"Do I need to "lengthen" this table by making every single forecast value it's own row with a new column to demarcate it's MONTH/YEAR value, or do I leave it "wide" and retain all these columns as measures?"
I believe making the table more vertical would be the correct way to go. You can still associate each forecast record in this fact table to your date dimension, such as using the first day of the month, and then utilize attributes of the date dimension to describe the forecast record year and month.
Chris,
So you think that normalizing is a must here? I thought it might be, since there are like 28 different columns that would be considered measures.. I am currently looking at developing an SSIS package with an Unpivot Transformation to make this data set more vertical.
Do you think that is the right solution?
Thanks,
Jerid