DMN Expression Types Reference¶

DMN (Decision Model and Notation) provides a powerful set of expression types that can be combined to model complex decision logic. When defining decision logic in the DMN editor, you'll select from these expression types through the "Select expression" menu.

Expression Types¶

Icon	Type	Description	When to Use
*FEEL*	Literal	The simplest expression type for direct values or formulas using FEEL.	Simple calculations, direct value assignments, math operations
	Relation	Creates a table structure for organizing data (not to be confused with a decision table).	Reference data, lookup tables, structured data
{}	Context	Defines a set of named variables that can be referenced within the context.	Multi-step calculations, intermediate variables, complex logic
	Decision Table	Organizes conditions (inputs) and conclusions (outputs) in a tabular format with hit policies.	Multiple conditions/rules, business rule tables, if-then scenarios
	List	Creates an ordered collection of values or expressions.	Collections of values, multiple results, aggregations
*f()*	Invocation	Calls a Business Knowledge Model (BKM) or function defined elsewhere.	Reusing logic, calling functions, BKM invocation
f	Function	Defines a reusable function with parameters.	Reusable calculations, parameterized operations
if	Conditional	Creates branching logic with if-then-else statements.	Conditional logic, different outputs based on conditions
	For	Iterates over items in a list, applying an expression to each.	Transforming collections, iterating through lists
	Every	Tests if a condition is true for every item in a collection.	Validation rules for all items, ensuring conditions are met
	Some	Tests if a condition is true for at least one item in a collection.	Finding matching items, checking existence
	Filter	Filters a collection based on a specified condition.	Extracting subsets, finding specific items

Detailed Expression Type Examples¶

Literal Expression¶

The simplest expression type in DMN, used for direct values or FEEL expressions.

Base Premium * 1.2

Relation¶

The most widely used DMN expression type, with columns for input conditions and output conclusions.

State	Abbreviation
"Florida"	"FL"
"North Carolina"	"NC"
"New York"	"NY"

Decision Table¶

The most widely used DMN expression type, with columns for input conditions and output conclusions.

U	Credit Score	Risk Rating
1	>= 700	"Low"
2	[600..700)	"Medium"
3	< 600	"High"

Context¶

Defines key-value pairs that can be used for multi-step calculations.

{
  Base Amount: 100,
  Tax Rate: 0.07,
  Tax Amount: Base Amount * Tax Rate,
  Total: Base Amount + Tax Amount,
  result: Total
}

List¶

Creates an ordered collection of values or expressions.

[ "Red", "Green", "Blue" ]

Invocation¶

Calls a Business Knowledge Model (BKM) or function.

Calculate Discount(
  customer: Customer,
  order total: Order.total
)

Function¶

Defines a reusable function with parameters.

function(age, income)
  if age < 25 then income * 0.1
  else if age < 65 then income * 0.15
  else income * 0.05

Conditional¶

Creates branching logic with if-then-else statements.

if
  Applicant.age >= 18
then
  "Eligible"
else
  "Not Eligible"

For Loop¶

Iterates over items in a list or collection.

for item in Order.items return item.price * item.quantity

Every¶

Tests if a condition is true for every item in a collection.

every item in Order.items satisfies item.price > 0

Some¶

Tests if a condition is true for at least one item.

some item in Customer.productHistory satisfies item.category = "Premium"

Filter¶

Filters a collection based on a condition.

Order.items[price > 100]

Hit Policies in Decision Tables¶

Decision tables use hit policies to determine how rules are applied:

Symbol	Name	Description
U	Unique	Each row must be unique and only one rule can match (most common)
A	Any	Multiple rules can match but must return the same output
P	Priority	Multiple rules can match, but the output of the first matching rule is used
F	First	Multiple rules can match, but the first matching rule (by rule order) is used
C	Collect	Multiple rules can match and all are collected
C+	Collect (Sum)	Collects and sums numeric results
C<	Collect (Min)	Collects and returns the minimum value
C>	Collect (Max)	Collects and returns the maximum value
C#	Collect (Count)	Counts the number of matching rules

FEEL Operators and Functions¶

When writing expressions, you'll often use FEEL operators and functions:

Comparison Operators¶

=, !=: Equal, not equal
>, <, >=, <=: Greater than, less than, etc.
[..], (..): Inclusive and exclusive ranges (e.g., [5..10])

Logical Operators¶

and, or, not: Logical operators

Collection Operators¶

in: Membership test (e.g., x in [1,2,3])
satisfies: Test if collection elements meet a condition

Common Functions¶

count(collection): Count elements
sum(collection): Sum numeric values
min(collection), max(collection): Find minimum/maximum
substring(string, start, length): Extract substring
contains(collection, element): Test for element
list contains(collection, element): Test if list contains element

Combining Expression Types¶

DMN's power comes from combining these expression types to create sophisticated decision logic. For example:

{
  High Value Items: Order.items[price > 100],
  Discount Eligible: some i in High Value Items satisfies i.category = "Electronics",
  Discount Rate: if Discount Eligible then 0.15 else 0.05,
  Discount Amount: sum(for i in Order.items return i.price * i.quantity) * Discount Rate,
  result: Discount Amount
}

This context expression combines filtering, some, conditional, for loop, and mathematical operations to calculate a discount based on complex criteria.

Was this page helpful?