full-text-search-specialist

To implement full-text search in PostgreSQL, create a generated column or trigger to maintain a tsvector from your target text fields, add a GIN index to that column, and query it using websearch_to_tsquery for fast ranked results.

The best way to add fuzzy search and autocomplete in Supabase is combining GIN-indexed tsvector fields for standard search with the pg_trgm extension for trigram similarity matching against user input.

Rank search results by applying field weights to title, content, and tags when building your tsvector, then use the ts_rank or ts_rank_cd functions in your ORDER BY clause to sort by relevance.

Use websearch_to_tsquery for parsing user search input when you need to support standard search engine syntax like quoted phrases, and use plainto_tsquery when you want to treat all user input as plain literal text without syntax errors.

PostgreSQL full-text search scales effectively for large datasets when you apply GIN indexes to your tsvector columns, often eliminating the need for external search engines by using EXPLAIN ANALYZE to tune query performance.

Standard full-text search fails for misspelled words because it relies on exact lexeme matching; enable the pg_trgm extension and apply trigram similarity checks to calculate fuzzy matches and resolve typographical errors.