If I misunderstood the problem, then I apologize. But if the rows numbers all match each other anyway, wouldn't it be easiest to just throw an identity(1,1) column on each...

Hmmm, perhaps I don't understand this. I thought that all stored procedures are compiled and stored by sql server, and as such offer good performance. I guess I don't understand...

The thought occured to me that there was some change from SQL Server 2000 to 2005 regarding Text, so I googled it and I was right.

http://www.awprofessional.com/articles/article.asp?p=327394&seqNum=6&rl=1

So I...

Thanks.

Actually, if just one of the answers are ever large enough to require that they be stored as text datatype, I would need to convert everything to text I...

In response to your suggestion about negative identity columns:

Yes, I like your idea about making them less nice to work with so as to discourage app developers from being...

First of all, your sql worked brilliantly. Thank you. I had to fix some column and table names and throw in a couple of commas but I got the...

And yes, to answer your question, any given question would always have the same answer datatype over the years. If we are asking a company how many widgets they sold...

Yeah, currently I have it split up as follows:

create table answers (

answerID int identity (1,1) primary key,

questionID int not null,

answerYear numeric(4,0) not null,

userID int not null, /* user...

Sure. Basically we do a lot of surveys. Survey's have questions and answers. Some answers are text, some are varchar, some are int, and some are decimal.

The surveys...

Ok, so it sounds like you guys are saying it's better to split it up into separate tables as the datasets grow large for performance reasons.

So you should set...

This is an issue I have gone back and forth on a great deal. Splitting everything up into separate tables is such a pain, especially when it's just one field....