute values from Collibra to tags in Databricks Unity Catalog, so values your stewards set in Collibra reach the tags Databricks users see.

**Good to know**

* Collibra's docs say the Unity Catalog integration lets you set up sampling, profiling and classification.
* Technical lineage for Unity Catalog through Edge has its own page in Collibra's documentation.

**Capabilities**

* Two-way metadata sync with Unity Catalog through one Edge integration.

**Choose Collibra if** your governance workflows already run in Collibra and you want Unity Catalog tags to carry the attribute values those workflows set.

**Official sources:** Collibra documentation (productresources.collibra.com): About the Databricks integration, accessed November 17, 2026.

---

## DataHub: a Unity Catalog ingestion source with opt-in Genie spaces

DataHub ingests Unity Catalog through its databricks ingestion source, which its docs point Unity Catalog workspaces to.

**Pros**

* Covers datasets, tables, views, schema fields and containers.
* Captures table- and column-level lineage, usage statistics, data profiling, ownership and stateful deletion detection.
* Maps Unity Catalog tags to DataHub tags.

**Good to know**

* Several entity types are opt-in. Metric views come in as datasets with include_metric_views, notebooks with include_notebooks, and Genie spaces as AI agents with include_genie_spaces.
* For Genie spaces, the ingestion identity needs CAN EDIT on each space it ingests.

**Capabilities**

* One ingestion source that brings Unity Catalog structure, lineage, usage and profiling into DataHub's metadata graph.
* Genie spaces cataloged next to the tables and metric views they declare, with those declared datasets as upstream lineage.

**Choose DataHub if** you want Genie spaces, notebooks and metric views cataloged as entities next to the tables they use. [DataHub pricing](https://atlan.com/know/ai-agent/context-layer/datahub-pricing/) breaks down what self-hosting and the cloud offering cost, and [open source vs commercial data catalog total cost of ownership](https://atlan.com/know/data-catalog/open-source-vs-commercial-data-catalog-total-cost-of-ownership/) puts those numbers in context.

**Official sources:** DataHub documentation (docs.datahub.com): Databricks ingestion source, accessed November 17, 2026.

---

## OpenMetadata: a Unity Catalog connector with lineage and tags

For Unity Catalog, OpenMetadata ships a dedicated connector with its own setup page.

**Pros**

* Supports three ways to authenticate: a personal access token, Databricks OAuth or Azure AD.
* Ships workflows for metadata ingestion, incremental extraction, data quality, lineage and dbt integration.

**Good to know**

* Lineage needs permissions on system.access.table_lineage and system.access.column_lineage. The docs describe it as table and column-level lineage from Databricks system tables.
* Tag extraction reads catalog, schema, table and column tags.

**Capabilities**

* An incremental extraction workflow alongside full metadata ingestion.
* A data quality workflow documented on the same connector page as metadata ingestion.

**Choose OpenMetadata if** you want data quality workflows and lineage to run from the same Unity Catalog connector. [What OpenMetadata is used for](https://atlan.com/know/ai-agent/data-for-ai/what-is-openmetadata-used-for/) describes the wider product.

**Official sources:** OpenMetadata documentation (docs.open-metadata.org): Unity Catalog connector, accessed November 17, 2026.

---

## When do you need a catalog beyond Unity Catalog?

Consider a catalog beyond Unity Catalog in three situations that follow from the scope the Databricks documentation describes: data that is not registered in Unity Catalog, platforms that stay outside Databricks, and definitions that have to work beyond it.

1. **Assets outside Unity Catalog.** Databricks says its discovery tools are optimized for data governed by Unity Catalog, and assets that are not registered "might not be discoverable" through some of them.
2. **A long-term hybrid estate.** Catalog federation connects Hive Metastore, AWS Glue or Snowflake catalogs read only, for incremental migration or "a long-term hybrid model."
3. **Definitions that must work outside Databricks.** BI tools such as Power BI, Tableau and Sigma can already query metric views in Databricks. A KPI that Snowflake also computes on its own data needs a definition the owners on both platforms agree on, and [warehouse-native vs universal semantic layer](https://atlan.com/know/ai-agent/semantic-layer/warehouse-native-vs-universal-semantic-layer/) lays out that trade-off.

### Adding a catalog on top of Unity Catalog

A second catalog can sit on top without displacing the first. [Can a data catalog sit on top of existing catalogs](https://atlan.com/know/data-catalog/can-a-data-catalog-sit-on-top-of-existing-catalogs-without-replacing-them/) explains how, and [how to extend a data catalog for AI readiness](https://atlan.com/know/ai-agent/ai-ready-data/extend-data-catalog-for-ai-readiness/) covers what to add first. When the estate spans clouds, the [multi-cloud context layer](https://atlan.com/know/ai-agent/context-layer/multi-cloud-context-layer/) guide covers ownership and portability.

Microsoft Purview also documents a scan for Unity Catalog, covering the metastore, catalogs, schemas, tables, views, tags and lineage from notebook runs, and the same Microsoft page notes that Microsoft Purview Unified Catalog and classic Purview data governance capabilities are being retired, with maintenance mode through November 12, 2027; [Unity Catalog vs Microsoft Purview](https://atlan.com/know/purview-vs-databricks-unity-catalog/) compares the two.

Check the cheaper fixes first. If registering the asset in Unity Catalog, federating the other catalog or pointing a BI tool at a metric view closes the gap, Unity Catalog still carries the load.

---

## How does lineage compare across these catalogs?

Lineage across these catalogs mostly starts from the same place, the Unity Catalog lineage system tables, and differs in what each catalog joins to it. The table lists only what each vendor's documentation names as its lineage source for Unity Catalog.

| Catalog | Lineage source named in its docs |
|---|---|
| **Databricks Unity Catalog** | Captured automatically for queries run on Databricks through Spark DataFrame or Databricks SQL, down to the column level, for tables registered in a Unity Catalog metastore |
| **Alation** | Unity Catalog lineage system tables (system.access.table_lineage, system.access.column_lineage) plus DDL statements from Compose |
| **Atlan** | The same system tables Databricks uses: system.query.history, system.access.table_lineage, system.access.column_lineage |
| **Collibra** | Technical lineage for Unity Catalog through Edge, documented on its own Collibra page |
| **DataHub** | Table- and column-level lineage from its Databricks source, with Genie space and notebook lineage when those are turned on |
| **OpenMetadata** | Table and column-level lineage from Databricks system tables, with permissions on table_lineage and column_lineage |

Several catalogs read the same system tables. A column-level lineage claim alone will not separate them. Unity Catalog lineage comes from queries run on Databricks and can include ML model versions, UDFs and external assets you register. For a third-party catalog, ask which other systems its graph connects on either side, and test it on a pipeline you know end to end.

[Data lineage for AI](https://atlan.com/know/ai-agent/data-for-ai/data-lineage-for-ai/) shows why that end-to-end path matters when an agent cites a number. [Automated SQL lineage vs manual lineage mapping](https://atlan.com/know/ai-agent/data-for-ai/automated-sql-lineage-vs-manual-lineage-mapping/) covers what parsing adds over hand-drawn maps, and [MCP for data lineage](https://atlan.com/know/mcp/mcp-for-data-lineage/) explains how agents query lineage directly.

---

## Can AI agents use Unity Catalog metadata through these catalogs?

Catalogs touch AI agents in two different ways: some deploy context into the agent's surface, and some catalog the agent itself. The native route comes first.

Genie One draws on Genie Ontology, the modeled knowledge from metric views, domains, Pages, wikis and certification, plus knowledge it learns from assets and usage. [Databricks Genie context requirements](https://atlan.com/know/ai-agent/databricks/databricks-genie-context-requirements/) lists what Genie needs to answer well, and [Genie One MCP](https://atlan.com/know/ai-agent/databricks/genie-one-mcp/) explains how agents reach Unity Catalog data over MCP.

### Third-party routes into Genie

Two third-party routes are documented:

* **Atlan** deploys a context repository from Context Engineering Studio into Databricks. It pushes descriptions into Unity Catalog, creates one metric view per table and configures a Genie Space backed by those views.
* **DataHub** ingests Genie spaces as AI agents with include_genie_spaces, including their text instructions and declared dataset lineage.

The two routes run in opposite directions. Atlan writes context into Databricks for Genie to use. DataHub reads Genie's configuration out into its own graph.

Runtime control is a separate layer. [Unity AI Gateway](https://atlan.com/know/ai-agent/databricks/unity-ai-gateway/) describes the Databricks side of agent governance. For the catalog side, [an MCP-connected data catalog](https://atlan.com/know/mcp-connected-data-catalog/) and [the agentic data catalog](https://atlan.com/know/agentic-data-catalog/) describe how a catalog serves agents directly, and [a data catalog for AI](https://atlan.com/know/data-catalog-for-ai/) lists what agents need from one.

An agent that only ever talks to Genie can live on Unity Catalog semantics. One that also answers from Snowflake, a CRM or a BI tool needs that context held somewhere both can read.

---

## How to choose the right catalog for your Databricks estate

Choose by the shape of your estate. Each row below names the options whose own documentation covers that situation; none of it is a score.

| If your estate... | Start with | Documented basis |
|---|---|---|
| Runs only on Databricks, with Genie as the main agent surface | Unity Catalog | Built-in discovery, lineage, metric views, domains and Pages |
| Spans Databricks and other platforms in one graph | A third-party catalog: compare each vendor's connectors for your other platforms | All five document a Unity Catalog lineage route |
| Needs edits made in the catalog to land in Unity Catalog | Atlan or Collibra | Atlan syncs tags and descriptions back; Collibra syncs attribute values to tags |
| Already runs Alation or Collibra | That vendor's Unity Catalog route | Alation's OCF connector; Collibra through Edge |
| Wants Genie spaces cataloged next to tables | DataHub | include_genie_spaces ingests spaces as AI agents |
| Wants governed context deployed into Genie | Unity Catalog semantics, or Atlan's Context Engineering Studio | Metric views and Pages natively; Atlan deploys metric views and a Genie Space |

### Before you pick a tool

Settle who owns each KPI and term before you pick a tool. [Context layer vs data catalog vs semantic layer](https://atlan.com/know/ai-agent/semantic-layer/context-layer-vs-data-catalog-vs-semantic-layer/) separates the three jobs, and [semantic layer vs business glossary](https://atlan.com/know/ai-agent/semantic-layer/semantic-layer-vs-business-glossary/) shows where a metric ends and a term begins. [How to manage KPI definitions and a business glossary in a data catalog](https://atlan.com/know/ai-agent/how-to-manage-kpi-definitions-and-a-business-glossary-in-a-data-catalog/) sets out the ownership model.

If you do add a layer, the [enterprise context layer](https://atlan.com/know/context-layer-enterprise-ai/) explains what it holds and [how to implement an enterprise context layer for AI](https://atlan.com/know/how-to-implement-enterprise-context-layer-for-ai/) gives the rollout steps.

Whatever you pick, decide in the trial who should see metadata for a restricted table, then test it: log in as a user without access to that table in Unity Catalog and check search, lineage and any API or agent tool in the other catalog. Before the trial ends, name an owner for each definition that lives outside Databricks.

[Get the AI Context Stack](https://atlan.com/resources/ai-context-stack/): a two-page overview of how context graphs, knowledge graphs, ontologies and semantic layers differ, and which one fits your stack.

---

## FAQs about data catalogs for Databricks

Short answers on Unity Catalog, third-party routes, lineage, Genie and Purview.

### 1. What is the best data catalog for Databricks?

Start with Unity Catalog, the governance layer built into Databricks and turned on by default for workspaces created after November 8, 2023. Add a third-party catalog when the estate reaches past Databricks. Alation, Atlan, Collibra, DataHub and OpenMetadata each document a Unity Catalog route, and the right one depends on your other platforms and the tools your teams already use.

### 2. Is Unity Catalog enough as a data catalog?

For a Databricks-only estate, often yes. Unity Catalog covers access control, discovery, lineage, auditing, data classification, quality monitoring, sharing and AI governance. Databricks notes that its discovery tools are optimized for data governed by Unity Catalog, so assets outside it may not appear in some of them. That boundary is where teams start to look at a second catalog.

### 3. Can a third-party catalog read Unity Catalog?

Yes. Each vendor documents its own route: Alation through an OCF connector, Atlan by crawling each metastore, Collibra through Edge, DataHub through its Databricks ingestion source and OpenMetadata through a dedicated Unity Catalog connector. Separately, Unity Catalog implements the Iceberg REST catalog API so engines such as Spark, Flink and Trino can read its tables.

### 4. Does a third-party catalog replace Unity Catalog?

No. Unity Catalog stays the layer that enforces access inside Databricks. Atlan's docs, for example, state that grants, row filters, column masks and audit logs remain in Unity Catalog while Atlan reads and enriches the metadata. Databricks also documents catalog federation for teams that keep a long-term hybrid model.

### 5. How does lineage differ between Unity Catalog and a third-party catalog?

Unity Catalog captures lineage for queries run on Databricks, down to the column level. Several third-party catalogs read that same lineage from Databricks system tables: Atlan and Alation name system.access.table_lineage and system.access.column_lineage. The difference is what each one connects it to, such as DDL statements in Alation or other platforms in Atlan's graph.

### 6. Can a data catalog work with Databricks Genie?

Yes, in two documented ways. Atlan's Context Engineering Studio deploys context to Genie by pushing descriptions into Unity Catalog, creating one metric view per table and configuring a Genie Space. DataHub can ingest Genie spaces as AI agents when include_genie_spaces is turned on.

### 7. What is the difference between Unity Catalog and the open-source Unity Catalog?

Unity Catalog in Databricks is the governance layer built into the platform. The open-source Unity Catalog is a separate implementation released under the Apache 2.0 license as an LF AI and Data project, with its own Iceberg REST catalog API.

### 8. How does Microsoft Purview scan Unity Catalog?

Microsoft Learn documents a Unity Catalog scan for Microsoft Purview. It covers the metastore, catalogs, schemas, tables, views and Unity Catalog tags, plus lineage captured during notebook runs. The same page notes that Purview Unified Catalog and classic data governance capabilities are being retired, with maintenance mode through November 12, 2027.

---

## Sources

1. What is Unity Catalog?, Databricks documentation (docs.databricks.com), last updated September 11, 2026, accessed November 17, 2026

2. Discover data, Databricks documentation (docs.databricks.com), last updated September 11, 2026, accessed November 17, 2026

3. View data lineage using Unity Catalog, Databricks documentation (docs.databricks.com), last updated November 4, 2026, accessed November 17, 2026

4. Unity Catalog metric views, Databricks documentation (docs.databricks.com), last updated October 29, 2026, accessed November 17, 2026

5. [Unity Catalog semantics, Azure Databricks documentation, Microsoft Learn](https://learn.microsoft.com/en-us/azure/databricks/uc-semantics/)

6. Access Databricks tables from Apache Iceberg clients, Databricks documentation (docs.databricks.com), last updated November 13, 2026, accessed November 17, 2026

7. Integrations with Unity Catalog, Databricks documentation (docs.databricks.com), last updated September 11, 2026, accessed November 17, 2026

8. What is query federation?, Databricks documentation (docs.databricks.com), last updated November 12, 2026, accessed November 17, 2026

9. [Unity Catalog open source project, unitycatalog.io](https://www.unitycatalog.io/)

10. Databricks Unity Catalog OCF Connector overview, Alation documentation (docs.alation.com), accessed November 17, 2026

11. Databricks Unity Catalog OCF lineage extraction, Alation documentation (docs.alation.com), accessed November 17, 2026

12. [Databricks connector, Atlan docs](https://docs.atlan.com/apps/connectors/data-warehouses/databricks)

13. [What does Atlan crawl from Databricks, Atlan docs](https://docs.atlan.com/apps/connectors/data-warehouses/databricks/references/what-does-atlan-crawl-from-databricks)

14. [Atlan vs Databricks lineage, Atlan docs](https://docs.atlan.com/apps/connectors/data-warehouses/databricks/faq/atlan-vs-databricks-lineage)

15. [Unity Catalog governance and connections, Atlan docs](https://docs.atlan.com/apps/connectors/data-warehouses/databricks/faq/unity-catalog-governance)

16. [Deploy to Databricks Genie, Atlan docs](https://docs.atlan.com/product/capabilities/governance/context-engineering-studio/how-tos/deploy-databricks)

17. About the Databricks integration, Collibra documentation (productresources.collibra.com), accessed November 17, 2026

18. Databricks ingestion source, DataHub documentation (docs.datahub.com), accessed November 17, 2026

19. Unity Catalog connector, OpenMetadata documentation (docs.open-metadata.org), accessed November 17, 2026

20. [Connect to and manage Azure Databricks Unity Catalog in Microsoft Purview, Microsoft Learn](https://learn.microsoft.com/en-us/purview/register-scan-azure-databricks-unity-catalog)

## Next steps

- To see Atlan's context layer in action first, watch a recorded demo: [Atlan context layer demos](https://atlan.com/context-layer-demo/)
- To see Atlan working on your own data stack, book a demo: [Book a demo with Atlan](https://atlan.com/forms/talk-to-sales-contact/)
