Connect to a semantic model

In this article, you look at the steps to connect to a semantic model from a PowerTable sheet. Connect to an existing semantic model and create a table app. The data table, along with any changes or updates, is saved to your preferred destination database.

In this method, the semantic model is the input source, and the created database is the destination that stores the changes written back from the app.

Prerequisites#

Before you begin, make sure that you have the following prerequisites in place:

• Connection established to the Fabric SQL database that stores the data for your PowerTable app.
• A plan item created in your Fabric workspace.

Create a new PowerTable sheet#

1. In your plan, select New PowerTable Sheet or select the PowerTable icon on the landing page. Enter a name for the sheet and select Create.

2. Select Create a New App to create your app. You can also select Explore PowerTable to experience a sample PowerTable app.

3. Select a Connection: Choose a Fabric SQL connection if available. If there are no connections, create one by selecting Create Connection.

4. Database Name: Select the Database Name field to open the list of databases available.

5. Select the destination Fabric SQL database to store the table data and then select Add.

6. Select Connect.

Create a table#

1. Select New Table.
2. Choose the database schema.
3. Enter a Table Name. Currently, table names support English characters only.
4. Choose Connect To Semantic Model in Import Data.
5. Select Semantic Model Item and then select the search field to choose a semantic model.

6. Select a semantic model from the list and select Add.

7. Select Next.

Map data

1. Map and assign your data to the table by selecting the required fields and corresponding values.

2. Set the primary key fields by selecting the three dots and selecting primary keys.

3. If needed, you can use filters to include only specific field values from the table.

4. Select Next.

Configure table

PowerTable sheets automatically detect column properties and rename any unsupported column names to match the supported format. You can review and modify them as needed.

1. Review the detected column settings. Modify them if required.
2. If the primary key isn't selected, select the appropriate column as the primary key.
3. Configure columns as required:
   • Mark fields as Mandatory if they must contain values.
   • Set fields with distinct values as Unique Key to allow unique values.
   • Set a column as an Identity Column when you need PowerTable to generate unique, sequential values automatically. Only numeric columns and primary key columns can be configured as identity columns.
4. Modify column properties, including:
   • Length
   • Data type
   • Input type
   • Display name
5. Enter default values for columns, if required.
6. Optionally, select Add Column to create additional columns.
7. To create a composite primary key, select Unique Combination, and then select two or more columns to define a unique combination key.

Finish

Select Finish.

The table app is created successfully in PowerTable with the configured columns and values. Select Save to save your table.

Refresh semantic model

Select Refresh > Semantic Model in the status bar to refresh and retrieve the latest data from the semantic model. Use this option when the source data changes and you want to update the data displayed in the PowerTable sheet before saving it to the destination database.

Refreshing the semantic model updates the data available in the PowerTable sheet, including newly added or updated records. This lets you work with the latest source data before saving the table to the database.

FAQ #

Do changes made in PowerTable update the semantic model or its underlying data source?

No. PowerTable uses the semantic model only to seed the destination table.

When you create a PowerTable app from a semantic model, PowerTable reads the semantic model and copies its data into the selected Fabric SQL database destination table.

After the initial load:

• All changes made in PowerTable are written directly to the Fabric SQL database destination table.
• Changes made in PowerTable don't update the semantic model or its underlying data source.

Are there any exceptions to this based on the underlying data source (Lakehouse, Warehouse, or external database) or on the semantic model storage mode?

No. The Save to Database option always writes data to the configured Fabric SQL database destination table, regardless of the semantic model's underlying data source or storage mode.

When you create a PowerTable app from a semantic model:

• The semantic model is used only to seed the PowerTable with initial data. The underlying data source, whether a Lakehouse, Warehouse, or external database, isn't written to.
• The semantic model's storage mode, whether Import, DirectQuery, or Direct Lake, doesn't affect the write path. It only determines how PowerTable reads the seed data.
• After the initial load, all changes made in PowerTable are written directly to the configured Fabric SQL database destination table.

When are newly added members in PowerTable reflected on a planning sheet? Is it supported?

Yes. PowerTable supports this scenario. Newly added members appear in the planning sheet in two ways:

• Direct connection: When PowerTable writes to a Fabric SQL database table, the table appears in the planning's data pane under Data > From Sheets, where you can assign its columns directly as dimensions and measures. There is no replication or intermediate semantic model. Newly added members are available in the planning sheet as soon as the changes are committed.
• Direct Lake connection: When the planning sheet uses a Direct Lake semantic model, PowerTable writes the data to the Fabric SQL database table, which is automatically replicated to OneLake. The Direct Lake semantic model reads the replicated data from OneLake. With automatic updates enabled, newly added members are automatically reflected in the planning sheet after the changes reach OneLake, although a short delay might occur because replication has near-real-time latency.

Next steps #

Configure access control and automated workflows and approvals for your new app.