Thank you for your post! I agree - SQL is infrastructure and should be treated like that. The CTE tip is very important - specially in the early exploration phases - a well defined CTE can reduce the processing load and make the discovery phase more effective! One thing that I like about BigQuery is the estimated memory/processing load per query. It gives a tangible metric of many of the tips that you have shared. It is all good until your query stars becoming VERY expensive. Architecture first approach can save time, money and head-ache.
Thank you for your post! I agree - SQL is infrastructure and should be treated like that. The CTE tip is very important - specially in the early exploration phases - a well defined CTE can reduce the processing load and make the discovery phase more effective! One thing that I like about BigQuery is the estimated memory/processing load per query. It gives a tangible metric of many of the tips that you have shared. It is all good until your query stars becoming VERY expensive. Architecture first approach can save time, money and head-ache.