We are planning to migrate Data from AuroraDB to AWS neptune (property Graph). we have table in AuroraDB with 3.5 million records, and which has relationship with other five tables. if I have to create all the records as Vertices result in 3.5 million vertices and it has relationship with other five table, each record will have 5 Edges for 5 other tables, then the edges will result in 3.5 million (records in one table ) * 5 (each records have relationship with other 5 tables) results in above 10 million.
How we can reduce this Edges and is there anything i am doing wrong?
In general it is reasonable to consider a row in a relational table as a vertex in a graph where each column becomes a property. It is likewise reasonable to consider converting what would be joins in the relational model to edges in the graph model. A graph with 3.5 million vertices and ~10 million edges is not that big, so in terms of size nothing of concern. However, data modelling generally works best if the approach is not just to mirror the exact relational model as a graph but rather to consider the types of questions (queries) you need to the graph to answer. It may be that changes to the data model will make that easier and more efficient. Without knowing more about the exact use case and questions you need the graph to answer that is about the best advice I can offer.