I understand that the order in which the tables are placed in the FROM clause of a SELECT statement can have a significant effect on performance. However, what I don't understand is how you decide on which is the most efficient order in which to place the tables. I have a query that will join 5 tables together, one of which has ~1.5 million records. I looked at the following article: http://www.oracle.com/oramag/webcolu...rf_082100.html, but it wasn't entirely clear how he got his "NSR" figure. If anyone could point me to any other online resources, or describe the procedure step by step I'd be very grateful.

Thanks,

Matt