Table of Contents

Table properties

  • 2026-09-23
  • Jeroen ter Heerdt
Applies to:
  • ✔ Tabular Editor 2
  • Tabular Editor 3
    ✔ Desktop
    ✔ Business
    ✔ Enterprise

This page covers the properties of the three kinds of tables in a semantic model: regular tables, calculated tables and calculation group tables. Calculated tables and calculation group tables are based on a regular table, so the shared properties are described once, under Table. The sections for the other two kinds cover what they add and which table properties the Properties view hides for them. For properties that most objects share, see Common properties.

Table

A table holds rows of data in columns, and it's also the container for the measures and hierarchies that belong to it. A regular table gets its data from one or more partitions, each with a query against a data source, for example a Power Query (M) expression or a SQL query. How the data gets there depends on the partition's mode: import, DirectQuery or Direct Lake.

In Tabular Editor 3, you add tables from a data source with Import Tables, and you look at a table's data with Table Preview.

Common properties

Basic

Table Group

TableGroup · string · stored as an annotation (TabularEditor_TableGroup)

The table group the table appears in, in the TOM Explorer of Tabular Editor 3. Table groups work like folders for tables and help you find your way in a large model, for example by keeping fact tables, dimension tables and calculation groups apart. Leave it empty to show the table at the top level.

Table groups can't be nested, and client tools such as Power BI and Excel ignore them. Tabular Editor stores the group in an annotation called TabularEditor_TableGroup on the table, so the grouping is saved with the model and other developers see the same groups. The Properties view only shows Table Group while table groups are turned on in the TOM Explorer. See Table Groups.

In the TOM Explorer, you can also right-click tables and use Move to group, or drag tables onto a table group (see @drag-drop). To put all tables in groups at once, based on their type and relationships, run the Create Table Groups script.

Incremental Refresh

Enabled

EnableRefreshPolicy · bool · shortcut: adds or removes the table's Refresh Policy

Turns incremental refresh on or off for the table. When you set it to true, Tabular Editor adds a refresh policy to the table, and the Refresh Policy property appears so you can configure it. When you set it back to false, Tabular Editor removes the refresh policy object from the table. It doesn't touch the table's partitions: any policy range partitions that the policy created stay in the table. To turn the table back into a table with a regular M partition, follow the steps in Removing Incremental Refresh.

Incremental refresh is only available for Power BI and Fabric models. The Properties view only shows Enabled and Refresh Policy when the model's compatibility mode is Power BI, the compatibility level is 1450 or higher, and the table either has no partitions yet or has M or query partitions in import or dual mode. For Analysis Services models, you create and manage the partitions yourself. See What is a Refresh Policy? and Set Up a New Refresh Policy.

The Setup Incremental Refresh script sets up a refresh policy based on a date column you select, including the RangeStart and RangeEnd parameters.

Refresh Policy

RefreshPolicy · BasicRefreshPolicy

The incremental refresh policy of the table. Expand it in the Properties view to set how much history to keep, how much to refresh, the source expression and the other settings of the policy. The individual settings are described under refresh policy on Partition properties. See also Incremental Refresh.

A refresh policy doesn't create partitions by itself. The engine creates the policy range partitions the first time the policy is applied, for example when you refresh the table in the Power BI service, or when you right-click the table in the TOM Explorer of Tabular Editor 3 and choose Apply refresh policy. Tabular Editor only applies the policy when it's connected to the model and you've saved all changes to it first.

In Tabular Editor 2, right-click the table in the TOM Explorer and choose Apply Refresh Policy. It has the same requirements.

To change an existing policy, see Modify an Existing Refresh Policy. In Tabular Editor 3, the Advanced Refresh Dialog dialog can apply the refresh policy with an effective date other than today, so you can test how the policy behaves at a different point in time.

Metadata

Source

Source · string · computed · read-only

The name of the legacy (provider) data source that the table's first partition queries. It's a quick way to see where a table comes from without opening its partitions. Tabular Editor only looks at the first partition, so if the partitions use different data sources, you only see the first one.

The Properties view only shows Source when the first partition is a query partition or an M partition. For an M partition it's empty, because M partitions don't refer to a data source object: the connection is part of the M expression. It isn't shown for calculated tables, calculation group tables, Direct Lake tables or tables with policy range partitions.

To go the other way and list the tables that use an explicit (legacy) data source, run the Show Data Source Dependencies script.

Source Type

SourceType · PartitionSourceType · read-only · shortcut to Source Type of the first partition

The kind of source the table's partitions use. It tells you at a glance whether the table is imported with M, part of a Direct Lake model, and so on. Tabular Editor takes it from the table's first partition, so for a table with partitions of different kinds, you see the kind of the first one. A table without partitions shows None.

The Properties view doesn't show Source Type on calculated tables and calculation group tables. In a C# script, those return Calculated and CalculationGroup.

Value Meaning
Query The partitions run a query, for example SQL, against a legacy (provider) data source.
Calculated The table is a calculated table. Its data comes from a DAX expression.
None The source isn't defined. The data is pushed into the table, for example in a push semantic model in Power BI.
M The partitions use a Power Query (M) expression. This is the most common value in Power BI models. Compatibility level 1400+.
Entity The partitions read a named entity, such as a table in a lakehouse or warehouse. Direct Lake tables use this. Compatibility level 1400+.
PolicyRange The partitions were created by an incremental refresh policy. Compatibility level 1450+.
CalculationGroup The table is a calculation group table. Compatibility level 1470+.
Inferred The engine generates the query that fills the partition. Compatibility level 1563+.

Options

Alternate Source Precedence

AlternateSourcePrecedence · int · compatibility level 1460+

The order in which the engine considers this table when it looks for an aggregation table that can answer a query. It's only relevant when the table is an aggregation table, that is, when some of its columns have Alternate Of set to a column in a detail table. When more than one aggregation table can answer a query, the engine uses the one with the highest value.

Typically you give a small, highly summarized aggregation table a higher value than a larger, more detailed one, so the engine tries the smaller table first. The default is 0. In Power BI Desktop, this is the Precedence setting in Manage aggregations. To set up an aggregation table in Tabular Editor, see User-defined Aggregations. The Implement User-defined Aggregations script automates those steps for a fact table you select.

Calendars

Calendars · collection of Calendar · read-only

The custom calendars defined on the table. A calendar tells the engine which columns of a date table hold the year, quarter, month, week and date, so that calendar-based time intelligence functions work with fiscal, retail or other non-standard calendars. Custom calendars need compatibility level 1701 or higher. See Calendar properties.

In Tabular Editor 3, right-click a table and choose Create > Calendar... to add a calendar, then map its columns in the Calendar Editor. See Calendars (Enhanced Time Intelligence).

Data Category

DataCategory · string

Tells client tools what kind of data the table holds. In practice, the only value that matters in Power BI is Time: it marks the table as a date table. Power BI then uses the table's date column (the column with Key set to true) for time intelligence and turns off the auto date/time table for it. When you use Mark as date table in Power BI Desktop, this is what it sets.

Leave it empty, or set it to Regular, for all other tables. The other values come from multidimensional models and only some Excel features and older client tools use them: Geography, Organization, BillOfMaterials, Accounts, Customers, Products, Scenario, Quantitative, Utility, Currency, Rates, Channel and Promotion. The dropdown in the Properties view lists these values and Unknown, but you can also type a value yourself.

In Tabular Editor, you can also right-click the table in the TOM Explorer view and choose Mark as date table.... The Best Practice Analyzer rule Date Table Should Exist flags models that have no date table. To create a date table from the date columns in your model, run the Create Date Table script, which also sets Data Category to Time.

Default Detail Rows Expression

DefaultDetailRowsExpression · string · compatibility level 1400+

A DAX table expression that defines the rows a user sees when they drill through on a measure in this table, for example with Show Details in an Excel PivotTable. It applies to every measure in the table that doesn't have its own Detail Rows Expression (see @object-properties-measures).

For example, to show the order number, customer and amount for each sales row:

SELECTCOLUMNS (
    Sales,
    "Order", Sales[OrderNumber],
    "Customer", RELATED ( Customer[Name] ),
    "Amount", Sales[Amount]
)

Power BI doesn't use this property. See Detail Rows Expression. In Tabular Editor 3, you write the expression in the Expression Editor of the DAX Editor, with auto-complete and parameter info. To remove the expression, clear the value.

Direct Lake Indexing Behavior

DirectLakeIndexingBehavior · DirectLakeIndexingBehavior · Preview compatibility level

Controls when the engine builds and saves the indexes it uses for a Direct Lake table. Building indexes takes time. Doing it during a refresh means the first queries after a refresh are faster, and doing it on demand means refreshes are faster. It only applies to tables in Direct Lake mode.

TOM doesn't assign this property a released compatibility level yet, so the Properties view in the current version of Tabular Editor 3 hides it for every model.

Value Meaning
Default Use the model's default, set with Default Direct Lake Indexing Behavior on Model properties.
Auto The engine may build in-memory indexes when it loads data or runs a query.
Full The engine builds and saves the indexes as part of a full refresh, a calculate refresh and a refresh of type indexes.
Explicit The engine only builds and saves the indexes when you explicitly run a refresh of type indexes. Other refresh types don't build them.

Exclude From Automatic Aggregations

ExcludeFromAutomaticAggregations · bool · compatibility level 1572+

When true, the table isn't used when Power BI builds automatic aggregations. Automatic aggregations are a Power BI Premium and Fabric feature for DirectQuery models: the service watches the queries and creates aggregation tables for you. Set this property on a table you don't want included, for example because its data changes too often for a cached aggregation to be useful.

Exclude From Model Refresh

ExcludeFromModelRefresh · bool · compatibility level 1480+

When true, a refresh of the whole model doesn't refresh this table's partitions if they already hold data. You can still refresh the table by selecting it explicitly. Use it for tables that rarely change or are expensive to load, such as a large historical table or a table filled with a separate process, so that a scheduled model refresh doesn't reload them every time.

Partitions that don't hold data yet are still refreshed, so a new or cleared table gets loaded.

In Tabular Editor 3, you refresh a single table by right-clicking it in the TOM Explorer and choosing Refresh table. See Refreshing, previewing and querying data.

On SQL Server 2025 Analysis Services, every type of model-level refresh skips the table when its partitions already hold data: Full, Automatic, Data only and Calculate. The exception is Clear values, which does clear the table's data. The next model refresh then loads the table again, because its partitions are empty.

Partitions

Partitions · collection of Partition · read-only

The partitions of the table. Each partition defines where a part of the table's data comes from, for example an M expression, and each can be refreshed separately. Most tables have a single partition. Large import tables often have one partition per period, either created by you or by an incremental refresh policy.

Click the ellipsis button to open the collection editor, where you can add, remove and edit partitions. See Partition properties.

The Properties view only shows Partitions when the table's first partition is an M partition or a query partition. It doesn't show it for calculated tables (including field parameters), calculation group tables, Direct Lake tables or tables with policy range partitions. To view and edit the partitions of calculated tables and calculation group tables anyway, run the Edit Hidden Partitions script.

Private

IsPrivate · bool · compatibility level 1400+

When true, client tools don't show the table at all, not even to developers who choose to see hidden objects. Power BI Desktop uses this for the date table template that the auto date/time feature uses to create its date tables (the table named DateTableTemplate_ followed by a GUID). You rarely need to set it yourself.

It's always available in Power BI models. In Analysis Services models it needs compatibility level 1400 or higher. SQL Server 2025 Analysis Services stores the flag but doesn't act on it: the table is still listed in the schema information that client tools read, you can query it with DAX, and measures in other tables can use it.

Like Hidden, this isn't a security feature. To secure a table, use object-level security (see Object Level Security below).

In Power BI Desktop, a private table doesn't appear in the Data pane at all, unlike a hidden table, which you can still show there. It still works in DAX: you can query it, and measures in other tables can use it. It's also still listed in the schema information that the engine returns.

The Best Practice Analyzer rule Remove Auto Date Tables flags the DateTableTemplate_ and LocalDateTable_ tables that auto date/time creates, so you can replace them with a single date table.

Sets

Sets · collection of Set · read-only

The calculated sets defined on the table. A set is a named DAX expression that returns a set of members, which client tools can offer as a predefined selection. Sets are only supported in Power BI models: the Properties view only shows Sets when the model's compatibility mode is Power BI and the compatibility level is 1400 or higher. Sets don't appear as objects in the TOM Explorer, so you add and edit them here, in the collection editor. See Function and set properties.

Show As Variations Only

ShowAsVariationsOnly · bool · compatibility level 1400+

When true, client tools don't show the table in the field list on its own. It only appears through a column variation that points to it. Power BI Desktop uses this for the date tables that auto date/time creates (the tables named LocalDateTable_ followed by a GUID): you see them as the date hierarchy under a date column, and not as separate tables. See Variations on Column properties.

It's always available in Power BI models. In Analysis Services models it needs compatibility level 1400 or higher. SQL Server 2025 Analysis Services stores the flag but doesn't act on it: the table is still listed in the schema information that client tools read, you can query it with DAX, and measures in other tables can use it.

System Managed

SystemManaged · bool · compatibility level 1562+

When true, the engine or the service owns the table: it creates and deletes it, and you shouldn't change it by hand. Don't set this property on your own tables. Tabular Editor doesn't stop you from editing a system-managed table, so take care not to change one by accident.

The auto date/time tables that Power BI Desktop creates aren't system-managed: Desktop marks the template table as hidden and private, and each local date table as hidden with Show As Variations Only.

Translations, Perspectives, Security

Object Level Security

ObjectLevelSecurity · per role · compatibility level 1400+ · shortcut to the table permissions of each role

The object-level security (OLS) permission of the table in each role of the model. Expand the property to see one entry per role. The Properties view only shows the property when the model has at least one role. The values are:

Value Meaning
Default No permission is set for this role. Members of the role can see the table.
Read Members of the role can see and query the table.
None The table is hidden from members of the role, together with everything that depends on it. Queries and measures that refer to it return an error for them.

Unlike Hidden and Private, OLS is a real security feature: the table and its metadata don't exist for members of the role. Use it with care, because every measure that refers to the table fails for those users. Tabular Editor stores the permission in the table permission of each role. See Setup or Modify OLS and Role and security properties.

To check that OLS and RLS work as intended, test them with impersonation in Tabular Editor 3. See Testing RLS/OLS.

Row Level Security

RowLevelSecurity · per role · read-only · shortcut to the table permissions of each role

The row-level security (RLS) filter of the table in each role of the model. Expand the property to see one entry per role. Each entry is a DAX expression that returns TRUE for the rows that members of the role may see, for example:

Customer[Country] = "Denmark"

The property itself is read-only, but you can edit each role's entry, and the entries accept multi-line expressions. The Properties view only shows the property when the model has at least one role.

Leave an entry empty to not filter the table for that role. Filters also flow along relationships, so filtering a dimension table also filters the fact tables it's related to. Tabular Editor stores the filter in the table permission of each role. See Roles and Row-Level Security, Setup or Modify RLS and Role and security properties.

Calculated table

A calculated table gets its data from a DAX expression instead of from a data source, for example a date table built with CALENDAR or a table that summarizes another table. The engine evaluates the expression when you refresh the table or the model, and stores the result like any other imported table.

A calculated table has all the properties of a table, plus Expression.

A calculated table has exactly one partition, which Tabular Editor manages for you, and it can't have an incremental refresh policy. That's why the Properties view hides Partitions, Source, Source Type, Enabled and Refresh Policy on calculated tables. Alternate Source Precedence is still shown.

Field parameters in Power BI are calculated tables too. The Create Field Parameter script creates one from the columns or measures you select.

Common properties

The same as on tables.

Options

Expression

Expression · string

The DAX expression that returns the table's rows. It must return a table, for example:

ADDCOLUMNS (
    CALENDAR ( DATE ( 2020, 1, 1 ), DATE ( 2030, 12, 31 ) ),
    "Year", YEAR ( [Date] ),
    "Month", FORMAT ( [Date], "mmm" )
)

The table's columns come from the expression: each column the expression returns becomes a calculated table column (see @object-properties-columns). In Tabular Editor 3, the columns change as soon as you change the expression: Tabular Editor works them out with its own DAX semantic analysis, so this also works when you edit a model file without a connection to Analysis Services or Power BI (see @creating-and-testing-dax). Adding, removing or renaming columns this way can affect other DAX expressions that refer to the table, and hierarchies that use its columns.

Tabular Editor 2 doesn't analyze the expression. It only updates the columns when you save the model to a database you're connected to: after the save, it reloads the table's columns from the server. When you edit a model file, the columns don't change with the expression.

Edit the expression in the Expression Editor. In Tabular Editor 3, the Expression Editor of the DAX Editor offers auto-complete and parameter info. The engine recalculates the table's data when the table is refreshed, and also when a table it depends on is refreshed.

After you change the expression, refresh the table before you use it in a report. See Refreshing data.

Calculation group table

A calculation group table is the table that holds a calculation group. It has one column, whose values are the names of the calculation items, and usually an ordinal column that sets the order of the items. Report authors use the column like any other column, for example in a slicer, and the selected calculation item changes how the measures in the report are calculated. See Calculation group properties for the properties of the calculation group and its items.

Microsoft documents calculation groups for compatibility level 1500 or higher, which includes all Power BI semantic models. Tabular Editor 3 already offers Create > Calculation Group from compatibility level 1470, the level at which TOM added calculation groups.

A calculation group table is a table, plus the properties below. Most of them are shortcuts to properties of the calculation group that the table holds, so you can edit them without selecting the calculation group itself.

Many table properties don't apply to a calculation group table, because its data doesn't come from a data source, so the Properties view hides them. Of the table properties, it only shows the common properties listed below, Table Group and Default Detail Rows Expression. Calculation groups don't support detail rows expressions, row-level security or object-level security, so leave Default Detail Rows Expression empty.

Common properties

Basic

Calculation Group Description

CalculationGroupDescription · string · shortcut to Description of the calculation group

The description of the calculation group. It's stored on the calculation group, not on the table, and it's separate from the table's own Description. Client tools don't show it. In the Data pane of Power BI Desktop, hovering over a calculation group table shows the table's own Description, not this one, and Excel's field list shows no descriptions at all. So put the text you want users to see in the table's Description, and use this property for notes to other developers.

Calculation Group Precedence

CalculationGroupPrecedence · int · shortcut to Precedence of the calculation group

The order in which the calculation group is applied when a query uses more than one calculation group. It's a shortcut to Precedence on the calculation group. See Precedence on Calculation group properties for how it works.

Metadata

Calculation Group Annotations

CalculationGroupAnnotations · collection of name/value pairs · read-only · shortcut to Annotations of the calculation group

The annotations stored on the calculation group, as opposed to the annotations on the table (see Annotations). Tools rarely use these. The property itself is read-only, but you can click the ellipsis button to open the annotation editor, where you add, change and remove the annotations. You can only do this with one calculation group table selected.

Options

Multiple or Empty Selection Expression

MultipleOrEmptySelectionExpression · string · compatibility level 1605+ · shortcut to the calculation group

A DAX expression that the engine uses instead of the measure when the filter on the calculation group selects more than one calculation item, or selects none because the filter doesn't match any item. Without it, the calculation group isn't applied in those cases. What measures then return depends on Selection Expression Behavior on the model: with Automatic or NonVisual they return their normal value, and with Visual they return a blank value. A normal value can be misleading, for example when a user selects two currencies in a currency conversion calculation group and sees a total that isn't in either currency.

Use SELECTEDMEASURE() to refer to the measure. For example, to return an error message:

ERROR ( "Select a single currency." )

Or to return a blank value:

BLANK ()

Multiple or empty selection expressions need compatibility level 1605 or higher. Below that level, the Properties view doesn't show this property or the other selection expression properties below. To remove the expression, clear the value.

In Tabular Editor 3, you write the expression in the Expression Editor of the DAX Editor, with auto-complete and parameter info.

Multiple or Empty Selection Expression Description

MultipleOrEmptySelectionDescription · string · shortcut to the calculation group

A description of the multiple or empty selection expression, for other developers. Client tools don't show it to report authors.

Multiple or Empty Selection Format String Expression

MultipleOrEmptySelectionFormatStringExpression · string · shortcut to the calculation group

A DAX expression that returns the format string to use when the multiple or empty selection expression applies. It works like Format String Expression on a calculation item. Inside the expression, SELECTEDMEASUREFORMATSTRING() returns the format string of the measure.

No-selection Expression

NoSelectionExpression · string · compatibility level 1605+ · shortcut to the calculation group

A DAX expression that the engine uses instead of the measure when there's no filter on the calculation group at all, that is, when the user hasn't selected any calculation item. Without it, measures return their normal value when nothing is selected.

Use it to set a default calculation that users can still override by selecting a calculation item. For example, in a currency conversion calculation group, you can convert amounts to the currency selected in a Currency table without the user having to pick a calculation item. Sales amounts are stored in US dollars, and a Currency Rate table holds the daily exchange rates:

IF (
    SELECTEDVALUE ( 'Currency'[CurrencyName], "US Dollar" ) = "US Dollar",
    SELECTEDMEASURE (),
    SUMX (
        VALUES ( 'Date'[Date] ),
        CALCULATE ( DIVIDE ( SELECTEDMEASURE (), MAX ( 'Currency Rate'[EndOfDayRate] ) ) )
    )
)

No-selection expressions need compatibility level 1605 or higher.

No-selection Expression Description

NoSelectionExpressionDescription · string · shortcut to the calculation group

A description of the no-selection expression, for other developers. Client tools don't show it to report authors.

No-selection Format String Expression

NoSelectionFormatStringExpression · string · shortcut to the calculation group

A DAX expression that returns the format string to use when the no-selection expression applies. It works like Format String Expression on a calculation item. Inside the expression, SELECTEDMEASUREFORMATSTRING() returns the format string of the measure.