Table definitions in Microsoft Dataverse
Each table provides the capability to store structured data. For developers, tables correspond to the classes you'll use when working with data in Microsoft Dataverse.
Unsure about entity vs. table? See Developers: Understand terminology in Microsoft Dataverse.
Each table has a unique name defined when it's created. This name is presented in several ways:
||Typically, a Pascal cased version of the logical name. For example, Account|
||A plural form of the Schema name. For example, Accounts|
||All lower-case version of the schema name. For example, account|
||All lower-case version of the collection schema name. For example, accounts|
||Used to identify collections with the Web API. By default, it's the same as the logical collection name.
It's possible to change the Entity Set name by programmatically updating the definitions. But you should only do this before any Web API code is written for the table. More information: Entity set name
EntitySetName is automatically generated by changing the
EntityName to the plural of that name. Example: EntityName: Car EntitySetName: Cars. When manually creating a new table the
EntitySetName can be changed. System created
EntitySetName will always be the plural of the
EntityName and these cannot be changed through the client interface. This includes system tables as well as tables created for many-to-many relationship intersects.
When you create a custom table, the name you choose will be prepended with the customization prefix value of the solution publisher associated with the solution that the table was created within. Other than the entity set name, you cannot change the names of a table after it is created. If you want consistent names for definition items in your solution, you should create them in the context of a solution you create associated with a solution publisher that contains the customization prefix you want to use. More information : Solution publisher
Each table also has three properties that can display localized values:
||Typically, the same as the schema name, but can include spaces. For example, Account|
||A plural form of the display name. For example, Accounts|
||A short sentence describing the table that is, Business that represents a customer or potential customer. The company that is billed in business transactions.|
These are the localizable values that are used to refer to the tables in an app. These values can be changed at any time. To add or edit localized values see Translate customized table, form, and column text into other languages.
PrimaryIdAttribute property value is the logical name of the column that is the primary key for the table.
By default, all tables have a single GUID unique identifier column named:
<table logical name>Id.
PrimaryNameAttribute property value is the logical name of the column that stores the string value that identifies the table record. This is the value that will be displayed in a link to open the record in a UI.
Example: The Contact table uses the
fullname column as the primary name column.
Not every table will have a primary name. Some tables are not intended to be displayed in a UI.
PrimaryImageAttribute property value is the logical name of the image column that has been chosen to represent the entity record. This image may be displayed in an application.
This is different from the icon displayed for a table in model-driven apps. The
IconVectorName property contains the name of the SVG web resource that sets this.
- Image columns
- Use image column data
- Sample: Image Operations using Dataverse SDK for .NET
- Sample: Image Operations using Dataverse Web API
Types of tables
The capabilities and behavior of tables depends on several table properties. Most of these properties are relatively simple and have descriptive names. Among four properties that require some more explanation are: Ownership, Activity tables, Activityparty table and Child tables.
Tables can be categorized by how the data within them is owned. This is an important concept related to how security is applied to tables. This information is in the
OwnershipType property. The following table describes the different ownership types:
|Business||Data belongs to the Business unit. Access to the data can be controlled at the business unit level.|
|None||Data not owned by another table.|
|Organization||Data belongs to the organization. Access to the data is controlled at the organization level.|
|User or Team||Data belongs to a user or a team. Actions that can be performed on these records can be controlled on a user level.|
When you create new tables, the only ownership options are: Organization or User or Team. Once you make a choice for this option, you can't change it. Choose User or Team for the most granular control over who can view or perform actions on the records. Choose Organization when this level of control isn't necessary.
An activity is a task performed, or to be performed, by a user. An activity is any action for which an entry can be made on a calendar.
Activities are modeled differently from other kinds of tables that store business data. Data about activities is frequently displayed together in a list, yet each kind of activity requires unique properties. Rather than have a single Activity table with every possible property, there are separate kinds of activity tables and each of them inherits properties from a base ActivityPointer table. These tables will have the
IsActivity property set to true.
|Appointment||Commitment representing a time interval with start/end times and duration.|
|Activity that is delivered using email protocols.|
|Fax||Activity that tracks call outcome and number of pages for a fax and optionally stores an electronic copy of the document.|
|Letter||Activity that tracks the delivery of a letter. The activity can contain the electronic copy of the letter.|
|PhoneCall||Activity to track a telephone call.|
|RecurringAppointmentMaster||The Master appointment of a recurring appointment series.|
|SocialActivity||For internal use only.|
|Task||Generic activity representing work needed to be done.|
Whenever anyone creates one of these kinds of activity table records, a corresponding
ActivityPointer table record will be created with the same
ActivityId unique identifier column value. You can't create, update, or delete instances of the
ActivityPointer table, but you can retrieve them. This is what allows all types of activities to be presented together in a list.
You can create custom activity tables that behave the same way.
This table is used to add structure to activity table
PartyListType columns that include references to other tables. You'll use this table when setting values for activity table columns like
PhoneCall.from. Within the ActivityParty table, you set the ParticipationTypeMask column to define the role that the reference is playing. Roles include
Organizer and more.
You can query the
ActivityParty table, but you can't create, retrieve, update, or delete an activity party outside of the activity that it's related to.
Tables where the
IsChildEntity property is true will never have any privileges defined and will never be User or Team owned. Operations that can be performed on a child table are bound to a table that they're associated to via a Many-to-one relationship. Users can only perform operations on child tables if they can perform the same operation on the related table. Child tables get deleted automatically when the table record they depend on is deleted.
PostRole are each children of the
Each alternate key definition describes one or more columns in combination that will uniquely identify a table. Alternate keys are typically only applied for integration with external systems. You can define alternate keys to uniquely identify a record. This is valuable if you're integrating data with a system that doesn't support GUID unique identifier keys. You can define a single column value or combination of column values to uniquely identify a table. Adding an alternate key will enforce a uniqueness constraint on these columns. You won't be able to create or update another table record to have the same values.
Most tables have two properties to track the state of a record. These are
StateCode, which is called Status in model-driven apps and
StatusCode, which is called Status Reason in model-driven apps.
Both columns contain a set of choices that display the valid values. The
StatusCode column values are linked so that only certain
StatusCode choices are valid for a given
This can vary by table but the common pattern for many tables, and the default for custom tables is this:
|StateCode Choices||StatusCode Choices|
|0: Active||1: Active|
|1: Inactive||2: Inactive|
Some tables will have different sets of choices.
|0: Open||1: Open|
|1: Completed||2: Made
|2: Canceled||3: Canceled|
The set of valid state codes for a table isn't customizable, but the status codes are customizable. You can add more
StatusCode choices for a corresponding
For custom tables, you can define more criteria for valid transitions between statuses.