An efficient file structure for specialized dictionaries and other “lumpy” data

摘要

There are many cases where it is necessary to store sets of data that are variable in length, and to search these in order to satisfy requests for subsets with a common characteristic. This article presents a file structure that holds an integrated English dictionary used to locate clusters of words for presentation to an intelligent spelling error correction system. Although the emphasis has been on misspelling, the structure presented is capable of handling any other types of lumpy data provided the characteristics used in search requests can be translated into a set of integer numbers.