This is a question of what is the best practice and best performance.

I have inherited a database that contains data for turbine engines.  I have found 20 data points that are calculated from several fields from the turbine.  The way it was done in the past is a view was create to pull data for some turbines and calculate some of the 20 data point.  Then other views for the same turbines but different data point and then other views for different turbines and data point.  So the same equations are used over and over.

I want to consolidate all of the equations (20 data point) into one place.  My debate is either creating a user function that will do all 20 calculations or creating them as computed columns in the table.  With a function it would calculate all 20 for each turbine even thou I might only need 2 or 3 for a view.  But as a computed column it would only calculate the columns the view pulled.

• Moved by Kalman TothModerator 7 hours 26 minutes ago Better fit

Is it the same calculation on all of the data points? If so why not just create a function and pass in the data for a single point and use the function as many time as necessary?

I wish they were all the same but each of the 20 data points has its on equation and some also have decision tree also to determine the correct equation to use.  Based on thinks like altitude, Air Temp and Fuel type...  These fields are updated quite often.  But only every few months is a new record added to the table Part of why I am considering Computed Columns.

>My debate is either creating a user function that will do all 20 calculations or creating them as computed columns in the table. 

I vote for UDF.