Poor data quality costs organizations at least $12.9 million a year on average, according to Gartner . Most teams trace it to the same root cause, which is analysts building on the wrong data because they had no reliable way to find the right data.
A financial analyst joins a new team, opens Microsoft Fabric, and sees 340 tables across 12 workspaces with no descriptions, no labels, and no sign of which are production-ready. Three hours later they’re on Slack asking engineers. The certified semantic model they needed was there the whole time.
That is the problem the Microsoft Fabric OneLake catalog is designed to solve. In this article, we’ll cover its core features, setup guide, Purview comparison, industry use cases, and migration best practices.
Key Takeaways OneLake catalog replaced the legacy OneLake Data Hub. It is a full governance platform with domains, quality scores, lineage, endorsement, and sensitivity labels, not just a search tool. Data quality scores are the most underused feature. Organizations that configure DQ rules first see quick gains in report reliability and analyst trust. OneLake catalog and Microsoft Purview are complementary, not competing. OneLake catalog is Fabric-scoped; Purview covers enterprise-wide, multi-cloud assets. Purview labels surface automatically inside the catalog. External catalog federation (connecting to Databricks Unity Catalog and AWS Glue) is in preview and positions OneLake catalog as a multi-cloud discovery layer. A basic catalog setup takes 1–2 weeks. A full enterprise implementation takes 6–8 weeks. The difference between those two is the difference between a governance checkbox and actual governance. Catalog adoption directly cuts report rebuild rates. When analysts build on certified, quality-scored datasets, errors fall sharply.
Looking for Help With Your Microsoft Fabric OneLake Catalog Setup? Get advice on catalog setup, governance configuration, and enterprise best practices.
Contact Us
What Is the Microsoft Fabric OneLake Catalog? The Microsoft Fabric OneLake catalog is a centralized governance and discovery layer built on top of OneLake , Fabric’s unified storage layer. Every user in a tenant gets one place to find, endorse, and govern data assets across all workspaces. The catalog is also embedded in Microsoft Teams, Microsoft Excel, and Microsoft Copilot Studio, so discovery happens where people already work.
It replaced the legacy OneLake Data Hub, and the difference is substantial. The Data Hub was a search tool scoped to Power BI datasets, with basic endorsement and a single flat view. The catalog covers every Fabric item type across three tabs (Explore, Govern, and Secure). It adds domains, data quality scores, sensitivity labels, end-to-end lineage, Copilot-powered discovery, and external catalog federation. Purview governance signals surface inside it natively. It’s a fundamentally different kind of tool.
Core Features of the Microsoft Fabric OneLake Catalog The catalog is built around ten capabilities that work together to make data discoverable, trustworthy, and governed. Each one addresses a specific gap that exists without a catalog layer in place.
1. Unified Data Asset Discovery Across 9 Fabric Item Types The catalog covers lakehouses, warehouses, SQL analytics endpoints, semantic models, KQL databases, pipelines, notebooks, dataflows, and reports, all in one place. Users filter by workspace, domain, item type, endorsement status, sensitivity label, and data quality score. Access is role-scoped, so people only see items they’re authorized to view.
2. The Govern Tab: Governance Posture Across 4 Health Indicators The Govern tab shows the health of everything you own in Fabric. It covers curation completeness, endorsement coverage, sensitivity label gaps, and recommended actions to close them. Fabric admins get a tenant-wide view of item counts, workspace counts, capacities, and domains. Note that insights refresh daily, not in real time, so factor that into same-day monitoring workflows.
3. The Secure Tab: Centralized Management Across All Workspace and OneLake Roles The Secure tab gives admins a single place to audit permissions, review user access, and manage security roles. It covers both workspace and OneLake security scopes. Most teams use it to catch permission drift before it becomes a compliance issue.
4. Business Domains: Organize the Tenant Into Logical Data Areas Domains group workspaces by business area, Finance, HR, Sales, Operations, Supply Chain, so analysts can filter straight to relevant assets without asking anyone. Each domain has its own admin, enabling federated governance without routing everything through central IT. Without domains, the catalog becomes a flat list that doesn’t scale past a handful of workspaces.
5. Catalog Endorsement: 3 Trust Levels for Every Item Endorsement assigns every catalog item one of three trust levels. Promoted means the workspace owner recommends it. Certified means formal sign-off against production standards. Master data marks the authoritative source of record. Data consumers filter to Certified items and see only the validated ones, with a full certified-by trail for accountability.
Only users with Certifier permissions, assigned by Fabric admins in tenant settings , can certify items. Workspace owners can promote. Set this hierarchy at configuration time to prevent endorsement from becoming a free-for-all.
6. Data Quality Scoring: 5 Rule Dimensions, 0–100% Scores per Item Data quality scores (0–100%) are generated by rules that check completeness, uniqueness, accuracy, consistency, and validity. Scores appear on every catalog item card, so analysts filter to items above a threshold before building a report rather than guessing. Despite this, scores are configured in fewer than 30% of Fabric deployments. Prioritize by consumption volume and business risk to make the effort manageable.
SKU note: Data quality scoring requires Fabric capacity . Check the licensing page for full SKU details.
7. Sensitivity Labels: 4-Tier Classification on Every Item Card Sensitivity labels from Microsoft Purview flow automatically into the catalog and appear on every item card before access is attempted. The four tiers are Public, General, Confidential, and Highly Confidential. Governance teams can filter to all Highly Confidential items tenant-wide in one view, which is essential for GDPR, HIPAA, and SOX workflows. Enable label inheritance so items created from a labeled source carry that label forward automatically.
8. Data Lineage: Full Upstream and Downstream Traceability The catalog traces the full path from data source through pipeline to lakehouse to semantic model to report. Teams use it for impact analysis before schema changes, audit traceability for compliance, and debugging data quality issues upstream. Lineage auto-generates for Fabric-native items. External sources like on-prem SQL Server and Salesforce need a Purview scanner configured, or they appear as dead ends in the lineage chain.
9. Copilot-Powered Discovery: 2 AI Capabilities Built Into the Catalog Copilot adds two capabilities to the catalog. The first is AI-generated descriptions and the second is natural language search. Descriptions are auto-generated from schema and metadata; stewards review before publishing. Search lets users query in plain English: “find all certified customer tables in the Finance domain.” Both are available on all paid Fabric SKUs with Copilot enabled. Treat generated descriptions as a first draft, since complex schemas with abbreviated column names frequently need steward correction.
10. External Catalog Federation: Connect to 3 External Platforms (Preview) External catalog federation connects OneLake catalog to Databricks Unity Catalog , AWS Glue Data Catalog , and Google Data Catalog, surfacing assets from those platforms alongside Fabric items in a single view. This feature is in preview as of early 2026. Don’t architect production multi-cloud governance strategies around it yet, but when it reaches GA, it turns OneLake catalog from a Fabric-only tool into a genuine enterprise-wide data catalog.
Microsoft Fabric Implementation and Governance End-to-end Fabric delivery from tenant setup to catalog governance and Purview integration.
Explore the Services
How to Set up the Microsoft Fabric OneLake Catalog: Step-by-Step Guide Each step builds on the last. Skipping ahead creates rework.
Step 1: Enable OneLake Catalog in Fabric Tenant Admin Settings Navigate to the Fabric Admin Portal under Tenant Settings → Governance and insights settings. Settings to configure:
OneLake catalog: toggle On (tenant-wide)Users can discover data using Microsoft Fabric: enable for all users or specific security groupsCertification: define who can certify items. Do this before any endorsement activity starts. Assign to a named security group, not individual users, so the permission set stays maintainableExternal data sharing: configure whether catalog users can see externally shared itemsSensitivity label policies: confirm Purview Information Protection is connected and labels are visible in Fabric items
Admin Portal path: app.fabric.microsoft.com → Settings gear → Admin portal → Tenant settings
One thing to set correctly from the start is role-scoped access. Fabric Admins see everything tenant-wide. Domain Admins see all items in their domain. Workspace Admins, Members, and Contributors manage items in their workspaces. Viewers get read-only access. Users with no workspace role cannot see an item in the catalog at all. If someone says they can’t find a dataset, it almost always comes down to workspace membership , not catalog configuration.
Step 2: Create and Configure Business Domains Navigate to: Admin Portal → Domains
Select Create domain Name the domain (e.g., “Finance”, “Operations”, “Supply Chain”) Assign domain admins, these users manage all workspace assignments and governance settings within the domain Assign workspaces to the domain, each workspace can belong to only one domain Set default sensitivity labels for the domain if desired
Start with five to seven domains mapped to how data is consumed, not how the org chart is structured. Finance and Accounting might be one org unit but two distinct data consumer communities. Subdivide later as usage patterns emerge.
Step 3: Configure the Endorsement Workflow and Certification Criteria Before enabling endorsement, define what “Certified” means in your organization. Write it down. This matters more than the permission configuration.
Navigate to: Admin Portal → Tenant Settings → Endorsement and discovery
Enable Certified endorsement Assign Certifier permissions to a security group Communicate the certification criteria to workspace owners
Then, in individual workspaces, workspace owners can begin promoting items. Certifiers can certify items that meet the defined standard via the item settings panel in the catalog.
Step 4: Configure Data Quality Rules and Scoring Navigate to: Fabric workspace → [Select a lakehouse or warehouse] → Data quality tab
Select the table to score Choose a quality dimension: Completeness, Uniqueness, Accuracy, Consistency, or Validity Define the rule logic (e.g., column X must be non-null in 99% or more of rows) Set threshold: score above X% means Pass Schedule rule execution Scores appear on the catalog item card after first run
Start with your top 20 most-consumed items. A catalog with quality scores on 20 items is more useful than one with empty scores on 200. Prioritize production semantic models, certified lakehouses, and financial reporting tables.
Step 5: Verify Sensitivity Label Flow and Information Protection Confirm Microsoft Purview Information Protection is configured at the Microsoft 365 tenant level In Fabric Admin Portal, verify Information protection settings are enabled Open the OneLake catalog and spot-check 10–15 items for label presence Identify label gaps, particularly tables containing PII, financial data, or health information Validate that label inheritance is working: create a derived item from a labeled source and confirm the label propagates
Under GDPR , knowing where personal data resides is a foundational requirement. Sensitivity label filtering in the catalog is exactly how you meet it.
Step 6: Enable Copilot and Drive Catalog Adoption Across the Organization Confirm Copilot is enabled in Fabric Admin Portal under Tenant Settings → Copilot and Azure OpenAI Service In the catalog, select high-traffic items → Edit description → Generate with Copilot Data stewards review, edit, and approve generated descriptions Run an analyst onboarding session: show how to filter to Certified and quality-scored assets Set “start in the catalog” as the team standard for every new report development task
Track the ratio of reports built on certified vs. uncertified datasets. That number tells the real governance story more than item counts ever will. That ratio tells the real governance story, not the number of items in the catalog.
Phase Step Timeframe Primary Owner Success Signal 1 Tenant admin configuration Days 1–2 Fabric Admin All settings toggled; certifier security group assigned 2 Domain creation and workspace assignment Week 1 Fabric Admin + Domain Admins All production workspaces assigned to a domain 3 Endorsement workflow and certification criteria Weeks 1–2 Governance Lead Written certification criteria published; first items promoted 4 Data quality rule configuration Weeks 2–4 Data Engineering + Stewards Top 20 items scoring; scores visible on catalog cards 5 Sensitivity label verification Weeks 3–4 Governance / Compliance Zero unlabeled PII items; inheritance validated 6 Copilot enablement and adoption program Weeks 4–6 Governance Lead + Business Leads Catalog search is the default start point for report development
OneLake Catalog Vs. Microsoft Purview Data Catalog A common governance question is whether teams that already have Purview still need the Microsoft Fabric OneLake catalog.
They are not the same thing. They are designed to work together. Microsoft Purview is an enterprise-wide governance platform covering Azure, AWS, GCP, on-prem, SaaS, and databases. The OneLake catalog is the Fabric-native discovery and governance experience that surfaces Purview’s signals at the consumption layer. Governance teams, CDOs, and compliance officers managing policy across an entire data estate use Purview. Data consumers, analysts, and engineers working inside Fabric every day use the OneLake catalog. Setup complexity reflects this difference. The catalog is built into Fabric and requires minimal configuration. Purview requires Data Map scanning setup and is a heavier investment.
A simple way to decide which to use is based on your data estate scope and governance maturity.
Fabric-first, early governance maturity: Start with OneLake catalog only. Build discipline before adding Purview complexity.Mid-enterprise, Microsoft stack (Fabric + Azure SQL, ADF, some on-prem): Add Purview Information Protection labels early, but defer the full Purview Data Map until you have the adoption baseline.Large enterprise, multi-cloud (AWS S3, Salesforce, Oracle, SAP): Run both. Purview as the governance backbone, OneLake catalog as the Fabric-layer experience.Regulated industry (HIPAA, SOX, GDPR): Both are required regardless of estate complexity.Migrating from a legacy environment: Pair OneLake catalog with FLIP migration accelerators to populate catalog metadata during migration, not after.
Purview also covers capabilities the OneLake catalog doesn’t handle natively. That includes a business glossary with term-to-asset assignment, automated sensitive data classification at scale, policy-based access control, governance of non-Fabric sources like Oracle, SAP, and Salesforce, and data stewardship workflows. For organizations in regulated industries, these aren’t optional. They’re where the compliance story gets built.
How OneLake Catalog Governs OneLake Shortcuts OneLake Shortcuts are a frequently overlooked interaction point with the catalog. Shortcuts created from external sources like ADLS Gen2, Amazon S3, or Google Cloud Storage appear as catalog items inside the lakehouse that hosts them. The governance behavior varies by signal type, as shown below.
Governance Signal Fabric-to-Fabric Shortcut ADLS Gen2 / Azure Storage Amazon S3 / GCS Action Required if Gap Exists Sensitivity label Inherited from parent lakehouse (if inheritance on) Inherited from parent lakehouse (if inheritance on) Inherited from parent lakehouse (if inheritance on) Enable label inheritance in tenant admin settings Data quality scoring Not inherited: must configure rules on shortcut target Not inherited: must configure rules on shortcut target Not inherited: must configure rules on shortcut target Configure DQ rules explicitly per shortcut item Catalog visibility Appears as item in the lakehouse hosting the shortcut Appears as item in the lakehouse hosting the shortcut Appears as item in the lakehouse hosting the shortcut No action needed: automatic Lineage (upstream) Auto-generated within Fabric Breaks at shortcut boundary without Purview Breaks at shortcut boundary without Purview Configure Purview scanner for external source Endorsement Must be set manually on the lakehouse Must be set manually on the lakehouse Must be set manually on the lakehouse Include in endorsement workflow for the workspace
OneLake Catalog in Action: 5 Industry Use Cases 1. Financial Services: Certified Regulatory Datasets and SOX Audit Trails Financial reporting teams spend significant time finding the right dataset and then proving it’s traceable. The catalog solves both. End-to-end lineage traces every number in a close report back to its source system automatically. Endorsement ensures analysts only build on formally validated production data. Sensitivity labels on all PII and account data block unauthorized discovery before access is even attempted.
In a recent bank engagement, KANComply mapped existing data assets to SOX control requirements inside the Purview governance layer, cutting audit preparation time by eliminating the manual evidence collection step. See the outcome in the sales financials KPIs case study .
2. Healthcare: PHI Protection and FDA-Ready Data Lineage Clinical data teams need to know which datasets are safe to use for model training and be able to prove the data chain for regulatory validation. The catalog makes both manageable. All PHI-containing items are labeled Highly Confidential. Data scientists filter to Certified datasets before training, removing the risk of using uncertified or synthetic data by accident. Catalog lineage produces a traceable data flow from source to analytical output, supporting FDA 21 CFR Part 11 documentation.
KANComply maps HIPAA safeguards directly to Fabric items in the catalog, so compliance officers can see which regulatory controls each asset is mapped to, not just that it exists.
3. Manufacturing and Supply Chain: Domain-Organized Sensor Data and AI Agent Governance Factories generate thousands of data streams across plants, lines, and sensors. Without a catalog layer, discovery becomes a spreadsheet problem. Domain structure organizes those streams by plant and production line, giving operations analysts a view that reflects how they think rather than how pipelines are built. AI agents like KARL that operate on production data require certified, quality-scored catalog assets as a prerequisite. When catalog governance breaks down, agent output breaks down with it.
Two native Fabric workloads debuted at FabCon 2026 on this exact premise. The approach held up in practice for a global packaging manufacturer and for a logistics client .
4. Insurance: Single Source of Truth for Claims and Underwriting Data Insurance teams deal with overlapping data sources across claims, underwriting, and actuarial functions. Without endorsement, analysts can’t tell which version of a dataset is authoritative. Master data endorsement in the catalog designates the source of record for each entity. Lineage then confirms everything downstream traces back to it. The result is a a single source of truth for claims and underwriting that satisfies state insurance regulation requirements.
5. Retail: Real-Time Inventory and Sales Analytics on Certified Data Retail analytics breaks down when inventory and sales data isn’t trusted, and the catalog’s DQ scores on certified semantic models give analysts a clear signal before they build. When teams migrate using FLIP, metadata populates the catalog during the migration itself, so they arrive at a governed catalog from day one rather than spending months backfilling documentation afterward. This played out directly for a retail and material handling client . For teams migrating from fragmented reporting environments, catalog-first delivery .
Why Most OneLake Catalog Deployments Stall and How to Fix Them The Three Mistakes That Kill Catalog Value Treating the catalog as a feature, not a governance program : Organizations enable OneLake catalog in the admin portal and consider it done. A proper Microsoft Fabric governance program is what separates configuration from value. But catalog value comes from coverage, endorsement discipline, and user adoption, not from the toggle being on. A common pattern is catalogs where 80% of assets have no endorsement, no quality scores, no descriptions. Technically configured; practically useless.
Building governance structure before there is data to govern : Some teams spend weeks designing domain hierarchies before any Fabric workloads are in production. Governance structure built for a dataset inventory that doesn’t exist yet ends up redesigned once real usage patterns emerge. Get two or three high-value workloads into production and certified first. Let governance structure follow real consumption patterns.
Skipping the data quality foundation : Without DQ rule configuration, every item in the catalog looks equal. Consumers cannot distinguish reliable from unreliable data and default to asking engineers, which is exactly the behavior the catalog is supposed to eliminate.
The Five Stages of Catalog Maturity Most implementations follow a predictable progression. Knowing where you are helps you know what to fix next.
Level 1: Configured (Days 1–5): Catalog enabled, domains created, some items visible. No endorsement, no DQ scores, no labels, no adoption yet.Level 2: Structured (Weeks 1–3): Workspaces assigned to domains, endorsement permissions set, top items promoted. DQ rules still missing; analysts still asking engineers.Level 3: Trusted (Weeks 4–8): Certified items with DQ scores of 90%+; sensitivity labels on all PII items; lineage visible for production pipelines.Level 4: Operationally Embedded (90+ days): Catalog is the default start for all report development; DQ rules on all Tier 1 items; Purview integration complete.Level 5: Governed at Scale (6+ months): Business glossary terms mapped via Purview; external catalog federation active; governance KPIs tracked monthly.
When analysts build on trusted, quality-scored data, they rebuild reports less often. As Amit Chandak, Kanerika’s Chief Analytics Officer and Microsoft MVP in Power BI, puts it: “The catalog doesn’t govern your data. Your team does. The catalog makes it possible for your team to govern at scale without it becoming a full-time job.”
What Happens to Your Metadata When You Migrate to Microsoft Fabric Existing catalog metadata doesn’t migrate automatically. The OneLake catalog starts from scratch, and for most organizations that is an opportunity. Legacy environments tend to accumulate years of inconsistent metadata, outdated endorsements, and governance debt. A Fabric migration is a natural forcing function to rebuild catalog governance with better structure than what came before.
The key is running a catalog population sprint in parallel with the technical migration rather than treating it as a post-migration activity that perpetually gets deferred. FLIP migration accelerators automate 50–60% of the technical migration work across 12 supported paths and generate structured metadata as part of the output. What that means in practice:
Item descriptions: generated from source metadata during migration, not written manually afterLineage connections: built into the FLIP artifact set for SSIS , Azure Data Factory , Informatica , Alteryx , SSAS, and Oracle pathsWorkspace organization: domain and workspace structure recommendations generated alongside the migration deliverablesEndorsement and DQ rules: not generated by FLIP, these require manual configuration post-migration, but with the metadata foundation already in placeThe practical difference is arriving at a governed catalog at migration completion versus spending three months post-migration trying to document what was just built.
What Transfers Automatically and What You Need to Configure Metadata Type Migrates Automatically? FLIP Accelerator Output Manual Action Required Table and column names Yes: from source schema Preserved during pipeline migration None Item descriptions / documentation No FLIP generates descriptions from source metadata where available Steward review and enrichment Lineage connections (Fabric-native) Yes: auto-generated for Fabric items post-migration Lineage connections generated in FLIP artifact set Verify completeness post-migration Lineage (external source connections) No Partial: FLIP documents source connections Purview scanner configuration for external sources Endorsement status No: starts at zero Not generated by FLIP Must re-certify migrated items per new certification criteria Data quality scores No: rules do not transfer Not generated by FLIP DQ rules must be configured post-migration (use DQ prioritization framework) Sensitivity labels No: labels do not transfer Not generated by FLIP Re-apply via Purview label policies; enable inheritance Domain assignment No Workspace organization recommendations in FLIP output Assign workspaces to domains per domain architecture
OneLake Catalog Limitations to Know Before Implementation External catalog federation is still maturing: Unity Catalog and AWS Glue connections are in preview as of early 2026. Not yet suitable for production governance architectures.Private Link incompatibility: OneLake catalog is not compatible with Private Link configurations. Tenants using Private Link for network isolation cannot access the catalog, this is a known limitation to assess before implementation in high-security environments.Lineage gaps for third-party sources: Lineage auto-generation works for Fabric-native items. Pipelines pulling from non-Fabric sources without Purview scanning will have broken lineage chains.Data quality rules require real investment: Quality scores don’t appear automatically, someone has to define the rules. For organizations with hundreds of tables, this is a material time commitment.No native business glossary: The catalog doesn’t include a business glossary with term-to-asset assignment. That capability requires Microsoft Purview.Limited access request workflow: The catalog surfaces assets but doesn’t handle access request workflows or data product packaging natively. Those require Purview or custom workflow additions.Copilot descriptions need steward review: AI-generated descriptions can be inaccurate for complex or poorly named schemas.
See How Kanerika Governs Fabric at Enterprise Scale Get a walkthrough of catalog setup, KAN suite deployment, and FLIP migration in action.
Schedule a Demo
How Kanerika Implements OneLake Catalog for Enterprise Clients Kanerika is a Microsoft Solutions Partner and one of the earliest Microsoft Purview implementors globally, with 100+ enterprise clients across financial services, healthcare, insurance, manufacturing, and logistics. Catalog and governance configuration is a core part of every Fabric delivery. In a recent banking engagement, KANComply mapped data assets to SOX controls , eliminating the manual evidence collection step from audit preparation entirely. The full delivery stack includes:
KANGovern: defines and enforces data governance policy across the Fabric estate, including domain assignment rules, endorsement workflows, and certification criteria that apply consistently across all workspacesKANComply: maps catalog assets to regulatory frameworks including GDPR, HIPAA, and SOX, giving compliance officers a direct line-of-sight from each data item to the specific controls it satisfiesKANGuard: monitors for policy violations and unauthorized access in near-real time, alerting data stewards when certified items are modified, relabelled, or accessed outside approved rolesFLIP migration accelerators: automate 50–60% of technical migration effort and generate structured catalog metadata in the process, so teams arrive at a governed catalog at completion rather than months after
A top North American healthcare organization managing electronic health records, medical imaging, and administrative databases across a network of hospitals, clinics, and specialty centres needed to bring fragmented, ungoverned data under control while meeting strict regulatory requirements.
Challenges: Data spread across Azure Blob Storage, SQL databases, and SaaS applications with no unified discovery layer No consistent framework for data classification or metadata management, leading to poor data quality and limited visibility Low stakeholder understanding of governance practices, blocking effective data integration and management
Solutions: Built a centralized data catalogue using Microsoft Purview with full metadata and data classification coverage Designed a classification framework covering steward roles, usage policies, and data handling standards Integrated Power BI dashboards for business insights and delivered user guides and process documentation for ongoing governance
Results: 57% reduction in data discovery time 35% increase in data accuracy 70% improvement in data accessibility 90% increase in compliance adherence
Wrapping Up The Microsoft Fabric OneLake catalog is one of the platform’s most underused capabilities. Most organizations configure it and move on. The ones getting real governance value treat it as a program, not a feature, defining quality standards, training data consumers, enforcing endorsement discipline, and connecting it to Purview governance. The difference between a basic setup and one that works shows up at 90 days, when analysts either self-serve or are still on Slack asking engineers.
Ready to Build a Governed Fabric Environment? Talk to Kanerika’s team about your catalog and governance requirements.
Book a Meeting
FAQs What Is the OneLake Catalog in Microsoft Fabric? The Microsoft Fabric OneLake catalog is the platform’s built-in governance and discovery layer, covering all item types across every workspace in your tenant. Unlike the legacy OneLake Data Hub it replaced, it goes beyond search to include endorsement workflows, data quality scoring, sensitivity labels, end-to-end lineage, and Copilot-powered descriptions. The official Microsoft Learn overview has the complete feature reference.
Is OneLake Catalog the Same As Microsoft Purview Data Catalog? No. The OneLake catalog is scoped to Microsoft Fabric items and built natively into the Fabric experience. Microsoft Purview covers enterprise-wide, multi-cloud assets and adds business glossary, automated classification, and access policy enforcement. They complement each other; Purview’s sensitivity labels and governance signals surface automatically inside the OneLake catalog.
Does the OneLake Catalog Cost Extra? The OneLake catalog is included in Microsoft Fabric capacity licensing, no separate catalog license required. Data quality scoring requires Fabric F-SKU capacity. Copilot descriptions require Copilot to be enabled. Basic discovery and endorsement are available at no additional cost for any Fabric tenant. Full details are on the licensing page.
What Are the Endorsement Levels in the OneLake Catalog? There are three levels: Promoted, Certified, and Master data. Certified is the most important for data consumers; it means the item has formal sign-off and meets production standards. The full permission model for endorsement is documented on Microsoft Learn.
How Do Data Quality Scores Work in the OneLake Catalog? Data quality scores (0–100%) are generated by rules you configure in Fabric’s Data Quality feature. Rules check completeness, uniqueness, accuracy, and consistency against defined thresholds. Scores appear on catalog item cards after the first rule run.
Can the OneLake Catalog Connect to Databricks Unity Catalog or AWS Glue? Yes, external catalog federation enables connections to Databricks Unity Catalog, AWS Glue, and other external platforms. This feature is in preview as of early 2026. Production multi-cloud governance architectures should not depend on it until it reaches full GA status.
How Do OneLake Shortcuts Interact With the Catalog? OneLake Shortcuts appear as catalog items within the lakehouse where they are created. Sensitivity labels inherit from the parent lakehouse when label inheritance is enabled. Data quality rules must be configured explicitly against shortcut targets, they don’t inherit automatically. Lineage for external shortcuts requires Purview scanner configuration to show the complete chain.
Does the OneLake Catalog Have a Business Glossary? No. Business glossary with term-to-asset assignment is not native to the OneLake catalog. That capability is provided by Microsoft Purview, which integrates with the catalog so governed terms can surface alongside catalog items.
