We have hit a situation where allowing passthrough of null values from SQL to CSV has resulted in the state of 'this field is blank' vs 'this field is ommitted by design as part of being a powerset' has resulted in these two fields being indistinguishable from each other. Since the dashboard uses the hash of all non-count fields as a primary key, it rejects this data from being uploaded.
We need to decide on a way in which to represent data that is null inside of count tables, and then apply that across the board, to ensure that the dashboard can correctly handle this. It would be nice to handle #131 at the same time and eliminate all static count tables as part of this, but it is not required.
We have hit a situation where allowing passthrough of null values from SQL to CSV has resulted in the state of 'this field is blank' vs 'this field is ommitted by design as part of being a powerset' has resulted in these two fields being indistinguishable from each other. Since the dashboard uses the hash of all non-count fields as a primary key, it rejects this data from being uploaded.
We need to decide on a way in which to represent data that is null inside of count tables, and then apply that across the board, to ensure that the dashboard can correctly handle this. It would be nice to handle #131 at the same time and eliminate all static count tables as part of this, but it is not required.