Dialects

FEEL has two entry points in the DMN spec. parse() accepts a dialect parameter to select the correct grammar.

'expression' (default)

Parses a general FEEL expression. This is the mode used for:

• Input expressions in decision tables
• Output expressions
• Decision logic (Boxed Expression)
• Context values
• Parameter defaults

parse('if score >= 700 then "approved" else "declined"')
parse('for x in 1..5 return x * 2')
parse('some o in orders satisfies o.amount > 100')

The expression dialect supports all 24 AST node types.

'unary-tests'

Parses a comma-separated list of unary tests — the entries you put in decision table input cells.

parse('"A","B","C"', 'unary-tests')
// UnaryTestList { negated: false, tests: [StringLiteral{A}, StringLiteral{B}, StringLiteral{C}] }

parse('[100..999]', 'unary-tests')
// UnaryTestList { negated: false, tests: [RangeLiteral{100..999}] }

parse('not("High","Medium")', 'unary-tests')
// UnaryTestList { negated: false, tests: [UnaryTestList { negated: true, tests: [...] }] }

parse('-', 'unary-tests')
// UnaryTestList { negated: false, tests: [] }  ← wildcard (any)

Unary test operators

Inside a unary-test list you can use prefix comparison operators:

parse('> 100, <= 200', 'unary-tests')
// tests contain BinaryOp nodes with the implicit input value as left operand

Wildcard -

A single dash means "any value passes" — represented as UnaryTestList { tests: [] }.

When to use each

Context	Dialect
Decision table input cell	'unary-tests'
Decision table output cell	'expression'
feel-runner unaryTest()	'unary-tests'
feel-runner evaluate()	'expression'
BPMN condition expression	'expression'