The trade-offs involved in using abstracted data in relational databases versus non-abstracted data in non-relational databases, relying on pre-abstraction versus self-abstraction of high-volume data streams, and so on are not obvious, will change as technology involves, and will depend on the company and the problem.

FORBES: Three Ways To Make Big Data Make Money

The Cascalog system has combined the Cascading system for abstraction of advanced data processing on Hadoop and other systems, the Datalog declarative programming language that is well suited for expressing database queries in an abstract form, and the Clojure language to create a declarative language for queries that can be run on Apache Hadoop.

FORBES: Why Adopting the Declarative Programming Practices Will Improve Your Return from Technology