Is it possible to use Enterprise Architect to model Graph databases the way they are usually drawn (round nodes with labeled arches), or what could be the best approximation? I’m primary interested in Apache TinkerPop / Gremlin.
Yes. We use Class Diagrams to accomplish this. As you know, graph databases deal with Nodes, Edges, and Properties. Classes represent nodes, Associations represent relationships, and Attributes represent the Properties of Nodes and Edges. We model our Class diagram(s) to represent a specific instance of the Graph, which then allows us to depict the Nodes (as objects) and the associations (edges) between Nodes in the Graph. There are limitations, so we usually do this side-by-side with query languages (e.g., Cypher, Neo4j) for our graph-specific models. This works as long as we understand the limitations. Thank you for the question! We have done quite a bit with UML related to things like horizontal and vertical data relevance (and traversal) in ML design and learning models (e.g., Personalization and Tailoring). These are advance subjects in CASE/UML, and will try to get to these later.
@@UMLOperator As regards how to make the shape of a Node represented by a Class to be round, I got the answer in your video [Sparx EA Custom Drawing Style (e18)]
Is it possible to use Enterprise Architect to model Graph databases the way they are usually drawn (round nodes with labeled arches), or what could be the best approximation? I’m primary interested in Apache TinkerPop / Gremlin.
Yes. We use Class Diagrams to accomplish this. As you know, graph databases deal with Nodes, Edges, and Properties. Classes represent nodes, Associations represent relationships, and Attributes represent the Properties of Nodes and Edges. We model our Class diagram(s) to represent a specific instance of the Graph, which then allows us to depict the Nodes (as objects) and the associations (edges) between Nodes in the Graph. There are limitations, so we usually do this side-by-side with query languages (e.g., Cypher, Neo4j) for our graph-specific models. This works as long as we understand the limitations.
Thank you for the question! We have done quite a bit with UML related to things like horizontal and vertical data relevance (and traversal) in ML design and learning models (e.g., Personalization and Tailoring).
These are advance subjects in CASE/UML, and will try to get to these later.
@@UMLOperator As regards how to make the shape of a Node represented by a Class to be round, I got the answer in your video [Sparx EA Custom Drawing Style (e18)]