Hash Join
The engine builds an in-memory hash table on the smaller input's join key, then streams the larger input probing it, giving roughly linear cost for equality joins on large unsorted inputs. If the build side exceeds the per-operation memory budget (work_mem in PostgreSQL), it spills to disk in batches and slows sharply.
This Concept is waiting for its first lesson!
The engine builds an in-memory hash table on the smaller input's join key, then streams the larger input probing it, giving roughly linear cost for equality joins on large unsorted inputs. If the build side exceeds the per-operation memory budget (work_mem in PostgreSQL), it spills to disk in batches and slows sharply.
Are you a teacher? Sign in to start contributing.
Sign In