Custom Operators

EvalEx comes with a predefined set of ready-to use operators. A standard collection is added to the default configuration, using the ExpressionConfiguration.StandardOperatorsDictionary.

Operation definition is made in two parts:

  • Defining the operator type and logic by adding a class that implements the OperatorIfc interface.
  • Adding the operator with a unique name to the operator dictionary, so it can be found and used in expressions.

Defining the Operator

To ease some common implementation routines, a custom operator usually extends the AbstractOperator class. As an example, we can look at the boolean “GREATER” operator:

public class InfixGreaterOperator extends AbstractOperator {

  public EvaluationValue evaluate(
          Expression expression, Token operatorToken, EvaluationValue... operands) {
    return expression.convertValue(operands[0].compareTo(operands[1]) > 0);

As you can see, the only method that has to be implemented, is the evaluate() method. It will be called by EvalEx during evaluation. It will pass the operand values to the method.

The expression itself and also the operator token are passed to the call, they can be used to access the configuration or to find out the operator name and its position in the expression.

Operator type

The operator type is determined through the corresponding annotation:

  • @PrefixOperator
  • @InfixOperator
  • @PostfixOperator

Infix operators will receive two operands in the evaluate() method, pre- and postfix operator will receive one operand.

Operator Precedence and associativity

The order of expression evaluation is determined by the operator precedence and its associativity.

The precedence tells EvalEx which operator have to be evaluated first. With this, it is e.g. defined that multiplication happens before addition.

The associativity tells in which order (left to right, right to left) operators of the same precedence are evaluated.

Precedence and associativity can be specified with the operator annotation.

There is a collection of predefined operator precedences in the OperatorIfc interface.

Lazy Operands Evaluation

Infix operators can optionally be defined to allow lazy evaluation. Without lazy evaluation, the value received by the operator would already have been evaluated.

Lazy evaluation can be helpful for certain situations where sometimes you may want to skip part of the expression without affecting the result. One example is where you want to implement short-circuit evaluation. Consider an expression “a != NULL && a > 0”, in case “a” has a NULL value, the right side of the expression can be skipped or an error will be thrown saying NULL is not comparable.

Note that currently only infix operators allow lazy evaluation. The “AND” operator demonstrates this:

@InfixOperator(precedence = OPERATOR_PRECEDENCE_AND, operandsLazy = true)
public class InfixAndOperator extends AbstractOperator {

  public EvaluationValue evaluate(
          Expression expression, Token operatorToken, EvaluationValue... operands)
          throws EvaluationException {
    return expression.convertValue(
                    && expression.evaluateSubtree(operands[1].getExpressionNode()).getBooleanValue());

Adding the Operator

You can always add the operator directly to the operator dictionary, using the addOperator(String operatorString, OperatorIfc operator) method.

But there is an easier way by using the expression configuration, which also allows you to add more than one operator:

ExpressionConfiguration configuration =
            Map.entry("AND", new InfixAndOperator()),
            Map.entry("OR", new InfixOrOperator()));
Expression expression = new Expression("(a > 5 AND x < 10) OR (y < 0)");

Copyright © 2012-2022 Udo Klimaschewski