There are a couple of phases in the ETL process when an analyst might want to transform their data. BigQuery is a very sophisticated data warehouse system. Because of this, BigQuery has already implemented a number of functions useful for transformation. With that said, BigQuery does not have every function that you might find in other data warehouses or programming languages. Also, BigQuery's automatic type detection sometimes might coerce a data type with unwanted results. Because of this, the analyst will need to make a decision as to how much transformation he or she wants to apply to the data prior to loading it in BigQuery.