我应该使用Neo4j寻找类似用户还是坚持使用像MySQL这样的东西？

Question

I'm starting a new project and although I'm used to MySQL, I'm worried about efficiency. I'm open to other options, and graph databases sound intriguing.

I will need to find similar users based on location and rating like values. In mysql I probably would have to join across 2 many to many relationships and order based on distance of both location and those values (euclidean distance probably). MySQL seems slow with things like that.

I will also need to do things like find 10 nodes with text that starts with a sub string, and has the largest number of connections (which is an autocomplete I guess).

Would Neo4j or another graph database do this easily and efficiently?

Answer 1

Yes, Neo4J is certainly more appropriate than MySQL. I've used it myself for similarity searches and continue to do so. Check out Cypher , or Gremlin depending on how complex your criteria are -- together with the inbuilt Lucene index, it's terrific.

Examples of what you may be trying to achieve: http://docs.neo4j.org/chunked/stable/data-modeling-examples.html