Skip to main content

Publish, Share, and Use Flow Outputs

Publish a version​

Choose Validate output and review the resulting schema and checks. Then choose Publish version to save a fixed definition, including its output and compiled query. Further draft edits are released as another version. Choose the version deliberately before sharing, attaching it, or configuring a write.

Use the published output controls for the kind of result you need:

OutputBehavior
Live datasetRuns the query for a particular published version when consumed.
Destination mappingWrites to an existing target table in an explicit mapping run.
Stored outputWhen enabled, builds and checks a physical result for a published version.

Use a flow in a dashboard​

  1. Open the flow and select the version in Published output. Choose Add to live dataset library if the version is new to the library.
  2. Use Add to a dashboard, choose a dashboard you own that uses viewer permissions, and attach the version. You can also find registered versions under Live datasets in Data Flows.
  3. In the dashboard editor, select the attached dataset for a widget. Map its fields and preview the result. Publish the dashboard when it is ready for viewers.
  4. To adopt a newer flow version, add that version to the library, attach it, and update the widget binding explicitly.

The dashboard binds to a specific published version. The query runs with the viewer's Engine permissions. Confirm that the viewer can open the dashboard, use the published flow version and connection, and read the source data.

Share a published version​

In the published version’s sharing controls, grant Consumer access for inspection and preview or Operator access when the recipient also needs to run a destination mapping. Both roles use the recipient's own Engine permissions. Check the recipient's access with the chosen version and target before relying on a shared workflow.

Write to a destination table​

Where destination mappings are available, select the target catalog, schema, and table and map the output fields. Review the schema and run the mapping preview before writing. Choose Append to add rows, Upsert to match existing rows using explicit keys, or Replace for a full supported target replacement. Check upsert keys for missing or duplicate values and confirm the target before running.

Run mapping writes to the target under your Engine permissions. Inspect the run record and read back expected counts or values. If the outcome is unconfirmed, inspect the run record and use verification controls where available before submitting another write.

Build or refresh a stored output​

When your deployment permits stored refreshes and the destination is configured, Refresh once runs the published version and checks the stored result. Publishing by itself saves the definition. If a refresh fails its checks, Studio retains the last successful output. Use Recent refreshes for status and check results, and Review retained outputs for retention details where available.

Scheduled refreshes require an enabled deployment, a published version, and approval for the execution identity and destination. Verify the first completed run and its results before relying on a schedule. You can still preview published definitions and use a live dataset in an authorized dashboard when stored refreshes are disabled.