If you are unable to create a new account, please email support@bspsoftware.com

 

News:

MetaManager - Administrative Tools for IBM Cognos
Pricing starting at $2,100
Download Now    Learn More

Main Menu

Creating Dimension recommended practice?

Started by dylsing, 19 Sep 2006 04:52:44 AM

Previous topic - Next topic

dylsing

Hi,

I am new to using Powerplay Transformer. I need to create a dimension and I am unsure what's the best practice.

My source data (see below) has a book code (ISBN), Title, Source Country as available data columns. Both the Title and Source Country column is tied to the unique ISBN

Eg:
[ISBN]Ã,  Ã,  Ã,  Ã, [Title]Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, [Source Country]
1Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, JokesÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, China
2Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, SailingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  US
3Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, DivingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, Australia

I created a dimension as "ISBN" with "Title" and "Source Country" as its levels and I think it is incorrect, as I have no need of drilling down to the Title and Source Country.

My main question will be how do I create a dimension so that when I look at Powerplay, the crosstab table's format will look like below (with the measure being from another data source) with all the measures displayed as columns.

[ISBN]Ã,  Ã,  Ã,  Ã, [Title]Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, [Source Country]Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  [Price]
1Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, JokesÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, ChinaÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $5
2Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, SailingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  USÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $11
3Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, DivingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, AustraliaÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $4

Hope this makes sense.

Thank you.

srikalyan

Hello,

Have your book code as one dimension, the title as another dimension and source country as the third dimension.
This will help you in answering queries like - "Sales of books by ISBN code and by title and by country" or "Sales of books by ISBN code" or "Sales of books by title" or "Sales of books by country" etc.

Hope this satisfies your requirement.

Blue

I think you would need to create 3 dimensions in the Transformer model - ISBN, Title, and Country.  If your fact data source is only tied to the ISBN then modifiy it to include the Title and Country.  Then derive the Title and Country dimensions from the fact data source.

Alternatively you could nest the levels from the 1 in the rows.

What user interface are you using with the cube?

Blue
Robert Edis
Principal
Robert Edis Consulting
Rotorua, New Zealand

dylsing

#3
QuoteAlternatively you could nest the levels from the 1 in the rows.

What user interface are you using with the cube?

I will be using Powerplay for Web to access the cube results

One question:
What does "nest the levels from the 1 in the rows" mean?


1) My fact is only tied to the ISBN in another data source.
2) The Title and Source Country is tied only to ISBN (which is a primary key)

If I make ISBN, Title, and Country a dimension, the display I get is something like this:

Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  [Price]
1Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, JokesÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $5
Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, SailingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $0
Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, DivingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, $0
2Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, JokesÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $0
Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, SailingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $11
Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, DivingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, $0

Question 2> How can I design this or arrange the crosstab so that I can get the below structure and still employ the various queries that Srikalyan suggested:
[ISBN]Ã,  Ã,  Ã,  Ã, [Title]Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, [Source Country]Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  [Price]
1Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, JokesÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, ChinaÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $5
2Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã, SailingÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  USÃ,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  Ã,  $11

Thanks guys, really appreciate your advice.

Blue

I don't have access to PowerPlay Web.  This example is from Analysis Studio but you should be able to do that same with PP Web interface (see attachment).

I got this by the using the Expand <category> option after pulling in the top level of the dimension into the rows.

For the result you got with creating 3 separate dimensions all you need to do is suppress zeroes!  I susgest you do a bit of reading on how to use the interface.

Robert Edis
Principal
Robert Edis Consulting
Rotorua, New Zealand

dylsing

Hi,

I have actually already tried the suppressing of zeros and it does work rather well.

Thanks for all your input and it did give me an idea what to look out for   :D