Browse Teach Curate

Predicate Selectivity

Selectivity is the fraction of rows a predicate passes, estimated from statistics (e.g. equality on a column with k distinct values ≈ 1/k). It drives both the scan choice — low selectivity favors an index — and the estimated size of every intermediate result.

Questions this Concept answers

Why does a highly selective predicate make an index worth using?

jeremy Text

69.1. Row Estimation Examples

This section of the PostgreSQL manual works through real examples of how the planner guesses how many rows a condition will match. It shows the arithmetic done against the stored statistics: looking …