Will Qserv be storing nulls or defined default values for attribute values that have no measurement (eg non available or calculation produces NaN).
For our datasets it makes sense to try and replicate the same structure so that users get familiar with building queries and interpreting results. For our other data holdings we’ve typically avoided nulls and used default values.
Apologies if this is documented someplace, I tried a couple of searches. HSC/postgres seems to store NaN values in places but I assume that’s not possible in Qserv (mySQL)?