Alexander Ioffe: I ran the same query twice and Postgres disagreed with itself

Alexander Ioffe reports that two EXPLAIN ANALYZE runs on statistically identical data produce different PostgreSQL plans, reading 73 versus 3,449 heap blocks despite a 1.5% margin. The analysis uses ExoBench to examine the cause and describes an index shape intended to remove that planner variability.

Coverage 1 publisher

  1. Planet PostgreSQL

    Alexander Ioffe: I ran the same query twice and Postgres disagreed with itself

Articles stay on their publishers’ sites; each link opens the original.