Tableau Desktop and Web Authoring Help

Data Cloud and Tableau Next support the creation of semantic models using the Tableau Semantics platform. For more information, see Create a Semantic Model in Tableau Semantics (Data Cloud) and About Semantic Models (Tableau Next).

These semantic models can be used directly within Data Cloud and Tableau Next, or they can also be used as a data source in Tableau.

Complete the following steps to connect to a Tableau Semantics model in Tableau Cloud or Tableau Desktop (version 2025.2 and later). The first steps are the same as the Data Cloud connector.

1. Open the Data menu, then New Data Source. (If you just launched Tableau Desktop, ignore this step.)
2. In the Connect pane, under the To a Server section, select Tableau Semantics.
3. Log in and allow access when prompted.
4. Select the Data space.

If you're using Tableau Desktop 2023.1 or earlier, see Tableau Desktop (version 2023.1 and earlier) in the Salesforce Data Cloud connector topic for driver installation information.

A semantic model, by definition, contains more than the data alone. The semantic information contained in the model is also retrieved by the connector using semantic metadata queries. Semantic queries use a query syntax that doesn’t specify how to get the data (what joins, how to aggregate, and so on) but instead requests the data that is needed and leaves the rest to the source semantic model.

Extracts are not available for Tableau Semantics due to how much functionality relies on the source data model.

Semantic connectors function slightly differently than standard connectors. Some Tableau features are affected when you use a semantic data source. This is similar to limitations or differences in functionality when using cube data sources and is due to the control maintained by the source technology.

Most functionality will work as expected, but there are some places where the behaviors are different. Sometimes Tableau functionality can be layered on top of the semantic model, but sometimes the Tableau functionality is fully blocked.

Many semantic features function normally, including (but not limited to):

• Changing data types - only for fields created in Tableau, fields from the semantic model can't have their data type changed
• Converting between measure and dimension, or continuous and discrete
• Creating and using calculated fields - with some limitations: ATTR and user functions are not available
• Creating and using parameters
• Creating hierarchies
• Custom dates
• Duplicating a field- with some limitations: edit a copy is not available for calculations from the semantic model.
• Creating and using geographic fields
• Grouping the data pane by table or folders
• Hiding fields
• Re-aliasing field members
• Renaming fields
• Sorting
• Setting default colors and shapes
• Splitting fields

Tableau functionality can sometimes be layered on top of the semantic model source. This usually applies to functionality created in the semantic model but can be modified in Tableau.

Functionality that's originally created in the semantic model but can be extended (created, edited, or modified) in Tableau includes:

• Field descriptions can be edited or created locally.
• Default aggregation is taken from the semantic model but can be changed in Tableau. This impacts only the default aggregation when the field is used in a Tableau viz.
• Default sorting, like all sorting in a viz, is carried out by Tableau, not sent back to the source
• Number format is taken from the semantic model but can be changed in Tableau
• Changing data types may be allowed, depending on where the field came from:
  • For fields from the semantic model, users have to make any changes to data types in the source rather than in Tableau
  • For fields created in Tableau, the data type can be changed in Tableau.
• Parameters from the semantic model source remain available, and new parameters can be added locally in Tableau.

If these semantic definitions are changed in the underlying model, the updates are not carried through to Tableau because Tableau claimed local control. To reset back to the definitions from the source model, clear or revert the definition (such as clearing the sort or revert the field name). In some cases such as retrieving a field description or parameter, clear the field description or delete the parameter, then refresh or reload the data source.

Semantic model definitions sometimes take precedence and block Tableau customizations. This usually applies to semantic-specific functionality, such as changing data types. Tableau respects the semantic model's definitions and doesn't allow modifications locally.

Functionality that's unavailable when using a semantic connector:

• Sets can't be created in Tableau, and sets created in the semantic model aren't available in Tableau.
• Groups can't be created in Tableau, and groups created in the semantic model aren't available in Tableau.
  • No support for groups means that while Clusters can be added from the Analytics pane, the groups created by clustering can't be saved in the Data pane.
• Bins can't be created in Tableau, and bins created in the semantic model aren't available in Tableau.
  • This prevents building histograms.
• Combined fields can't be created in Tableau, and combined fields created in the semantic model aren't available in Tableau.
• Changing data types may be allowed, depending on where the field came from:
  • For fields from the semantic model, users have to make any changes to data types in the source rather than in Tableau
  • For fields created in Tableau, the data type can be changed in Tableau.
• Custom calendars aren't available.
• Disaggregated measures aren't available because the semantic connector passes aggregated data into Tableau.
• View underlying data isn't available because the semantic connector passes aggregated data into Tableau.
• Data transformation functionality such as pivots or merging mismatched fields after a union aren't available.
• Context filters can't be created, though other filter types can.
• Specific calculations are unavailable, including the ATTR() function and user functions.

Some functionality of data sources as a whole is also unavailable, including:

• Data Blending
• Extracting a data source
• Publishing a data source
• Replacing a data source

There are some ramifications of not having sets that extend beyond what might be thought of as semantics or what is obviously tied to sets. Some native Tableau functionality relies on sets behind the scenes. With sets being unavailable for semantic connectors, some additional functionality is unavailable in Tableau. This includes:

• Filter actions
  • Note that filter actions may create their underlying set but it will be invalid.
• Keep Only/Exclude filter options
• Using a viz as a filter in a dashboard
• Viz in tooltip filters