Camunda BPM: User Task Assignment based on a DMN Decision Table

By
Process with Business Rule Task and User Task
  • Blog
  • >
  • Camunda BPM: User Task Assignment based on a DMN Decision Table
TOPICS

30 Day Free Trial

Bring together legacy systems, RPA bots, microservices and more with Camunda

Sign Up for Camunda Content

Get the latest on Camunda features, events, top trends, and more.

TRENDING CONTENT

In business processes involving human workflow the task assignment logic can become quite elaborate. For instance the processing of insurance claims, or other variants of approval processes, may require many or complex task assignment rules. The Decision Model and Notation (DMN) decision tables are an excellent tool to manage such rules — outside of code, in a business user-friendly way.

Camunda Release 7.13.0 will ship in Early June with the FEEL Scala Engine,
which already supports the DMN standard version 1.3, released in March 2020. Friendly Enough Expression Language (FEEL) expressions now allow us a more elegant mapping between decision table results and task assignment parameters.

Example Process

This example uses a trivial BPMN Process consisting of a business rule task, implemented by a DMN file, and a user task, which gets assigned based on the output variables of the rules task. The data object reference shown in the diagram is only for documentation purposes. It has no technical implications and can be removed.

Process with Business Rule Task and User Task

User Task

A Camunda User Task has three parameters relevant to task assignment:

  • assignee (a specific user who must perform the task)
  • candidate users (a list of specific users who can perform a task)
  • candidate groups (a list of user groups who can perform the task)

In this example these parameters get bound to a process variable using ${variable_name}. This allows us to let the task assignment depend on the output of previous elements (here the business rule task).

User Task Details: Assignee, Candidate Users, Candidate groups

Decision Table

The example DMN file uses three product attributes as input to determine the task assignment. The Hit Policy is set to Collect because the intention is to evaluate all rules and create lists of all possible task assignments. (The task assignment logic will later automatically assign the task to an individual user in case a specific assignee is set by a rule.)

Example DMN Table for Task Assignment

Business Rule Task

Due to the hit policy (C) the result will be a list of rules (rows) which fire for a given instance. As the decision table produces several outputs (columns) each item in the result list will contain a collection of up to three values (assignee, candidateUsers, candidateGroups). In the properties tab of the business rule task, the decision result needs to be set to reflect this expected data structure (a List of collections), as shown below. The result variable name, here dmnResult, can be chosen freely. The variable name is subsequently used in the mapping expressions.

Setting Business Rule Task Properties to store the Result

The final step is to map the output of the business rule task to the three variables used for the task assignment. This can be done on the “Input/Output” properties tab of the business rule task using feel scripts.

To determine the list of candidate users / candidate user groups we:

  • go through list of results (rows), picking the “candidateUsers” / “candidateGroups” value from each row
  • filter empty entries from the resulting list
    • (for r in dmnResult return get value(r, “candidateUsers”))[item != null]
    • (for r in dmnResult return get value(r, “candidateGroups”))[item != null]

To determine the assignee we:

  • go through list of results (rows), picking the “assignee” value from each row
  • filter empty values from the resulting list
  • pick the first assignee in the list (assignee can only be one value, not a list)
    • ((for r in dmnResult return get value(r, “assignee”))[item != null])[1]

Mapping of DMN result to process variables on the Business Rule Task

Value Mapper Configuration

CustomValueMapper registration in META-INFservices

For this approach to work, as of version 7.13.0-alpha4, Camunda still requires the registration of a customer value mapper. The mapper is registered by placing a text file called org.camunda.feel.valuemapper.CustomValueMapper
with the content: org.camunda.feel.impl.JavaValueMapper
into the folder: _METAINF/services.
This step will likely no longer be required in newer versions.

Unit and manual Tests

The project can be Unit tested as shown in AssignmentTest.java in a few tests for different task assignments. You can run it in your IDE or using mvn clean test.

To step through the process manually, you can start the server using your IDE or mvn spring-boot:run. After the server has started you can access the Camunda Tasklist via https://localhost:8080/app/tasklist/default. Use the credentials demo / demo to login.

The complete example project is available on GitHub:
camunda-consulting/code

Credits and many thanks to Philipp Ossler for developing the Feel engine and teaching me about the available FEEL expressions.

Want to learn more? Check out this blog for a better understanding of how FEEL 1.2 integrates with Camunda BPM and test it out with a complete example prepackaged as a Spring Boot application.

Camunda Developer Community

Join Camunda’s global community of developers sharing code, advice, and meaningful experiences

Try All Features of Camunda

Related Content

We're streamlining Camunda product APIs, working towards a single REST API for many components, simplifying the learning curve and making installation easier.
Learn about our approach to migration from Camunda 7 to Camunda 8, and how we can help you achieve it as quickly and effectively as possible.
We've been working hard to reduce the job activation latency in Zeebe. Read on to take a peek under the hood at how we went about it and then verified success.