An efficient file structure for document retrieval in the automated office environment

A file system tailored to the general needs of the office environment is proposed. This system supports large numbers of a wide variety of documents and inexact ‘fuzzy’ queries on the documents. The file system is based on a multilevel file structure; the structure combines and extends multikey extendible hashing and signature files to create a document retrieval system that is more time efficient than other previously proposed systems, and also space efficient.