Major Description
Following on to the issue I mentioned CRUNCH-458, I'd like to allow null as a valid value in the "value" side of the key-value pair in an Avro-based PTableType. Allowing null-valued keys would be a bit more work and felt more dangerous to me (i.e., the logic for things like sorts and secondary sorts would need to change), but there shouldn't be an issue with allowing nulls as values.