The 80-20 Retriever search engine is no longer sold separately: it's now part of the 80-20 Document and Records Management Solution, which is designed for enterprise-wide management and search of documents within an organization.
Dieselpoint is a pure Java search engine that indexes metadata and field attributes as well as text and documents, allowing users to navigate within search results - faceted metadata search and browse. It indexes HTML, XML, pdf (Adobe Acrobat), Microsoft Office, and text files, as well as JDBC connectors for DB2, Oracle, Informix, MS SQL, MySQL, PostgreSQL, and Cloudscape, and near-real-time incremental index updating. It can handle up to forty languages and provides stemming for European languages. Additional features include spellchecking, synonyms, special weighting of search results, and extensive search administrator reports and controls. Definitely worth looking at for faceted metadata, at a lower price than most others.