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

Framework Model

Started by Divya, 24 Jun 2009 12:39:34 AM

Previous topic - Next topic

Divya

Hello all,

I have got a need to create 3 measures like count of cust, count of child and count of cust child combo that provides unique count over pairs of attributes of facts. its important that these measures are count distinct as the combinmations are expected to appear against multiple dimensions, but should only be counted once for the summary.

also i need a method to calculate these measures.

Please anybody help me with your valuable suggestions

Thanks
divya

BIsrik

I guess you need to use count distinct function and create new query items...

Srik

Divya

Thank you very much for your reply.

Can you be bit detail. what is the procedure involved in it.
should i need to create model query subject for the measures like  count ofcust id, count of child id......

actually my team asked me to create three measures
count of cust
count of child
count of cust child combo

i can create only two measures count of cust and count of child from the logical layer query subjects, but unable to create count of cust child combo as there is no query item like that in logical layer.

how to create count of cust child combo?
please give me the procedure for it? thanks

BIsrik

I suggest to create the measures in the model query subject. For Cust Child, create a query item for this combo (like concat both the query items) and use that for the count.

Srik

Divya

The question to solve is how do we get a count of the combinations of Customer and Child. It seems reasonable to expect we'll need to derive some sort of key first that incorporates the customer and child, and then aggregate (count) over that key. But there might be a better way (something with a separate Query Subject with determinants perhaps), again, not sure the best way to do it,

We need the count of Customer+Child combinations to be broken down by other dimensions.

Please give me some suggestion so that it would be great

blom0344

No idea how your data looks like, but assuming you have meaningless keys for both customer and child dimensions why not create a composite out this pairing and perform a distinct count on the new data-item?