3 Replies Latest reply on Jul 13, 2016 9:34 AM by Luciano Vasconcelos

# LOD - Cluster problem

Hello everyone,

I am a new community user and i would like to ask how to solve this issue. I have a set of data about several customers among different years and I used table calculation in order to calc if one customer has bought more goods respect previous year or not. At the end I found several options, for example a customer could be called as "new" if he boughy 0 in the previous year and if he bought something in the current year or viceversa i could called it as "lost" if he didn't bought anything in the current year respect the previous etc etc.

Then i create a table in order to check if the results are correct and i used two headers: Cluster and Customer Name.

Cluster is a dimension (there are three possible cluster in which a customer could be insered) and this is the result:

So for example in the 2015, in Cluster 1, Allibert Benelux i a client type 5 (Client type is the calculated field explained before) and in the 2016 the same customer is a client type 4.

Now i would like to create a table with only cluster in which i can see the total number of client with "client type" = 1,2,3,4,5 ..... So for example for the Cluster 1 i would like to see this result:

Cluster
Client Type 2015 = 1
Client Type 2015 = 2
Client Type 2015 = 4Client Type 2015 = 5

Cluster 1

1002
Cluster 23001
Cluster 35112

How can i create this?

Thanks to all in advanced for help

Andrea

• ###### 1. Re: LOD - Cluster problem

What is client type? Is a dimension?

Post a sample of data.

• ###### 2. Re: LOD - Cluster problem

Yes Client Type is a dimension, For example 1 = New Customer, 2 = Lost Customer, etc etc

• ###### 3. Re: LOD - Cluster problem

Your print has client dimension. If you retire client and use countd(client) as metric it should work.