Transformational

It all begins with SELECT. Modeling through transformational logic is a powerful and highly maneuverable method for modeling data that comes with one serious drawback: it needs existing data to SELECT from. Transformational modeling is rarely done in transactional databases because, in such systems, data is created and modified through the company’s operational processes (e.g., purchases and profile updates)—with which expensive transformational processes should not compete for critical system resources. However, in data warehouses, where conformed datasets are extracted and duplicated with fresh timestamps for each load, transformational modeling becomes a necessity.

Because transformational modeling selects from existing structured data, the result set is already structured. Selecting the SUPERHERO_NAME and HAS_MASK columns and creating a table will preserve their structure (VARCHAR and BOOLEAN, respectively). However, as with all modeling, transformations...