NEW YORK — What happened: The PostgreSQL global development group has released PostgreSQL 18, featuring major enhancements to structured JSON data types, including native vectorized indexing for deeply nested document fields.
Why it matters: Many modern applications store semi-structured data alongside rigid relational records. Previously, querying JSON fields efficiently required cumbersome extensions or external document databases, adding architectural complexity.
Deep dive: PostgreSQL 18 implements a new storage layout for JSON values that compresses repetitive keys and allows the query planner to construct cost-based execution paths across both relational columns and JSON keys simultaneously. This turns JSON query performance closer to native tabular lookups.
Report check: Claims vs what is verified vs still rumor. It is verified that JSON query execution speeds have improved significantly in benchmark tests. Rumors that PostgreSQL 18 can fully replace dedicated document databases for all extreme write-heavy workloads are exaggerated.
Open questions: Will cloud database providers pass these performance gains through without raising managed instance pricing tiers?