Re: Creating the database
toggle quoted messageShow quoted text
I totally agree with this, and some of what the OP said too.
While it's not a perfect analogue, any book on (my/ms)sql (or similar) would have entire chapters early on about creating and connecting to databases (graphs). All the info I find on graph databases is either about traversals, connecting to DBs/indexers, or very low-level stuff for advanced users. I've spent more time on this than I care to admit (PAID WORK time, not spare time) and most of that time is going around in circles.
It's almost like everyone has their server of choice (janus or whatever) set up using whichever single graph the server is set up with. I have managed to create a new graph via the console, and have altered the config to get a traversal source from it on startup, but this is hardly dynamic. I want to be able to created named, empty graphs on demand, and select them arbitrarily, both in and outside the console.
While Kelvin's online book is excellent, he also confuses me by showing how to open a graph in Janus, but not showing anything about naming or persistence (how would I get access this new graph in the future? How do I get back to the air-routes graph again that was also opened, populated, but seemingly unnamed? That was assigned to graph too). ConfiguredGraphFactory appears to be the only way to do this(?)
I note the link to the graph of the gods setup source code. This could be useful, but I'm using gremlin.net which doesn't seem to have Factory classes and only lets me connect to existing traversal sources (which I have to create manually on the console etc etc). I'm wondering if I'd be better off running the server as a REST API if I could use it for open/create commands as well as traversals.
I'd like to end on a positive note, so I'll say that in fairness, once I have connected to a graph, I have no issues with traversals, and there seems plenty of info on these.
On Thursday, 22 March 2018 15:53:38 UTC, Kaidong Wu wrote: