On Sat, Mar 13, 2010 at 02:45:37PM -0500, Phpster wrote: > I'd go with a mysql data modelled approach as it will allow mysql to > do lots of the heavy lifting during analysis as you've mentioned. If > there are a lot of entries, it's gonna get complex and expensive > memory-wise to manage XML or session based datasets. > > Plus having each question as it's own record should give you greater > flexibility in packaging the data for analysis and reporting. >
+1 I invariably find that the original design for a project needs to be tweaked. The customer wants to add or delete questions, they want to add/change reports for the data. So I nearly always approach this kind of project this way. Maybe: vote table: id serial/sequence not null primary key voter_id int references voters (voter_id) question_id varchar(10) answer varchar(10) You can easily subset by voter, or by question ID. Or analyze answers in relation to other answers, etc. Paul -- Paul M. Foster -- PHP General Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
