2 of 2 people found this helpful
Rather than worrying about this at the AI stage, why not let the recon.job take care of the null values? So as the AI loaded data moves into BMC.ASSET (via a recon. job) you can tell it to not merge NULL values coming in from AI if BMC.ASSET has a non-null value (i.e. you can preserve the existing non-null values). This is the Defer If Null setting on the Recon. merge activity.
Agree with Carey,
You need a 3 step reconciliation here.
Step 1 - Dataset where AI is putting data, in this dataset data may or may not be null, we don't really care. This data is then merged to Step 2 using the DEFER if NULL option, thus making sure we only update non-NULL data that came from AI.
Step 2 - Dataset where we have the TRUE data coming from the source. Here we have good values, and a NULL in here means we really do not know what this data is about. NULLs are important in this dataset as well as all the other data. This dataset will be merged to Asset.
Step 3 - ASSET Dataset.
So basically you need two reconciliation jobs and 2 datasets to achieve this.