Graph-Like Data Models

As connections within your data become more complex, modeling them as a graph becomes more natural.

Property Graphs (e.g., Neo4j)

In a property graph:

1. Vertices: Have a unique ID, set of incoming/outgoing edges, and key-value properties.
2. Edges: Have a unique ID, head/tail vertices, a label, and properties.

Triple-Stores and SPARQL

Information is stored as statements: (subject, predicate, object).

• (Jim, likes, bananas) SPARQL is the standard query language for triple-stores.

Datalog: The Foundation

Datalog is a much older language (1980s) that provides the foundation for later graph query languages. It defines rules that tell the database about new predicates derived from facts.

• Fact: name(namerica, 'North America').
• Rule: within_recursive(X, Y) :- within(X, Y). Datalog is powerful for complex queries because rules can be combined and reused.