Full-text search - seek and ye shall find
summary
Demonstrates PostgreSQL full‑text search working on a million Open Library rows and 40 years of student newspapers ingested from archive.org, using TSVector columns maintained by triggers, GIN indexes and optional GiST, weighting with setweight and relevance via ts_rank_cd, returning top results under ~100 ms in tests, shows stopword removal by custom text search configuration, highlights with ts_headline, and explores trigrams for spell‑correction and prefix matching; examples include build times and where phrase matching forces a LIKE pass.
AI-written summary · about
chapters (10)
- 0:00Start
- 6:22Create the Weighted Keyword Column
- 8:27A Weighted Search
- 11:40Relevancy Ranking Thresholds
- 15:32Full Text Parser
- 17:17Text Search Data Types
- 19:14Prefix Matching
- 49:04Example of Synonyms via Dictionary
- 50:16Postgres Trigram Extension
- 53:18Auto Completion
related talks
43mFull-Text-Search Explained from A to Z with French Food
34mFull-Text Search in PostgreSQL
24mSearch From the Ground Up
27mMulti-table Full Text Search with Postgres
53mFull-Text Search in PostgreSQL in Milliseconds
49mBreakout Session
22mFrom Postgres full text search to Retrieval Augmented Generative search
28mElasticsearch-Quality Full-Text Search in Postgres via Tantivy