A.6 Using Off loaded Subgraph
This topic describes instruction to load graphs into the in-memory graph server.
There are two options for loading graphs into the in-memory graph server (PGX).
- In-memory: It collects all the data from the nodes and edge tables and it will be transformed into a single global graph in the Oracle Graph Analytics Engine (PGX).
- Offloaded: It collects all the data from the nodes and edges and keeps it ready to be loaded in Oracle Graph Analytics Engine (PGX) so that users can load subgraphs while doing graph analytics.
The following are the advantages of using subgraph:
- This feature provides a much more flexible way to declare the scope of the data to be loaded into PGX server.
- For global graphs with a high volume of records, the PGX server may need to be very large to hold the full global graph. Subgraph loading allows for many smaller graphs to be loaded which requires a smaller PGX server size.
Creating and defining Graph Pipelines
- For information on how to create a graph, see the Graphs section.
- For information on how to define the graph, see the Updating Graph Definition section.