My 2 cents here , which are similar to above and some more.

Approach 1 :  Follow an INSERT only model.

Insert all incoming rows with a flag mentioning the latest ones...

Thanks for the help guys.

What I am interested to understand and learn is , the issues that one might face in going down this path .

Performance
Etl load time
Report run time 

The revisios

Posting some more updates  on one of the the fact tables : 
Some of the properties about the fact table are 

1. wrong - columns were added over time without thinking...

Hi All

Thanks for your replies and thanks Jeff Moden for the extensive research.

Yes the question was related to a fact table design. 

The background :...

Bringing an old topic alive , this is what I am thinking .

We have a fact table that is , say 1000 columns wide and then lets say...

Yes Jez we have that implementation too . Thanks for suggesting that .

I like Jeff's idea.Will do it that ways .

we created a primary key on the table, and looks like, it has helped .

studied the mcts book thoroughly for a month and solved some dumps from net.

your variable in app must be geting overriden by some other value assignment. check the variable that you are using to call the proc.

Turn on the SQL Server Profiler and then run your app. You will get to see the exact statement that is throwing error. It will help debugging the issue.

Thanks for your reply Sam . However I do want to know if there is a performance difference between updating all columns and updating 1 single column in a...

Production stats are pretty much uptodate.

Yes that is what is a good starting point.

Steps:

1. Analyse the query plan in case its a single query: Go...