We have a model wit Customers (about 4 million) and the scores of those customers.
About the Customers we have about 30 characteristics like Age, Region, Married, ...
For each cutomer we have 40 scores.
So first we made 2 tables of it:
* Customers with 4 million rows and 30 columns
* Scores with 160 million rows and 4 columns (Customer_ID, Score_Domain, Score_Name, Score)
But this model was very slow.
The next thing we did was adding the scores to the cusomer table. Each score was an extra column in the Customer table.
In this way we had 1 table with 4 Million customers and 70 columns. This was much faster but now we have lots of problems creating charts on scores.
For instance : How do we make (in this model) a chart with an overview of the scores for 1 specific domain for a certain group (eg all the male cust) of customers?
Uptil now the only possibility i see is making an expression for each socre, but this is a very time consuming an inflexible way to make charts...
Any other suggestions to handle these wide tables?