As a reminder, Looker has three core components: first, it provides a physical modeling layer (LookML), which one uses to abstract SQL generation and define complex transforms; second, Looker consumes the LookML model and generates dialect-specific SQL, which it then issues to the database via JDBC; third, it provides a web-based UI which acts as both a LookML IDE as well as the primary means of exploring data.
Let's consider a simple schema containing two tables, job
and warehouse
, which we'll use as an example. In Looker, our representation of these tables would look like this:
- view: job
sql_table_name: public.job