RCubes
From: abombss <abombss_at_hotmail.com>
Date: Tue, 17 Dec 2002 02:34:55 GMT
Message-ID: <j3wL9.383238$QZ.58704_at_sccrnsc02>
I have been doing some investigating about data warehousing and I wanted to ask this groups opinion about the idea of RCubes. It looks like Sybase is the only vendor using this approach, from what I understand it is a way to denormalize your data but not experience the typical data explosion with traditional RDBMS using a start schema. Sybase IQ stores data by column rather than by row. All columns are compressed and indexed using different bit wise techniques so lookups based on range and aggregates are still very quick.
Date: Tue, 17 Dec 2002 02:34:55 GMT
Message-ID: <j3wL9.383238$QZ.58704_at_sccrnsc02>
I have been doing some investigating about data warehousing and I wanted to ask this groups opinion about the idea of RCubes. It looks like Sybase is the only vendor using this approach, from what I understand it is a way to denormalize your data but not experience the typical data explosion with traditional RDBMS using a start schema. Sybase IQ stores data by column rather than by row. All columns are compressed and indexed using different bit wise techniques so lookups based on range and aggregates are still very quick.
Has anyone had experience with Sybase ASE IQ Multiplex, or does this idea sound promising.
The funny thing is, a couple weeks ago, prior to learning about IQ, I kicked around the exact same idea of storing data by columns and using bits to save space. However, I couldn't get past the problem of high cardinality data, but I guess Sybase did.
Cheers,
Adam
Received on Tue Dec 17 2002 - 03:34:55 CET