There are three ways you can test a decision table.
-
The unit test is similar to that of the rule set. You have to give input for each column and then click run test.
-
The object test is also similar to the rule set. You need to select a data object and then run a test.
-
And then there is Simulation.
Simulation
Simulation is useful for checking that there are no gaps in decision logic. A gap is when there is no result for a specific segment. Sometimes a part of a segment or a whole segment can be without an assigned result from the decision table. That can happen when a combination of values is missing.
To prepare the definition, you will see a list of columns (attributes) of your decision table. The simulation window is titled "Synthetic data definition". Each condition column has its own field. Use the small drop-down arrow at the right end of the field to switch between "Interval" with a start and end and "List" with values typed one by one. A column with no definition is left out of every record. The engine will use the definition to create a sample of synthetic data based on the definition.
The "Sample size" is a drop-down with 10, 100, 1000 and 10000. For an interval, each record gets a random number between the two ends. If both ends are whole numbers, the generated numbers are whole numbers. Every value is equally likely. If the ends are in the wrong order, they are swapped. If one end is missing or is not a number, the column is left out of the record. For a list, each record gets one listed value at random, with every value equally likely.
When simulating, the engine will create random combinations using your definition. It will use the defined columns to create a sample. The engine will then run the sample through the table and will show you summary results. Select "Run sim" to start the simulation and view "Simulation results".