Understanding this term is easy. One liner like "Table can provide up to which level of detail is known as its granularity" can make anybody understand what is it. But its easiness stops here.
Why am I saying 'its easiness stops here" ?? if you want to get a report coming out of 2 tables with different granularity then you will find the reason.
If number of tables increasing in your query having different granularity then you are just heading for some tough times if you don't know how to handle granularity.
If you are just writing queries manually then you can handle granularity issues easily but when queries is being created by cognos FM then your result would be wrong if you have not handled granularity carefully. You would start getting double counting of fact data and all those kind of issues that you don't want to see.
When I am talking about granularity then I am talking about facts and not dimensions.If you have a requirement where you need to join 2 facts. How do you join 2 facts with altogether different granularities ?
If somebody thinks that he can directly join 2 facts on common keys then he is planning for disaster.This is not the best idea and no body suggest it. There is one more reason you should not do it,because you can not keep customizing your model dedicated to a requirment.Anything that we do in the model we should try that we are generalizing it for other future requirements as well.
Getting a report out of 2 facts with different granularity means helping 2 people talk having different mother tongue.So what we do here, we have to bring a person who knows both the mother tongues,right ? Yes that is what we will have to do with these 2 facts. We will have to bring a commong dimension between these 2 facts.We also say these common dimension, conformed dimension. I hope you got the idea how to join 2 facts.Thanks
Vishwas
No comments:
Post a Comment