FULLTEXT

Full text, in information retrieval, refers to the complete textual content of a document. It contrasts with metadata, abstracts, titles, or structured fields. A full-text search system builds an index over the content to support queries that search the words contained in the documents themselves.

In typical implementations, the text is tokenized, lowercased, and normalized; common words (stop words) may be removed; terms may be stemmed or lemmatized. The core data structure is an inverted index that maps terms to the documents containing them. When a user submits a query, the system retrieves candidate documents and ranks them using a scoring function such as TF-IDF or BM25, possibly incorporating term frequency, document frequency, and field weights.

Full-text search is provided by search engines, document management systems, and database systems with native full-text indexing. Examples include search features in web search engines; PostgreSQL's tsquery/tsvector, MySQL/MariaDB FULLTEXT indexes, and SQLite FTS modules. Some systems support phrase queries, proximity searches, boolean operators, and ranking controls, as well as support for multiple languages and stemming.

Benefits include improved recall for keyword-based queries and the ability to search entire documents. Limitations include potential noise from common terms, the need for linguistic processing, update and storage costs, and imperfections in ranking. Quality depends on indexing choices, language support, and the handling of synonyms, stop words, and document length.