Forum Discussion

AndreaF's avatar
AndreaF
Contributor III
2 years ago

Exclude members from rule expression

Hi, I have two sets of derivative source rules (see below).  What I am looking for is to exclude the Accounts starting with 9203 from the second set of derivative source rules.   Each deriv...
Transformation
  • JackLacava's avatar
    JackLacava
    2 years ago

    I'm not good at Transformation rules, but I don't think you can do a difference easily - you'd need a buffer, which is only available on BiBlend derivatives. However, it might be enough to place something like this on your 92* line, to nullify any value coming from 9203:

     

    if args.GetSource("A#").StartsWith("9203") then
   return 0
end if
return args.ColumnValue

     

     

     

MarcusH's avatar
MarcusH
Contributor III

In situations where it is difficult/impossible to reference the source members, I use another UD as a helper. Typically there is a UD that is not used for the data import and load (if there are no free UDs then enable an Attribute or Label dimension). What I do then is apply a BR on the data source for this working UD - let's say it is UD7. In your case the BR will set the value of UD7 to be 9203 for accounts that begin with 9203 and then 92 for all other accounts beginning with 92 (simple to do in a BR). Then in the Rule expression you add .U7#[9203]=None for HoldCo and .U7[92]=None for Loans. The transformation rules for UD7 map everything to the valid UD7 member.

I find that derivative rules are hard to understand in the first place. You could possibly do it by changing the Logical Operator and Derivative Type somehow but I prefer to keep the process of creating extra data records as transparent as possible.

HTH

AndreaF's avatar
AndreaF
Contributor III
2 years ago

Thank you, all of 3 are good ideas! The one I ended up implementing is Jack's solution, setting the derivative rule as complex expression and using the proposed formula. Only downside is that the rule creates zero rows that are not really needed: each [92*] row creates a row with the value of ([92*] excluding [9203*]) and a zero row (because of the Return 0 in the [92*] formula), but that should not be an issue since the amount of zero rows is limited.

  • JackLacava's avatar
    JackLacava
    Honored Contributor
    2 years ago

    I think there might be ways to avoid that too - try returning None or False instead of zero, or throwing an exception.