I have some use cases which better fit graph db (neo4j), but most of the feature are handled in MySql.
I'm thinking of how to put a glue between both.
Would you duplicate some data to bind MySql with Neo4j, e.g nodes in neo4j keep reference to IDs from MySql or vice versa?
I'm using Spring and thought I'd use chained transactions to handle rollback gracefully of transactions which span both dbs.
For entities on the edge of both sql and graph worlds i plan to duplicate entities (not all properties, just the ones which are used in each db world + IDs)on both and on queries to join data on IDs.
In DDD it's a known technique to split domains and use separate DB for each domain, where certain DB technology fits better. Bit in my case it's really the same domain but different views.
I'm totally aware it all depends from the context and business use cases, but are you aware of some resources/examples of how people are doing it?
I know, "Google is my friend", but I had difficulties to find useful resources.