That’s getting there. Thanks.
The following omissions at present all would require recursive parsing of the JSON returned value - which sort of defeats the purpose of having the graph engine add its value and smarts. Thoughts on these?
- depth
- cycle detection
- paths: paths are needed ANSI SQL in order to keep the parent/child rows together; Oracle’s CONNECT BY method doesn’t need this. As a graph engine, DGraph may also work more like Oracle’s implementation, keeping the parent and children together. But even so, generation of the path is very useful in many reporting outputs. In Gremlin this is done by the “simplePath()” step.
Over the weekend, I did some reading and have begun to see that the differences between a normal directed property graph and an RDF-kind of graph are substantial. I’ll ask on another thread about this so that discussion is buried here.
Thanks for the sample!