
7.8 Range Partitioning and Clustering Indexes 139
7.8 Range Partitioning and Clustering Indexes
Clustering indexes dramatically improve the efficiency of scans by reducing the I/O
required on the base table by ensuring that records with similar key data are located near
each other within the table storage. As described earlier, range partitioning implicitly
clusters data according to the range partition. However, the design of range partitioning
is often driven predominantly by administration needs (utility isolation, roll-in, and
roll-out), while clustering design is often driven by performance goals. The goals can
therefore conflict. Optimal scans are ...