A big problem (IMHO) with graph databases is building "the" graph model, and the fact that it's easy to be faced with problems that don't suit a graph database. Something as simple as returning the distinct set of values for an attribute and count of vertices containing each value require going outside the graph model, so aren't composable very well in a property graph system. (There are other graphs besides property graphs, they will have their time someday.)

What you really want is to apply graph processing to data as it is. The SQL 2023 additions are a step in the right direction. I need to find a good detailed description of the constraints and semantics to assess how good it is.