on compiling DMN file in Kogito (Quarkus) the compile time is very long more than 15 minutes. Where I am getting it wrong.
The DMN is attached herewith
I was expecting it to be less than a minute for this DMN
on compiling DMN file in Kogito (Quarkus) the compile time is very long more than 15 minutes. Where I am getting it wrong.
The DMN is attached herewith
I was expecting it to be less than a minute for this DMN
As explained in this post, the reported long time is not related to any compilation, but to the decision table validation capabilities, specifically due to the non-methodologically compliant decision table in question.
In general for Drools and Kogito, the Decision Table Analysis is configured by default to attempt Overlap analysis even when you are using string values which should have been enumerated for your columns. In your case, it seems to me you have not enumerated types for Dealer.IsPolicyIssuingDealerCode
, Vehicle.RTOCode
, Vehicle.ModelCode
, Vehicle.VehicleType
.
Looking at the table, it looks to me that decision table is also failing methodology-wise, especially for columns Dealer.IsPolicyIssuingDealerCode
and Vehicle.ModelCode
.
In my view, you have 2 options.
Option1.
If you want to follow methodology best practices, likely values in Dealer.IsPolicyIssuingDealerCode
ought to be represented in a subdecision, something ~like "is special dealer code", so that would transform into a boolean check in your current table.
Then, Vehicle.ModelCode
can also be refactored in another subdecision, for "PCP" classfication and segmentation in their own categories, looks to me you have 3 of them; so that would transform into 3 different category values in your current table.
You can find example of categorization in this webinar, video recording is also available: https://blog.kie.org/2021/08/how-to-capture-business-decisions-using-dmn-introduction-to-some-basic-patterns-and-their-value.html#:~:text=PATTERN%3A%20CLASSIFICATION
Option2. If you want to keep the decision table as-is, despite lacking DMN methodology best-practices, you can always disable validation. Ref.: https://docs.kogito.kie.org/latest/html_single/#con-dmn-validation_kogito-dmn-models:~:text=Configuration%20to%20disable,decisions.validation%3DDISABLED
I'm glad to hear the explanation in the linked blog post was helpful.