My prodn. databases have got tables ranging from a count of 30 million rows to 70 million rows, we have a analyze compute statistics command which is run as a cron job, which takes about 5hrs. to complete, i am looking to reduce this time by doing a analyze estimate statistics with a particular percentage, now here is my problem, what would be the best
percentage of rows to be used for estimating statistics on these very large tables?
reply soon
soren