Data Fabric is now generally available​

Entity types: Native and Federated​

• Native: the data is stored in UiPath. A Native entity supports create, read, update, and delete operations, and holds only its own fields.
• Federated: a read-only view that reads its data from its sources each time it is queried. A Federated entity stores nothing of its own.

• Entities that combine locally stored fields and external fields in a single entity can no longer be created. To model the same result, keep the locally stored data in a native Data Fabric entity and reference that entity as a source of the Federated entity.
• Existing entities that combine local and external fields continue to work without disruption. They can still be consumed in Apps, automations, Maestro, and the APIs, and their locally stored fields remain editable. Their external fields remain read-only, as before.
• Existing entities are classified automatically. An entity with no external fields is Native, and an entity whose fields all come from external systems is Federated. No action is required, and no existing entity changes behavior.
• Record-level role-based access control and Insights ingestion do not apply to Federated entities, and are no longer offered for them. The records of a Federated entity are held by its source systems, so there is no stored record to secure or to ingest.

Reference a native Data Fabric entity as a source of a Federated entity​

• Choice Set and File are available as destination field types, in addition to the basic types. A choice set is resolved end to end, so the Federated entity returns the choice display value rather than its identifier.
• Search and filter settings are preserved for each field, so a referenced field remains filterable and sortable.
• A Federated entity that references a native Data Fabric entity can be exported and installed through Automation Solutions.

Predictable limits on queries over external data​

• A filter is required on a query over an entity that has external fields. The requirement does not apply to queries where every entity involved is stored entirely in UiPath, to browsing data in the Data Fabric interface, or to a sample read of 200 records or fewer. A join condition does not count as a filter.
• Data warehouse sources return a single page. Redshift, Snowflake, Databricks, and warehouse databases reached through Database Hub do not enforce primary keys, and page through results in an order the warehouse does not guarantee, so a second page could repeat records or skip them. A request for a second page is refused. A filter that narrows the result to fit in a single page avoids the limit.
• Result size is limited. A query whose intermediate result exceeds the limit returns an error asking you to narrow it.

Clearer errors on queries over external data​

• A query that cannot be parsed or validated names the clause at fault.
• A missing permission on the connection, or a missing scope on an external application, is reported as such.
• A rejection by the source system is returned with the status and message of the source system, so the reason comes from the system that refused the request.
• A query that exceeds its time limit, a source system that is unavailable, and a deleted or disconnected Integration Service connection are each reported distinctly.

Faster queries over external data​

• Filters and sorting are applied by the source system wherever it supports them, so a query returns the records you asked for instead of retrieving a larger set and narrowing it afterwards.
• A request for one page of results reads only as far into each source as that page needs, instead of retrieving the full result set first.
• When a query draws on more than one system, those systems are read at the same time rather than one after another.

Data Fabric entities can now be used as agent context​

Folder-level entities and Solutions integration are now generally available​

Supported surfaces​

• RPA (Studio Web and Studio Desktop 26.10 LTS)
• Maestro BPMN, Maestro Case, Maestro Flow (via native entity nodes)
• API Workflows
• Coded Apps
• Agents (Preview)

Scope selectors in Data Service activities​

• Tenant — Shows entities available across the entire tenant.
• Folder — Shows entities across all Orchestrator folders where you have access.

Query Multiple Entity Records is now generally available​

Schema refresh in Studio Web​

Folder-scoped choice sets in Solution Builder​

Improved Solutions validation​

• Install and upgrade — Validation identifies specific conflicts — such as an entity name collision with a different schema, or a choice set conflict — before deployment begins.
• Upgrade with existing data — A merge offering lets you keep existing data while the schema is updated to match the new version.
• Uninstall — Entities with existing data are preserved to prevent accidental data loss. Only empty entities are removed.

Schema export and import across scopes​

Data Fabric UI improvements​

• Entity filtering — You can now filter the entity list by folder and by entity type, making it easier to find specific entities in tenants with large entity catalogs.
• Choice set NumberId visibility — Choice set values now display their NumberId in the Data Fabric UI, so you can verify you are selecting the correct values in workflows and identify errors caused by mismatches.
• OpenAPI spec — The downloadable OpenAPI specification now includes folder-scoped entities, in addition to tenant-scoped entities.

Previously resolved limitations​

• Studio Desktop is now supported for folder-level entities in a Solutions context (Studio Desktop 26.10 LTS).
• Folder-scoped choice sets are now included in Solution Builder packaging.
• Schema refresh no longer requires reselecting the entity from the dropdown.
• Maestro, API Workflows, and Agents now support folder-scoped entities via scope selectors.
• Federated entities and entities with external source fields are now supported in Solutions packaging.

Model Data Fabric entities inline in VS Code (Preview)​