>> I doubt that it makes a difference in SQL Server either (although I
don't that for sure). <<
Unfortunately, it does. SQL Server uses a simple B-tree, so you get a
different depth of tree with (state_code, city_name) versus
(city_name, state_code). They concatenate the columns in the order
given, so if the state_code appears first, you get a "tall, deep tree"
for the index. If you give the city_names first, you get a "flat,
wide tree" for the index. The heuristic is to start with the column
with the most values in its range.
|
