How can I create a decision table in Red Hat Decision Manager with only metadata information about the rule

Question

I am trying to map a set of rules defined as decision tables in FICO rule engine to Red Hat decision manager. The existing decision table contains metadata information about the rule and the actual rule construction happens inside a 'function'. My sample DT looks like:

Id	Field Name	Field Length	Is Manadatory	Message Length	Message Mandatory
1	a	5	true	"Invalid length	"Field is mandatory"

This Decision table is used to evaluate a form field in UI. Upto column 4 are metadata information about that form field such as datatype of the field, minimum length,isMandatory etc. From 5th column are the actions which has the validation message to show if the conditions are not met (eg if user is giving a wrong data type, a validation message should come as "Invalid Data Type"

Since this table has only metadata about the rules(not the actual rules), the rule construction based on these columns happen inside decision service. How can I achieve the same using Redhat Decision Manager?

Answer 1

You could use these "validation table data" to build something akin to the kie-dmn-validation rule base: example.

In your sample, you can consider translating your data into a rulebase something ~like:

rule r1mandatory
when
  $m MyDomainModel( a == null )
then
  reportError("Field a is mandatory in "+$m);
end

rule r1length
when
  $m MyDomainModel( a str[length] < 5 )
then
  reportError("Field a Invalid length in "+$m);
end

The way you generate the rules from your data, you could use FreeMarker/Velocity to generate DRL, or directly generate the Java Exec Model equivalent directly.

In other words, you need a way to "template" from your validation data, to some rules, as hinted above for their structure. There are several way to achieve it, as explained.