3 ms·

I hate these types of posts. They do nothing to help the graph database space. How good are they at bidirectional traversals, what about deep traversals in comp

I hate these types of posts. They do nothing to help the graph database space. How good are they at bidirectional traversals, what about deep traversals in complex networks with more than 50 relationship types? What about x? Every vendor is going to be better at one thing than another, there is always a trade off, but it is irrelevant. Last week we took a very deep and very wide sales hierarchy query from 10 seconds in Oracle to 4ms in Neo4j. If Arango could it in 3ms, or Orient in 5ms doesn't really matter. The point is graph databases are much better at some queries than relational. They should be blogging about use cases and customer successes, not benchmarks.

ifcologne 9y ago
>They should be blogging about use cases and customer successes, not benchmarks. They do, just recently the cases of AskBlue (Location aware recommendations) and Thomson Reuters (Fast & Secure Single-View of Everything) https://www.arangodb.com/why-arangodb/case-studies/ https://www.arangodb.com/why-arangodb/case-studies/ I am very interested in Graph-Cases and appreciate the support of Neo4J for projects like the Panama- or Paradise Papers that show how connected graphs help to understand former hidden relationships. Graphs are very powerful and a great addition to relational or document approaches. (full disclosure: I worked for ArangoDB 2 years ago.) And, I've used Neo successfully for a fraud detection PoC recently...