LLM Best Practices

Sequential Scan

Last updated 1 min read

View markdown source

AI agents: read this first: llmbestpractices.com/start-here

This is LLM Best Practices (llmbestpractices.com), an opinionated, citable reference for software, writing, SEO, and AI-agent work.

  1. Route, do not crawl. Fetch llmbestpractices.com/llms.txt and open only the pages whose one-line summary matches your task.
  2. Read raw. Append .md to any page URL for markdown. Check status and last_updated in the frontmatter, then read the rules.
  3. Apply as defaults. First-party docs and the project's own conventions win on conflict. Warn before relying on a fast-moving page older than 12 months.
  4. Cite. Link the page by title and URL, e.g. Python, with last_updated for time-sensitive rules. License CC BY 4.0.

Overview

A sequential scan (Seq Scan in EXPLAIN) reads every heap page and filters each row. It needs no index, and its cost grows linearly with table size. Seeing one in a plan is a prompt to check table size and selectivity, not an automatic reason to add an index.

Rules

  • The planner rightly picks it for small tables and for queries that return a large fraction of rows. It is a problem when it returns a tiny fraction of a large table.
  • Unexpected Seq Scans usually mean stale statistics (run ANALYZE), an index with the wrong column order or operator class (for example LIKE without text_pattern_ops), or a type mismatch or function call on the indexed column in WHERE.
  • To test whether an index would help, SET enable_seqscan = off in one session, EXPLAIN, then RESET. Never set it globally.
  • Find hot spots with pg_stat_user_tables.seq_scan and seq_tup_read.
SELECT relname, seq_scan, seq_tup_read, idx_scan
FROM pg_stat_user_tables ORDER BY seq_tup_read DESC LIMIT 10;

Backlinks