SELECT count(*) FROM table WHERE col1>'abc' AND col1<'xyz';In the original query, the result was indeed a count(*) so no access to the data we required there. But access to the data was required in order to evaluate the WHERE clause. So it is still O(NlogN).
What is different between his where clause and the one you cite as an example that only takes O(N)? Is it just that in your example col1 is (part of) the index? So, wouldn't Ken be able to do the same, except that he needs to step through two indices? That, it seems to me, is the crux of the issue.
I don't mean to belabor this issue, but I am curious as to the workings.
michael
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
