Posted by Scott Marquardt on 10/01/45 11:25

Consider a table that holds Internet browsing history for users/machines,
date/timed to the minute. The object is to tag all times that are separated
by previous and subsequent times by x number of minutes or less (it could
vary, and wouldn't necessarily be a convenient round number). This will
enable reporting "active time" for users (a dubious inference, but hey).

There are a lot of derivative ways of seeing this information that might be
good to get to. What's the fist and last of these sets of times? What
percentage of a given period is spanned by active times, and not? What is
the average duration of such periods? What is the average interval between
web hits during such periods? During other times?

Blah, blah. The basic problem is my principal problem. I don't have much
experience with cursors, but from what I understand it would be very good
indeed to spare them, given the number of records I anticipate working
with.

I'd be glad of any pointers.

--

Scott

[Back to original message]