RediSearch: Features, Benchmarks, Installation, and Usage Guide

RediSearch is a Redis module that provides full‑text search, secondary indexing and query capabilities on top of Redis data.

Key features include multi‑field document indexing, high‑performance incremental indexing, sortable fields, complex Boolean queries with AND/NOT, optional query clauses, prefix search, field weighting, autocomplete suggestions, exact phrase search, stemming‑based query expansion, custom scoring functions, numeric filtering, geo filtering, Unicode support, and document update/delete with garbage collection.

Benchmark tests show RediSearch builds indexes 58% faster than Elasticsearch (221 s vs 349 s) and achieves higher query throughput (12.5 K ops/s vs 3.1 K ops/s) with slightly lower latency (8 ms vs 10 ms) in a simulated multi‑tenant e‑commerce scenario.

Installation can be performed from source:

git clone https://github.com/RediSearch/RediSearch.git
cd RediSearch
make setup
make install

or via Docker: docker run -p 6379:6379 redislabs/redisearch:latest After loading the module, verify with: 127.0.0.1:0> module list Typical CLI workflow:

Create an index:

ft.create "student" schema "name" text weight 5.0 "sex" text "desc" text "class" tag

Add a document:

ft.add student 001 1.0 language "chinese" fields name "张三" sex "男" desc "这是一个学生" class "一班"

Search examples – full‑text, fuzzy, prefix, field‑specific queries, and highlighting.

Delete documents or drop the index with ft.del student 002 and ft.drop student.

Manage index aliases using FT.ALIASADD xs student, FT.ALIASDEL xs, etc.

The article also notes that RediSearch uses a compressed inverted index, supports Chinese tokenization via friso, and integrates with Redis’ replication and pub/sub mechanisms.