Defining Datatypes

User-defined datatypes, called item definitions, are created from built-in types by adding constraints or components. Each item definition has a unique name in the model namespace.

There are multiple ways to add a New type to your model:

Any of these present a dialog where you enter the type name, click the link icon at the right and select New type….

This assigns the default type Any, which you modify by again clicking the link icon and selecting either a Simple type or a Structure.

To change the data type of a type or one of its attributes, clicking on the icon will offer different sources for the type.

A new simple type is typically a FEEL base type with added constraints, i.e. limitations on allowed values. To define a constrained simple type, first select the base type from the Simple dropdown and then click on the base type name, such as Text, Number, Boolean, etc. Then select the Constraint type.

The following type of constraints are available

An "Allow null" checkbox can be checked for each constraint except the "Not Null" one.

Clicking the "Add Expression" button allows to enter an additional expression constraint that must be met along with the main one. To delete it, click on the at the top right. Changing the constraint type to "None" will remove the main constraint. If there are additional expression constraints, the first one will be set as the main one.

To refer to the data type itself in an expression, use the question mark "?".

An Expression constraint is a Boolean FEEL generalized unary test expression, meaning any literal expression in which the constrained variable is represented in the expression by the ? character. Only values for which the expression is true are allowed. For example, to define tInteger as an integer, use the base type Number and the Expression constraint

? = floor(?)

Since the FEEL function floor() extracts the integer part of a number, this expression is true only for integers.

Siblings (same level) of a data type can be referred in an expression using its name.

When a type is a collection, collection constraints can be defined by clicking at the icon in the namecell. The constraint will apply on the collection itself.

The following type of constraints are available

Optionals validation code and validation message can also be entered for each constraint. It allows to show a custom error message when the constraint is not respected.

Clicking the "Add Constraint" button allows to enter an additional constraint that must be met along with the main one. The constraint could be Expression, Size or Not null ( Size and Not null are limited to one of each for a collection constraint). To delete it, click on the at the top right.

To refer to the collection itself in an expression, use the question mark "?".

Since DMN 1.5, it is possible to report validation errors at execution time when a variable value does not conform to its type. This option can be turned on or off in File/Preferences.

In the Constraints dialog, you have the option to enter a Validation code and Validation message for each constraint. Both the code and message are returned in the Test action of the Execution ribbon of Decision Modeler if a variable does not conform to its constraint, and are returned in the error message of a deployed decision service as well.

A FEEL variable may be defined as a data structure containing components. Each component has an assigned type, which could be simple, a collection, or a nested structure.

When adding or editing a structured type, the dialog below defines the structure. Numbered rows define the name and type of each component.

To add a new component to a structure, press the icon at the end of the last component.

Datatype details can be expanded ( ) or collapsed ( ).

Attributes can be reordered using drag and drop.

Right-clicking on an attribute (or using the icon) brings up a contextual menu with localized actions.