It was a very simplified version of the actual situation. Which might mean I created a XY Problem unintentionally.
So for the report I had, I was pulling a patients orders, and nurse documentation. The only thing that ties all of this together is the patient and dates. So for instance there was an order placed and then you want to get all of the documentation between that order and the next order. In a graph database, you would create an edge from the documentation to the order, but that is not how it is done in the SQL DBMS DR. So if you were to extract the data and load it into Dgraph and then try to run mutations to transform the data, you would have to solve the problem above.
Does that make sense at all?
The xid’s only purpose here was to point back to the data’s original source. So if I went and looked for the PK for the orders, and documentation in the original DBMS I could verify the accuracy of the data.