DB Issues - compression

With the increasing size of database files and the need to store a variety of file types (eg. images, sound, video) as well as text data, compression techniques are becoming more and more an essential issue in a database.

Compression eliminates unnecessary information, such as empty fields and redundant data.

Compression can occur;
- On a disk drive to overcome hardware limitations (eg. Drive space/Double space), and compact datafiles.
- On a database to delete unwanted space (such as occurring with deleted records)
  - Physically remove records that have been logically deleted. In some databases a record is marked as deleted but not necessarily actually removed from the database
  - MS-Access files grow quickly when forms/reports are being built and should be compressed after programming has occurred.
- When backing up a datafile onto tape files are compressed. A decompression program is need to restore the files.
  - In many databases where records or fields are of fixed length, enormous amounts of space may be unused. (eg. in a name field of 25 charactres "Tui Smith" only fills the first 9 characters leaving 14 unused). Tape backup will copy the database eliminating the spaces. Note that a restore program is needed to retrieve the information. Common compression programs used on the PC is PK-ZIP (.zip), on the Macitosh Stuffit (.sit) and on Unix GZip.
- To transfer files from one computer to another, or compress files to save space programs like PK-ZIP or ARJ are used.
- Saving an object(such as a multimedia file) in a different format can compress the data significantly ( eg. A bitmap file saved as a gif file can be compressed up to 10 times, a video file is compressed using MPEG-2, a sound file saved as an MP3 file can be compressed 12:1.)

(McFadden, F. et al 1999, pp257-259)