Yep, we went there. You can now select "Dataframes" as a data source for SQL cells, and query any dataframe in a project like a table in a database. Queries are fast, in-memory, and use a well-featured flavor of PostgreSQL.
You can use Dataframe SQL to break up large queries into modules that are "chained" together, join tables across databases and CSV sources, or just pop back into SQL halfway through an analysis to do a simple filter or case statement. You can mix and match SQL and Python on the same data as many times as you'd like!
There's a lot to say about this one, so we wrote an entire blog post with the "why" behind the feature. Just in the mood to get started? Check out the docs for the full details and jump right in.