In the previous chapter, I warned you that the
clustering_factor was an important factor in the cost of using a B-tree index for a range scan and could easily be the biggest cause of errors in the calculation of cost. It's time to find out why things can go wrong.
This chapter is very important because it describes many of the sensible strategies that DBAs adopt to improve performance or avoid contention, only to discover side effects that can leave the optimizer ignoring indexes that it ought to be using. Almost invariably, the sensible strategy has caused problems for some queries because of the impact it has had on the
clustering_factor is a single number that represents the degree to ...