SQL can be used as a query layer outside of an RDBMS. Spark is a good example of this. RDBMS as a class of databases has nearly as many variations as “NoSQL”.

SQL can be used as a query layer outside of an RDBMS. Spark is a good example of this. RDBMS as a class of databases has nearly as many variations as “NoSQL”. So presenting SQL and RDBMS as the same thing or implying NoSQL is a single thing is oversimplifying to a point that any interesting nuance is lost. You need to pick a database that suits your need but to make that choice you’re going to have to look deeper than “RDBMS vs NoSQL”. Look at concepts like key-value stores, document databases, graph databases, shared nothing architecture, shared everything architecture, data normalization, CAP, and ACID. This is where the meaningful tradeoffs happen.