-
The issue you’re experiencing with missing NULLs in columns in the downloaded JSON is due to how JSON handles NULL values. In JSON, if a field has a NULL value, it’s often simply not included in the output. As a result, when you download your BigQuery results as JSON, columns that only contain NULLs seem to disappear. As for the changing order of columns, JSON objects are an unordered collection of key-value pairs. Upon re-importing your JSON data, BigQuery determines the column order based on the order it encounters the keys in the JSON file. One workaround to maintain a steady column order could be to create a view in BigQuery with columns in a specified order and then upload your JSON data into that view. This problem has much to do with the way BigQuery deals with data imported from JSON files. To maintain the same structure, some users apply extensive preprocessing before uploading it to BigQuery or utilize an intermediate system that is able to better handle JSON data.
