Numerous SQL-on-Hadoop, NewSQL and NoSQL databases provide semi-structured data model and query language capabilities, but it is difficult to compare these capabilities. Many differences between the data models and (especially) between the query languages are superficial, but nonetheless distract from the essential differences. Other query language differences are direct derivatives of corresponding data model differences. Finally, many query language differences correspond to fundamentally different query capabilities of the databases. To focus on the fundamental differences of the surveyed databases, while avoiding the distraction of superficial ones, the authors present a two-step systematic classification and comparison of their data models and query languages.