One document is made up of one or more fields. A field is the smallest definable unit of a data index in Lucene. There are 3 buckets of that field types split into:
Inverted Index — term-based search/filtering
Inverted Index
Link to original
Field Type Analyzed? Use Case [[7k4si5yyy4nqk3tzueci7sl tech/lucene/concept/entity/field/indexed/TextField]] Yes (tokenized) [[6cn5n3jd6wlwi35fear4iw0 tech/lucene/concept/entity/field/indexed/StringField]] No (exact)
BKD-trees — range filtering on numbers/geo
BKD-Tree (points) — efficient range filtering
Link to original
Field Type Use Case [[2e1p3ti5ti81f6bh5s1gx7j tech/lucene/concept/entity/field/BKD-trees/LongPoint]] [[qhx1lzbmzrhujml9gxlb0pp tech/lucene/concept/entity/field/BKD-trees/IntPoint]] [[i5kdcs6k84sw2ejw46w5cua tech/lucene/concept/entity/field/BKD-trees/LatLonPoint]]
HNSW-Graph-Vectors — similarity/semantic search
HNSW Graph (Vectors) — similarity/semantic search
- KnnFloatVectorField
- KnnByteVectorField
(added in 9.x lucene)
Link to original
DocValue: Sorting, Grouping, Aggregations (column-oriented)
DocValues (column-oriented)
Link to original
Stored: for retrieval/display (row-oriented)
Stored (for retrieval)
Link to original
- Row-oriented blocks — StoredField