I've been faced with an interesting design dilemma.. I have a pretty
straight forward table such as:
+------------+------------------------+----------+------------+-------------------+
| keyid (INT) | site (VARCHAR 250) | hits (INT) | bytes (INT) | dts
(DATETIME) |
+------------+------------------------+----------+------------+-------------------+
This table will keep track of 7 days worth of data.. What I'm tasked
with, is to come up with a method that will take the 7 days of data, and
produce a top X sites over the 7 day period based on either cumulative
hits / bytes.
For example, to show the stop 25 Websites by way of hits for the last 7
days.
In perl/php I could easily select * from table WHERE (dts BETWEEN
start_date AND end_date), push the results into a hash and do my
additions from within the hash. Unfortunately, the language I am
programming this in, is not perl/php but is cold fusion.
My question, does it make sense to create a stored procedure that does
this, and just returns the data I want to present to the user? Or,
should I figure out some way for Cold Fusion to emulate the perl/php
hash method? (And no, I can't ditch Cold Fusion -- I tried already. ;) )
Thanks for any tips / help.
-Sam
