Bringing GraphQL Natively To Relational Databases With AI
summary
Oracle AI Database 26ai runs GraphQL natively via a new GraphQL table function that accepts standard GraphQL strings inside SQL, automatically infers joins from FKs or uses a link directive, supports generated fields, GraphQL variables mapped to SQL bind variables, and a GraphQL-to-SQL PL/SQL function that emits a single SQL JSON query for diagnostics. A public IBM benchmark shows NL→GraphQL has a lower error profile than NL→SQL, and an open Agent Spec plus a private AI services container lets agents run inside the DB and expose AI‑enriched views over the same GraphQL surface.
AI-written summary · about
chapters (10)
- 0:00Introduction and agenda
- 1:04Why developers prefer GraphQL
- 2:51Relational schema as a graph
- 3:55Declarative GraphQL over relational data
- 6:51GraphQL table function examples
- 9:16GraphQL to SQL translation
- 11:00GraphQL vs SQL for MCP tools
- 14:30Agentic AI and GraphQL integration
- 18:41Key takeaways and summary
- 19:23Q&A session
related talks
47mHow to query data using natural language - intro to AI features in Oracle 23ai
27mAI-Native Databases: Four Innovations That Matter
48mOracle Database Directions: Innovate, Scale, and Sleep Easier
31mWhy AI needs GraphQL
23mBuild Natural Language-Powered Apps with Select AI
23mRevolutionizing Data Access Using GraphQL in the Oracle Database
06mOracle AI Database
12mTackle Any Workload with Oracle Database, AI, and Multicloud