Elasticsearch

Mapping Mistakes

• Always define explicit mappings—dynamic mapping guesses wrong (first "123" makes field integer, later "abc" fails)
• text for full-text search, keyword for exact match/aggregations—using text for IDs breaks filters
• Can't change field type after indexing—must reindex to new index with correct mapping
• Set dynamic: "strict" to reject unmapped fields—catches typos in field names

Text vs Keyword

• text is analyzed (tokenized, lowercased)—"Quick Brown" matches search for "quick"
• keyword is exact bytes—"Quick Brown" only matches exactly "Quick Brown"
• Need both? Use multi-field: "title": { "type": "text", "fields": { "raw": { "type": "keyword" }}}
• Sort/aggregate on title.raw, search on title

Query vs Filter Context