Page 1 of 1
Budget and Forecast for Profit and Loss
Posted: Mon Apr 29, 2013 6:31 am
by winsonlee
There is a dimension call GL_Code and different group of cost is being forecast using different calculation. There are some similarities in the formula when the calculation is used. My question is comparing having rules that is easy to maintain vs rules that is faster, which one will you go for ?
Is it the case that each if statement exist in the rules is going to slow down the cube ?
Easy to maintain approach will be using attribute to determine the grouping of these cost. Eg if GL_Code attribute = A then use this formula else use some other formula.
Faster approach will be having calculation formula for each of this GL code.
Re: Budget and Forecast for Profit and Loss
Posted: Mon Apr 29, 2013 7:13 am
by cgaunt
Based on your description, I would go for the attributes option every time. I dont have any evidence to empirically prove that either way is noticably faster or slower....
I would observe that from everything I have ever built, neither way would be considered slow. I would instead be looking at the overall direction of the install that you have and the direction that you want it to take. If other code is 'high maintenance' and that is the way that the company wants it, the go with that. If you already have a lot of this attribute look up code, which overall saves time, then go with that.
Re: Budget and Forecast for Profit and Loss
Posted: Mon Apr 29, 2013 9:27 am
by rmackenzie
winsonlee wrote:Easy to maintain approach will be using attribute to determine the grouping of these cost. Eg if GL_Code attribute = A then use this formula else use some other formula.
Faster approach will be having calculation formula for each of this GL code.
It's impossible to say whether it is faster, or not - this is totally dependent on your cube design. Perhaps you could share that and you may get more information.
However, it is generally true that the attribute-based approach is going to be less maintenance. You need to think about how frequently the GL code dimension is going to be updated. If lots of accounts are added every month that are similar to existing ones, then yes, maintenance is probably going to be lower. However, have you considered that new accounts may be added for which new attributes need to be created, meaning new rules? In that case, you still have a maintenance problem...