Amazon Neptune - Schemas and Data Models

Neptune is a "schema-on-read" database and does not provide a way to define a schema or constraints. That being said, there are one or two built-in constraints that you can take advantage of:

1. If using Neptune to store Property Graphs, every node and edge must have a unique ID. You can use this to your advantage if there are concepts in your data that must be unique.
2. Every edge must have a node at either end of the edge. You cannot have "dangling edges". This ensures that when you delete a node, all adjacent edges to the node are also deleted.