On The Look-Out For Your Data
summary
Compares database approaches (PostgreSQL ILIKE/contains and trigram indexing) with external engines (Elasticsearch, Solr, Whoosh, Xapian) and shows working code for each, including a trigram GIST index that makes ILIKE searchable, Django class-based indexes, and search-backed apps that use transaction.on_commit to sync writes. Covers practical gotchas: .update and QuerySet.delete skip save so you must handle PKs before deletion, search results can be stale (eventual consistency), and the reliable option is a full search index storing URL and metadata so reads can come straight from the search engine.
AI-written summary · about
chapters (10)
- 0:00Introduction to data search
- 4:33Searching in Django
- 6:21Text search challenges
- 10:59Efficient text search with trigrams
- 13:52Full-text search features
- 14:34External search engines
- 16:42Synchronizing data with search
- 18:35Maintaining full search index
- 20:44Why search matters
- 23:16Q&A session
related talks
24mSearch From the Ground Up
42mYou don't need Elasticsearch! Fuzzy Search with PostgreSQL and Spring Data
26mA pythonic full-text search
33mHow to Make Your Backend Roar
34mFull-Text Search in PostgreSQL
43mFull-Text-Search Explained from A to Z with French Food
29mPostgres Index types and where to find them
45mSQL Indexing Beyond Simple Columns: Expressions, Text Fragments, Attributes J...