One of the major blunders in SQL Server 2000 is in the calculation of data distribution statistics. In theory, it should be possible to take just a small true random sample to get a reasonable estimate of the true distribution. Because of the page organization
Read More...