Hi,
Is there any recommendation on limit of rules in a single decision table? What if i’ve 50,000 rules along with 10-15 input, 5-8 output columns? I hope that will consume too much memory if i run an embedded DMN engine. What are the options i’ve? can’t i run embedded engine in this case?
Hi,
No specific recommendation, except the more there are, the more memory etc.
Do you actually need 50k rules in one set of rules? It’s not a manageable number of rules for a human to work with, so may be they’re generated rules? Also, it’s unusual for that many rules to be so interconnected that you’d actually manage them as a single ruleset. Usually you find that there are more natural sets of rules that come to intermediate conclusions and then another st of rules become appropriate. You can break things down into separate DMN tables that may be more manageable for maintenance as well as execution. You can user DRDs to bring outcomes from different decision tables together, or just sequence with a process.
Cheers
Paul.
Hi Paul,
Thanks a lot! I’ll give DRDs a try.