Data in: documents and indices

• Elasticsearch Overview:

  • Elasticsearch serves as a distributed document store.
  • Stores complex data structures serialized as JSON documents.
  • Documents are indexed and fully searchable in near real-time (within 1 second).
  • Utilizes an inverted index for fast full-text searches.
  • Each unique word in any document is listed in the inverted index, linking to the relevant documents.

• Index and Document Structure:

  • An index is an optimized collection of documents.
  • Each document contains fields (key-value pairs) with data.
  • Elasticsearch indexes all data in every field by default.
  • Indexed fields have dedicated, optimized data structures (e.g., inverted indices, BKD trees).

• Schema Flexibility:

  • Elasticsearch is schema-less; documents can be indexed without explicit field definitions.
  • Dynamic mapping automatically detects and adds new fields to the index.
  • Easy exploration: Start indexing documents, and Elasticsearch maps data types (booleans, floats, dates, etc.) automatically.

• Custom Mappings:

  • Define your own mappings to take control:
  • Distinguish between full-text and exact value string fields.
  • Perform language-specific text analysis.
  • Optimize fields for partial matching.
  • Use custom date formats and non-detected data types (e.g., geo_point, geo_shape).

• Flexible Indexing:

  • Index the same field differently for various purposes (e.g., full-text search vs. sorting).
  • Apply multiple language analyzers to process user input in string fields.

• Analysis Chain:

  • Full-text field analysis during indexing is also used at search time.
  • Query text undergoes the same analysis before term lookup in the index.