Full-text analyzers

Analyzers create fulltext indexes by converting field content into a stream of tokens. They consist of:

1. Char filters – applied first, modify or filter characters.

2. Tokenizer – splits character streams into tokens.

3. Token filters – transform, remove, or add tokens.

Analyzers can be built-in or custom, and they allow language-specific processing for better search accuracy.

Analyzer	Type	Description	Parameters
standard	standard	Uses Standard tokenizer + standard, lowercase, and stop token filters. Unicode text segmentation (UAX#29).	stopwords, max_token_length
default	default	Same as standard.	–
simple	simple	Uses Lowercase tokenizer.	–
plain	plain	Alias for keyword analyzer. Cannot be extended.	–
whitespace	whitespace	Uses Whitespace tokenizer.	–
stop	stop	Lowercase tokenizer + stop filter.	stopwords, stopwords_path
keyword	keyword	Produces a single token from the entire content.	–
pattern	pattern	Splits text based on regex.	pattern, lowercase, flags
language-specific	<language>	Supports many languages (English, French, German, etc.) with stopwords and optional stem exclusion.	stopwords, stopwords_path, stem_exclusion
snowball	snowball	Standard tokenizer + standard, lowercase, stop, snowball filter.	stopwords, language
fingerprint	fingerprint	Lowercases, normalizes, sorts, deduplicates, concatenates tokens. Useful for clustering.	separator, max_output_size, stopwords, stopwords_path

---

T