Row Zero and Databricks: Share a Workbook or Publish a Dataset?

If you have analyzed Databricks data in Row Zero, decide what the recipient needs before sharing: access to a working spreadsheet, or a public table of approved results. Those are different deliverables. A colleague may need to refresh a connected report; a reader of your research may need only a documented monthly summary they can filter without an account.

For example, a public release might contain one row per month and region, with columns such as month, region, completed_requests, and median_wait_days. That is a proposed publication structure—not a reason to expose the underlying request-level records.

What the Databricks acquisition changes

On September 24, 2026, Databricks announced its acquisition of Row Zero. The release says Genie will gain a native governed spreadsheet experience and that Row Zero will continue supporting sources beyond Databricks. It describes permission-aware queries, live-source refreshes and export restrictions.

The announcement uses future-facing language for native integration and availability across major clouds. As of October 3, 2026, do not treat it as a rollout checklist or assume every described capability is enabled in your workspace. Use the documented Row Zero sharing and connection controls for the workflow you have today.

Share a Row Zero workbook with colleagues

1. Check which identity runs the query

Row Zero’s Databricks connector documentation supports personal access tokens and OAuth. OAuth is the documented choice when users have their own Databricks credentials: queries respect the permissions of the person running them, including row-level security and role-based access control.

For reusable shared data sources, the distinction is explicit:

• OAuth: the query runs through the end user’s connection; warehouse row-level security applies, and the user needs database access.
• Service account: the query runs through the connection that created the source; users in the organization who receive access to that source can run it.

These are different access models. Do not assume a shared service-account query applies each recipient’s personal Databricks permissions. Also distinguish permission to run a query from permission to view an already-populated workbook; review both before sharing.

2. Assign the workbook role

In the workbook, select Share in the upper-right corner, enter the recipient’s email address, assign the appropriate role and click Invite:

Role	Documented capability
Viewer	View, but not modify, the workbook
Editor	Modify the workbook; needs access to the backing connection or data source to modify or rerun a connected-table query
Owner	Manage workbook access and advanced sharing policies

Named recipients are prompted to log in or create an account. These mechanics and role limits are described in Row Zero’s sharing documentation. Use Viewer when the recipient only needs to read the report.

3. Check link scope and export policy

A Row Zero link is not necessarily private. Row Zero’s sharing documentation says individual Free and Pro accounts can create public links accessible without login. Business link sharing defaults to workspace users; Enterprise customers can restrict it to the workspace or disable it. Named sharing can also be restricted by Enterprise policy. Check the actual audience rather than inferring privacy from the word “Share.”

Row Zero also documents Enterprise restrictions on exports, duplication and copy/paste. Those are controls to inspect—not a guarantee that every workbook has them enabled.

Before sending a link, verify the intended recipient’s access, role, visible data and ability to refresh the report. If access fails, resolve the permission issue rather than sending an unrestricted export as a workaround.

Publish approved results as a public dataset page

Public publication is a separate release decision. Neither warehouse access nor permission to view a workbook is approval to publish its contents.

1. Create a publication-only table. Select approved fields and rows, remove confidential information, and review whether aggregates could reveal individuals or sensitive activity. Do not upload sensitive data to a public publisher.
2. Export only that table. Where export is authorized and enabled, select a range or data table in Row Zero, right-click and choose Export to CSV. CSV export is documented for Pro, Business and Enterprise accounts, not Free accounts. Enterprise policy can disable it, and a range mixing ordinary cells with data tables cannot be exported together. See Row Zero’s export documentation.
3. Validate the finished file. Compare headers, record counts, totals and sampled values against the approved table. Preserve the distinction between zero and missing values. Use the CSV export validation checklist before upload.
4. Publish and inspect the reader view. TablePage accepts CSV, TSV, XLSX and XLS files and generates public dataset pages with filterable tables and charts, without reader signup. Upload only the reviewed release, then check labels, values and chart interpretations on the generated page.

Accompany the link with the source, coverage period, extraction date, field definitions and any exclusions or missing-value conventions. Treat the exported file as a dated snapshot: the Row Zero workbook’s connection does not make that separately uploaded file live.

Use the Row Zero workbook for authorized analysis and collaboration. Use a separately reviewed public dataset for readers who need to explore the released results—not the underlying workspace.