I need some advice more than a specific solution.
As we roll out, we are getting a lot of requests for alternative navigation and visualizations. Those have been "encapsulated" by relationships. One of our DBs has a:Stops:Drives:Rest all as relationships that users select. On our next project, we see this same pattern emerging.
Some of our DBs have millions of nodes. Building them on the fly has been slow even on 16 cores. So we try to predefine those. What are the upsides/downsides for adding the specialized relationships ?
For the avoidance of doubt, our process is
- Load the data with a default relationship
- Clean and normalize
- Build secondary relationships (e.g. Drives:Rest)
- release to users
yFiles/Outsystems uses those pre-defined secondary relationships to build the GUI which the users love. So the speed has not been an issue.