@Chris_Knoll I really think for our CDM purposes at Janssen most of what we have are unmapped values; we very rarely get any NULL values from our sources. As you know, many of our data assets are US based claims or EHR systems which are centered around billing and it is extremely unlikely that we will see a NULL value anywhere in a field that relates to how a provider bills for a claim. The only place we ever really see NULLs would be somewhere related to an actual billable field, like measurement units or things like provider specialty, etc.
One of the benefits of our current approach to using 0s for NULLs and unmapped values is that we don’t have to tell the builder what to look for when mapping to zero, we can just say “map everything you can and anything you can’t, map to 0”. If we were decided to map all NULL values from the source to a NULL concept_id, we would then have to parse through the source each time we received an update so we could accurately tell builder what is a NULL and what isn’t. Considering the infrequency with which NULLs are actually present, this seems like it would take more time than it would save us in the long run.