Archives of the TeradataForum
Message Posted: Wed, 30 May 2001 @ 14:51:50 GMT
Compression in Teradata works quite differently than compression in other RDBMS's such as DB2. DB2 uses an algorithm (Zev-Lempl, I believe) to substitute short binary values for all occurrences of the same value in a tablespace. Teradata does not use an algorithm at all. It uses compression bits in the row header that are flipped on if the value is to be compressed. The actual value is not stored in the row at all.
In DB2 you take a small performance hit with uncompressing the compressed data. In Teradata, you never really uncompress the data, thus no performance hit. We have only seen huge benefits in using compression in our large tables.
|Copyright 2016 - All Rights Reserved|
|Last Modified: 28 Jun 2020|