You are not logged in.

Dear visitor, welcome to Palo Community Forum. If this is your first visit here, please read the Help. It explains in detail how this page works. To use all features of this page, you should consider registering. Please use the registration form, to register here or read more information about the registration process. If you are already registered, please login here.

Bassanio

Trainee

  • "Bassanio" is male
  • "Bassanio" started this thread

Posts: 2

Date of registration: Apr 21st 2011

  • Send private message

1

Thursday, April 21st 2011, 4:26pm

Merging and analyzing several data inputs

Dear All,

still quite a new user I am so far quite ok with working with basic Palo, but now that I go into more compley topics, I do need your advise basically on how to link different sources of information in one or several cubes.
Lets say I do have several input sources like:

Source 1: Dimensions (Country, Month (consolidated to year), Customer Number, Transaction Currency and Transaction Describtion) --- Sales Quantity and Revenue per transaction (several per month per customer)

Source 2: Dimensions (Country, Month, Customer Number) --- Outstanding payments at the end of the month

Source 3: Dimensions (Country, Year) --- Market Size per year

So my point where I am now is that, as you see, I do have repeating dimensions in the 3 sources, however never the same dimensions.

My aim is later to compare via a subset e.g. in month May which customer had the highest revenue but also the highest outstanding payments, or per year in which country was the highest revenue compared to the market size.

Thank you for your help and advise.
Bassanio

  • "holger_b" is male

Posts: 271

Date of registration: Jun 1st 2010

Location: Freiburg, Germany

Occupation: BI Consultant

  • Send private message

2

Thursday, April 21st 2011, 11:45pm

RE: Merging and analyzing several data inputs

Hi Bassanio,

your case is a perfect example of how to work with three cubes. Provided each client is located in just one country, you could even have just one dimension which depicts both your clients-by-region and the regions in one dimension, you would just need sort of a "dummy" element to hold data which are related to a country instead of a customer.

So, you would have three cubes, the dimensions of each being exactly the ones you described in your example. Do not hesitate to come back with more questions as you go along the way of implementing this.

Have fun
Holger

Bassanio

Trainee

  • "Bassanio" is male
  • "Bassanio" started this thread

Posts: 2

Date of registration: Apr 21st 2011

  • Send private message

3

Sunday, April 24th 2011, 8:33pm

RE: Merging and analyzing several data inputs

Dear Holger,

thank you for your fast and helpful answer. All cubes are now created and data is uploaded. Just to make sure: I have so far put all "key numbers" like revenue or market size in one dimension called measures. Is that for future analysis the correct way or shall I create 3 different measures dimensions where I will upload the data from every input source?
And how will it be possible for me to spread the subset over 2 cubes?
Or would it be more helpful to create only one cube in which, in case that e.g. "Transaction Currency" is not available in Source 3, I here refer this value in the Setdata forula to a value called "Total".
Thank you and a Happy Easter Holidays to everybody in the Forum...

Bassanio

This post has been edited 2 times, last edit by "Bassanio" (Apr 26th 2011, 3:03pm)


  • "holger_b" is male

Posts: 271

Date of registration: Jun 1st 2010

Location: Freiburg, Germany

Occupation: BI Consultant

  • Send private message

4

Tuesday, April 26th 2011, 9:02pm

RE: Merging and analyzing several data inputs

Hi,

actually there is no law for this. On the one hand, it may be easier to query just one cube instead of three, on the other hand, you may have to fiddle around with many meaningless dummy elements in case you put everything into one cube. And: several smaller cubes with a higher filling rate may perform better than one big cube with only few data in it. So actually it depends.

Maybe you could start playing around with two or three solutions and try to figure out which one suits best to answer the kind of questions you will put, I believe that is what I would do.

Regards
Holger

Rate this thread