5.4.2 Parallel Loading of Graph Data

A Java API is provided for performing parallel loading of graph data.

Oracle Spatial and Graph supports loading graph data into Oracle Database. Graph data can be loaded into the property graph using the following approaches:

• Vertices and/or edges can be added incrementally using the graph.addVertex(Object id)/graph.addEdge(Object id) APIs.

• Graph data can be loaded from a file in Oracle flat-File format in parallel using the OraclePropertyGraphDataLoader API.

• A property graph in GraphML, GML, or GraphSON can be loaded using GMLReader, GraphMLReader, and GraphSONReader, respectively.

This topic focuses on the parallel loading of a property graph in Oracle-defined flat file format.

Parallel data loading provides an optimized solution to data loading where the vertices (or edges) input streams are split into multiple chunks and loaded into Oracle Database in parallel. This operation involves two main overlapping phases:

• Splitting. The vertices and edges input streams are split into multiple chunks and saved into a temporary input stream. The number of chunks is determined by the degree of parallelism specified

• Graph loading. For each chunk, a loader thread is created to process information about the vertices (or edges) information and to load the data into the property graph tables.

OraclePropertyGraphDataLoader supports parallel data loading using several different options:

• JDBC-Based Data Loading
  
• External Table-Based Data Loading
  
• SQL*Loader-Based Data Loading