Decimal precision for numbers

The decision engine interprets any literal numbers that you use in your rules as double-precision floating-point numbers.

When you execute your rules with the decision engine, the engine interprets literal numbers as double-precision numbers. You must avoid the float type for numbers and use double in all circumstances.

If you use values of type float in your Business Object Model (BOM), they are likely not to match the literal values that are used in BAL rule conditions. This is also true for ruleset parameters and variables within a variable set, when they are compared to literals in decision tables or decision trees.

In case you cannot change your BOM to use the double type instead of float, you must symbolize your literal values. You can do so by introducing BOM constants of type float for each literal value in use. Instead of BOM constants, you might want to choose variables of type float, which you define in a variable set.