Thenne entryd in to the bataylle Iubance a geaunt and fought and slewe doune ryght and distressyd many of our knyghtes.
—Sir Thomas Malory (d.1471) Le Morte D’Arthur, V, 11
This chapter deals with the particular challenges that are facing us when data volumes swell. Those challenges include searching gigantic tables effectively, but also avoiding the sometimes distressing performance impact of even a moderate volume increase. We’ll first look at the impact of data growth and a very large number of rows on SQL queries in the general case. Then we’ll examine what happens in the particular environments of data warehousing and decision-support systems.
Some applications see the volume of data they handle increase in considerable proportion over time. In particular, any application that requires keeping online, for regulatory or business analysis purposes, several months or even years of mostly inactive data, often passes through phases of crisis when (mostly) batch programs tend to overshoot the time allocated to them and interfere with regular, human activity.
When you start a new project, the volume of data usually changes, as shown in Figure 10-1. Initially, hardly anything other than a relatively small amount of reference data is loaded into the database. As a new system replaces an older one, data inherited from the legacy system is painfully loaded into the new one. First, because of the radical ...