Full-text search engines help users find the right text quickly.

Real users do not search like database queries. They type half-remembered phrases, make spelling mistakes, use different words for the same idea, add filters, and still expect useful results at the top.

A product search like running shoes size 10 is more than "find this exact text." The engine has to break the query into words, find possible matches, rank them, apply filters, handle typos, and often mix in business rules such as popularity or availability.

Relational databases can provide useful full-text search for many applications.

Dedicated search engines become a better fit when search is a core product feature, the dataset is large, ranking quality matters, or users expect features like autocomplete, highlighted matches, filter counts, and fast log exploration.

Search engines are usually secondary indexes. The main copy of the data still lives in a database or object store, while the search engine keeps a searchable copy shaped for fast lookup and ranking.

This chapter covers how search engines build inverted indexes and rank results.