Repository navigation
Feature completeness of different protocols etc. #4472
|
Hello @lvca @robfrank and community! @venkatramanam and myself are from IBM and are currently evaluating ArcadeDB as a Neo4J alternative for our Graph Traversal needs. Firstly, let me applaud you for keeping the project under Apache 2.0 license forever 🫶 I have been scouring over the docs to understand which query language over which protocol would suit our needs the best. However, I found through the documentation that perhaps not everything might be feature complete. e.g. snippets from docs.
I am just looking for some guidance here as to what would be a good combination of Query Language + Protocol to begin with? I have seen the arcadedb-usecases repository and most examples are in ArcadeDB dialect of SQL. Would that be the best choice here? Getting devs to learn yet another sql might be challenging though :) |
Replies: 1 comment 5 replies
|
Hi, and welcome aboard! Right now the top 2 query languages are SQL and Cypher. If you are going to do traversals, I'd suggest to go with Cypher: in ArcadeDB is now native (zero conversion) and super fast. On top of that, you'll find tons of docs about Cypher and Open Cypher on Neo4j and other websites. I suggest SQL only if you're coming from OrientDB or it you really hate Cypher syntax :-) |
Hi, and welcome aboard!
Right now the top 2 query languages are SQL and Cypher. If you are going to do traversals, I'd suggest to go with Cypher: in ArcadeDB is now native (zero conversion) and super fast. On top of that, you'll find tons of docs about Cypher and Open Cypher on Neo4j and other websites.
I suggest SQL only if you're coming from OrientDB or it you really hate Cypher syntax :-)