A generalized file structure consists of a file and its directory from which some of the existing file structures such as inverted, index-sequential, and multilist files can be derived and specialized. The directory acts as a reference for information within the file. It consists of keywords (i.e., attribute-value pairs) and addresses of records which are characterized by these keywords. The directory and its access routines simulate an associative memory in that records can be located in the file by merely providing the keywords which are associated with the records. The directory is initialized when the file is created and is updated for each addition of a keyword or a record to the file. For a given attribute, directory decoding can be done for one specific value or a range of values. Decoding involves the search of the keywords within the directory and the addresses of the records characterized by these keywords.