Understand Full Text Search Save

Understand full-text search with Postgres

Piotr Lewandowski, @constjs

Table of content

0. Create demo DB tables
1. Stemmer — Building documents
2. Search — Building queries
3. Performance practises
4. Setting weight and ranking
5. Improve search quality
6. Further reading

When full-text search?

• Search by content created by people (not programmers)
• Divide more and less important fragments of document
• Searching database dumps from WikiLeaks

Why just not RegEx?

• RegEx is good to find only simple, finite languages

• Helpless for grammar

• Slow _{(Can be improved with Trigram Indexes)}

• Lots of pitfalls even for simple languages like HTML

• Complicated to maintain, e.g.

  ^(?=[A-Z0-9][A-Z0-9@._%+-]{5,253}$)[A-Z0-9._%+-]{1,64}@(?:(?=[A-Z0-9-]{1,63}\.)[A-Z0-9]+(?:-[A-Z0-9]+)*\.){1,8}[A-Z]{2,63}$
  

Why Postgres?

• Pretty rich in features
• Flexible and extensible
• Maybe you already have it
  • Low entry point
  • If your technology stack is already over-engineered