DMN Rule - Single Rule "Any of" Value Count Limit?

Is there any limit to the DMN values that are added into a single rule?

Could you have 200+ strings in a single DMN input rule?

Hi Stephen,

do you mean that you want to have a large input entry?
Can you please describe your use case?

Best regards,
Philipp

Hi @Philipp_Ossler

Yes i am speaking of the input entry.

You can add a do a string expression of multiple strings. Is there a limit of the number of strings? and is there a limit of a length of a individual string. (The length of a individual string is a edge use case, just curious).

I am looking at different scenarios based on the work from this thread: Pattern Review: DMN Looping for Array Input - #4 by Ingo_Richtsmeier

and given two competing scenarios, where i understand them to be the same from a execution standpoint? But may have different limitations, performance considerations, etc.

Considering two scenarios:

Scenario 1:

and Scenario 2:

My understanding is that these two would output the same set of information.

Is there a preference from a performance standard point?

From a usability view, I think scenario 2 is more realistic for end-user management when the list is large and complex names that are not a nice pretty incremental list. Especially if they manage it in Excel/CSV.

Hi Stephen,

there should be no limit of the number of strings or the length of a string. But at some point, it can be hard to understand the decision table easily :wink:

Regarding your scenarios, both are semantically equal and should have a similar performance.

Best regards,
Philipp

@Philipp_Ossler Thanks! Much appreciated as always!

Hi Philipp,

I wish to know if there will be a performance hit if I have 100K records in a single DMN.

What should be the split strategy or best practice?

Scenario Explained Below.

  1. In Scenerio2 as shown above, I have 100K records in the single DMN.
  2. The record towards the end of the table is the matching record.
    Will there be a performance hit in this case during decision making?

There is a similar thread here: How to process huge DMN file?

In general, a DMN decision table should be maintainable by humans (i.e. not more than 100 rules). If you have a huge number of rules or inputs then you should think about how to model the decision differently, and if DMN is the best option.

Hi Philipp,

Here I have a very limited set of columns(3 input columns only).
Since my application deals with global travel, I have a huge matrix that makes a large data set/ rules/ rows that needs to be put in this dmn. I have explored the DRG part of it, but am still unsure if it fits my my use case here.

@Satya_Kumar please share (parts of) your decision table. It would help to get a better understanding of your use case. Otherwise, it is hard to give any hints :slight_smile:

You can refer the benchmark for DMN.