Create a virtual table

A share must be established between CONNECT data services and Databricks before you can create a virtual table.

Timestamp precision

Virtual tables currently support timestamps with a maximum precision of microseconds (1 × 10⁻⁶ s), which is aligned with limitations in data lake providers.

When data from CONNECT data services or upstream industrial sources includes timestamps with nanosecond precision, the additional fractional digits (beyond 6 decimals) are truncated or rounded during the virtual table creation process. This limitation originates from the Apache Spark and Delta Lake storage formats used to materialize the data (TIMESTAMP_MICROS in Parquet).

As a result:

• The original nanosecond precision from CONNECT data services sources will not be preserved in the Delta table.

• Downstream consumers such as Databricks, Power BI, or other BI tools via delta sharing will observe timestamps rounded to the nearest microsecond.

Example:

2025-10-21T14:23:15.123456789Z is stored as 2025-10-21T14:23:15.123456Z.

Create a virtual table in CONNECT data services

To create a virtual table, perform the following steps:

1. In the left pane, select Analytics > Virtual Tables.

2. Select the Virtual Tables tab.

3. Select Create virtual table.

4. Select a data view for use in your virtual table.

   For more information on data views, read Data views.

5. Select Next.

6. Type a name for your virtual table in the Name box.

   The name of your chosen data view is visible for reference.

7. Select an Interpolated or Stored retrieval mode.

   • Interpolated: Returns data between specified intervals.

     If you select Interpolated, provide an Interpolation interval (dd.hh:mm:ss).

   • Stored: Returns stored or window data.

   For more information on data retrieval types, see Preview and save the data view.

8. Select a Retrieval interval and provide a time window.

   • Time-Fixed Window: contains a beginning and an end time.

   • Time-Extending Window: only contains a start time.

     Provide a Refresh interval for the Scheduler configuration when you choose a Time-Extending Window.

9. Select Next.

10. Select the Share(s) you want in your virtual table.

    If you do not have an available share, select Create share. For more information on how to create a share, see Create a share.

11. Select Next.

12. Review the summary of your virtual table and select Save.

Create a catalog in Databricks

Once the virtual table has been shared with the target Databricks metastore, your Databricks metastore administrator will need to create a catalog for your workspace from the share. The virtual table will then appear in the catalog.

For information about data access, permissions, and creating a catalog in Databricks, read the Databricks documentation.