Single‑Level Partitioning Case Studies
This topic presents several case study scenarios to evaluate various approaches to
row partitioning a primary index on one level in such a way that optimal benefit is
The first scenario (see “Scenario 1” on page 427) is set in a retail sales environment. The effects of row partitioning a sales table
on months is examined from several different perspectives.
The second scenario (see “Scenario 2” on page 431) is a continuation of the first and attempts to boost the benefits of using a partitioned
table further by partitioning on days rather than months. The concept is to investigate
how the number of row partitions affects query workload performance.
The third scenario (see “Scenario 3” on page 432) is set in a telecommunications environment. The effects of partitioning a telephone
call tracking table on call date and phone number are compared and contrasted for
the same query workloads.
The final scenario (see “Scenario 4” on page 434) illustrates how the decision whether to use row partitioning is often multidetermined,
requiring a solution based on numerous, carefully weighted factors.