Using the released version of FIM 2010 R2. (Don't know if FIM 2010 functions the same or not.) I have a scenario that reads If MVattribute1 = "somevalue" then flow MVattribute2 otherwise flow NULL to MAdestinationAttribute. NULL means <not set> if viewed in ADSIEdit. I understand the functioning of the Null() function that it doesn't flow a null value. I have created a nullString attribute bound to the person object in FIM Metaverse. I am using only declarative sync rules with outbound scoping filters. As reference http://social.technet.microsoft.com/Forums/en-US/ilm2/thread/9c0ac6ee-4795-4aef-96f1-663cb50dc7d2 contains a similar discussion about FIM 2010 that suggests IIF should work. In the sync rule I created a function IIF(Eq(MVattribute1,"somevalue"),MVattribute2,nullString) -> MAdestinationAttribute This rule works just fine when the condition is met, and MVattribute2 will flow to the destination. However when the condition ceases to be met it will not flow the nullString attribute to the destination; the previous value from MVattribute2 remains. I know the rule is constructed correctly because if I actually put a value in the nullString attribute, that value will flow. But that obviously defeats the purpose of needing to flow a NULL value. So why isn't IIF flowing the nullString attribute simply because nullString happens to have no value? If I do a direct flow of nullString to destination it will also succeed. It's just when IIF is being used that the NULL value won't flow.  

Hi David, I've ran into this in the past. I believe the hack that I found to work was to add an empty string to the nullString in your custom expression. For whatever reason, that seemed to actually flow a Null ... So, try changing the flow to IIF(Eq(MVattribute1,"somevalue"),MVattribute2,nullString+"") -> MAdestinationAttribute Good luck, Marc Marc Mac Donell, VP Identity and Access Solutions, Avaleris Inc. http://www.avaleris.com

Thanks for the suggestion, but adding the + "" doesn't make a difference on FIM 2010 R2. It still won't flow the null MV value as a CS attribute delete. It generates no export at all so whatever happens to be the previous value remains in the CS.

I know a few PFEs who are saying this is likely a regression error in R2 RTM.