I’m seeing work on the BigQuery denormalized connector recently that uses a value of

big_query_array

to account for nulled arrays. This seems to then be adding this value of

big_query_array

into table schemas. Was this necessary to handle nulled arrays? A large value proposition of the Airbyte BigQuery-denormalized connector, was that it was able to match schemas directly to JSON coming out of the source API. I’m concerned interjecting this hard-coded value into the created schemas causes many more problems downstream than it solves upstream. (edited for clarity/tone)

@Zach Brak do you mind creating an issue reporting your concern about this in Github?