Local vs Global Schema

Hello I am confused by the concepts of local and global schemas. In particular I understand a Tigergraph instance can have many graphs each with their own local schema. I also understand the benefit of a schema is to strongly “type” the graph vertexes/edges/attributes.

Now’, I also understand that a local schema can inherit/add to/drop/change patterns from a global schema. I.e., they could be completely different.

In this context why do we have a global schema? Is it necessary?

A big benefit of the global schema is that any vertices that exist across a given global vertex type, will persist across all graphs that include that global vertex.

You can use the global schema to define the big picture, for example a supply chain all the way from raw material supplier, through manufacturing and storage, all the way to the customer’s home.
All of this is important, but your sales team really only cares about customer orders and demographics. So you could create a local graph that includes just those customer centric vertices from the global schema and give your sales team access to that local graph.

2 Likes

Got it thanks!.. plus there is a short discussion on the practicality here: Loading Data into TigerGraph - YouTube

1 Like