Compression for Fast Read-only Access of a Large Database
Paul D. Filipski, Jonathan J. Hull, Sargur N. Srihari · 1992
An approach for the compression and high speed a1:cess of a large read-only database is described. This method uses a set of pre-compiled search trees to access compressed source data. A technique known as attribute coding uses local characteristics of the .sourcedata to determine an optimal compression strategy. This paper describes the overall appr9ach, including the design of the search trees and the attribute coding. An application of this method using the ZIP+4 file of the United States Postal Service is also discussed. This application achieves an effective compression ratio Qf25:L Also, efficient access to the data for a real-time address recognition system. is aChievable.