Hello! 👋 First of all, thanks to everyone working on Airbyte...I just discovered it, and it has been really awesome so far! I have a question about a Shopify integration, in particular, but the question applies more generally to how load and transform happen. I am trying to load Shopify orders + order_refunds + transactions into our BigQuery instance, and I was hoping to get the orders into one table (i.e. not normalized); however, when I turn off the ‘normalization’ option in the UI, I only seem to be able to see the JSON blob in _airbyte_data. Is there some middle-ground option where the data can be spread into columns in BigQuery, but not completely normalized into separate tables? I did notice this documentation here, which was interesting: https://docs.airbyte.io/tutorials/connecting-el-with-t-using-dbt Is the ‘transformation’ step something that we, as the consumer, are responsible for? It would be awesome if the data could fit into one table, given that BigQuery supports the RECORD type, so the nesting works pretty well in that destination

@Charles Lariviere hello! if i understand correctly you want to essentially have all the top-level fields written into BQ with the correct types/columns, but nested fields should be in their own column as JSON blobs?

Hi @s! Actually, I was thinking more of having the nested fields be loaded as the RECORD or REPEATED type in bigquery….we do this currently for nested JSON for various data sources that we were trying to port over to using Airbyte

Docs on what I’m referring to here: https://cloud.google.com/bigquery/docs/nested-repeated

The BigQuery schema structure(s) that we have map exactly to the imported JSON…all made possible by these RECORD types, basically

So there are no JSON blobs, in the end for us. I’m happy to expand further!

@Charles McMillan if i follow you, i think you want to subscribe, comment or react on this issue to add your requirements there: https://github.com/airbytehq/airbyte/issues/1927#issuecomment-779291005