[tabtab]

Most NOSQL databases are non-relational such that you have to deal with "impedance mismatches". They can have duplicate columns and duplicate rows, for example, which SQL itself is not designed to handle without adding ugly kludges. It's even messier for writes. I didn't mention it due to posting size limits, but each Dynamic Relational table has an auto-generated "ID" column so that each row is unique. Good question!

[PaulHoule]

see https://www.arangodb.com/ and also http://www.w3.org/TR/sparql11-query/

[tabtab]

I'm not sure what those solve. Dynamic Relational (DR) is relational so that one can use SQL on it without big impedance mismatches. Your links appear to be tying to invent a NON-relational version(s) of SQL, but they are too different from SQL and existing RDBMS to be easy to learn.

DR's goal is to match/fit relational, existing RDBMS, and SQL as much as possible, yet still be dynamic. Non-relational databases and non-relational query languages probably will fail at that goal. They deviate from RDBMS too much because they are changing too many things. DR only deviates just enough from traditional RDBMS to get dynamism so that users of RDBMS can easily learn/use it withOUT tossing out most their RDBMS, relational, and SQL knowledge. It just reinvents one nation, not the entire database world.