i get raw data quarterly. Essentially they are dumps from relational database tables. I import them into mysql bounce them against a reference database to get some validated columns then export the completed table with raw and validated columns.
right now i do it all manually via a workflow of the same sql queries over and over again. i would like to at minimum automate my mapping process but would also like the opportunity to build in analysis and possible breakpoints for custom transformation, etc. However, these are very large tables so the work should still be done by the database engine I would think.