分布式存储过程 Neo4j

Question

I have some experience with Neo4j, but now we are discussing both the ideas of employing custom stored procedures and a sharded Neo4j database (so using Neo4j Fabric) that I both haven't used, so far.

What I have been searching for a while on the internet but couldn't find an answer to is the actual possibility of mixing a custom stored procedure with a sharded database. I guess the procedure would have to be stored in the Fabric instance to be able to work with the whole graph. If that is true, still - is this even possible? Can I write a stored procedure and run it on a sharded database via the Neo4j Fabric?

If so, would it be automatically handled in a distributed transaction, or is it something I'd have to take care of?

Thank you very much, If I missed a source that talks about it. I'd be happy to be redirected. Or even tell me I just talk rubbish if you think so.

Answer 1

Good questions

Fabric itself doesn't execute any Cypher on a physical database it just distributes the queries and aggregates the results.

The procedures that access graph data from the store will have to be installed on all shards.

And then can be called on each database to be queried.

You can also install procedures on the fabric proxies but those can only be used in the parts of the query that is actually executed on the fabric instance, ie outside the CALL { USE... } subqueries.