Bewerken

Delen via


Tabular Object Model (TOM)

Applies to: SQL Server 2016 and later Analysis Services Azure Analysis Services Fabric/Power BI Premium

The Tabular Object Model (TOM) is an extension of the Analysis Management Object (AMO) client library, created to support programming scenarios for tabular models created at compatibility level 1200 and higher. As with AMO, TOM provides a programmatic way to handle administrative functions like creating models, importing and refreshing data, and assigning roles and permissions.

TOM exposes native tabular metadata, such as model, tables, columns, and relationships objects. A high-level view of the object model tree, provided below, illustrates how the component parts are related.

Because TOM is an extension of AMO, all classes representing new tabular objects are implemented in a new Microsoft.AnalysisServices.Tabular.dll assembly. General-purpose classes of AMO were moved to Microsoft.AnalysisServices.Core assembly. Your code will need to reference both assemblies. See Install, distribute, and reference the Tabular Object Model (Microsoft.AnalysisServices.Tabular) for details.

The API is available for managed .NET code. To learn more about specific AMO/TOM classes, see Microsoft.AnalysisServices Namespace reference. To review the full list of programming options for tabular models, including script and query language support, see Tabular Model Programming for Compatibility Level 1200.

Tabular object model hierarchy

From a logical perspective, all tabular objects form a tree, the root of which is a Model, descended from Database. Server and Database are not considered tabular because these objects can also represent a multidimensional database hosted on a server running in Multidimensional mode, or a tabular model at a lower compatibility level that does not use tabular metadata for object definitions.

With the exception of AttributeHierarchy, KPI, and LinguisticMetadata, each child object can be a member of a collection. For example, the Model object contains a collection of Table objects (via the Tables property), with each Table object containing a collection of Column objects, and so on.

The lowest level descendant of any parent object in this hierarchy is an Annotation object that can be used to optionally extend the schema as long as you provide the code to handle it.

object hierarchy diagram

TOM is built on top of the AMO infrastructure, which also accommodates multidimensional and tabular databases at compatibility levels below 1200. This has some practical implications. When you manage objects not specified in tabular metadata (such as a Server or Database), you need to leverage parts of the existing AMO stack that describe those objects. Along with the legacy API is the concept of major and minor objects that provide granular descriptions of object state as discovered from the server, or when saved to the server. The MajorObject class under Microsoft.AnalysisServices namespace exposes methods for Refresh and Update. Minor objects are only refresh or saved via the major object that contains them.

In contrast, when you manage objects that are part of tabular metadata, such as Model or Table, you leverage a completely new tabular stack. Within this stack, updates are fine-grained, which means every metadata object, derived from the MetadataObject class under the Microsoft.AnalysisServices.Tabular namespace, can be individually saved to the server. Typically, you would discover the entire Model. You then make changes to individual metadata objects under it, such as Table or Column. You then call Model.SaveChanges() method which understands changes made by you at fine-grained level, sending commands to the server to update only those objects that changed.

TOM and XMLA

On the wire, TOM uses the XMLA protocol to communicate with the server and to manage objects. When managing non-tabular objects, TOM uses ASSL, the Analysis Services Scripting Language extension of XMLA. When managing tabular objects, TOM uses the MS-SSAS-T tabular protocol, also an extension of XMLA. To learn more, see MS-SSAS-T SQL Server Analysis Services Tabular protocol documentation.

TOM and JSON

Tabular metadata, which is structured as JSON documents, has a new command and object model definition syntax via the Tabular Model Scripting Language (TMSL). The scripting language uses JSON for the body of requests and responses.

Although both TMSL and TOM expose the same objects, Table, Column and so forth, and the same operations, Create, Delete, Refresh, TOM does not use TMSL on the wire. TOM uses the MS-SSAS-T tabular protocol instead, as previously noted.

As a user, you can choose whether to manage tabular databases through the TOM library from your C# program or PowerShell script, or through TMSL script executed through PowerShell, SQL Server Management Studio (SSMS), or a SQL Server Agent Job.

The decision to use one or the other will come down to the specifics of your requirements. The TOM library provides richer functionality compared to TMSL. Specifically, whereas TMSL only offers coarse-grained operations at the database, table, partition, or role level, TOM allows operations at a much finer grain. To generate or update models programmatically, you will need the full extent of the API in the TOM library.

Using TOM with Power BI

Power BI Premium, Premium Per User, and Power BI Embedded workspaces support open-platform connectivity through the XMLA endpoint. With the XMLA endpoint, custom tools, script, and automated processes can be used for data modeling and to perform workspace and semantic model administrative tasks.

Before creating a .Net application using TOM to work with Power BI semantic models, be sure to read Semantic model connectivity with the XMLA endpoint in the Power BI documentation. This article describes how to enable the XMLA endpoint for read-write access, get a workspace connection URL, and other important aspects for semantic model management with custom apps, external tools, and scripts.

To learn more about using the Tabular Object Model for semantic model administration and management, see Programming Power BI semantic models (TOM).

See also

Compatibility level for tabular models
Analysis Services client libraries
XML for Analysis (XMLA) Reference
Analysis Management Objects (AMO)