graph database
/graf DAY-tuh-bays/ · noun · Development · Origin: 2000
Definitions
A database that stores data as nodes and the edges connecting them, treating relationships as first-class entities with their own properties rather than as joins computed at query time. This makes traversal queries, such as finding friends-of-friends or tracing a path through a supply chain, run in time proportional to the subgraph touched rather than to total table size, which is where relational databases degrade as join depth grows. Neo4j with its Cypher query language is the best-known implementation; alternatives include Amazon Neptune, TigerGraph, and the vendor-neutral Gremlin traversal language. Graph databases suit fraud detection, recommendation engines, knowledge graphs, and network topology, and suit poorly the aggregate-heavy analytical workloads where columnar stores win decisively.
In plain English: A database built around relationships between things — perfect for social networks or fraud detection where you need to trace connections between people or entities.
Example: Finding all fraud rings within three degrees of a flagged account takes milliseconds in the graph database but would be a nightmare of recursive CTEs in PostgreSQL.