Problems with Campaign ans a huge set of Data

[#1]

Dear all

I have a problem designing the data model in QV.

In de appendix (pdf) I wrote down a simple draft of my database.

The second page was the idea i had, but it didn't work, because to many records.

The result of this report should be:

1.     Which products and customers where in the campaign (not buying, only promoted)

2.     Which products where purchased from which customer (really bought the products)

These two Questions should be answered but there is another thing:

1.     In one case, we know the campaign and products, but no customers

     -> Unknown amount of customers

2.     In the other case, we know the campaign and customers

     -> Unkown amount of products

3.     In the thirth case, we know the campaign, the products and the customers

     -> The full set is clear

4.     And in the last case, we only know the campaign (we promoted all products to all costomers in a range of time)

     -> So we have two unknown set's

I really running out of ideas, because either there are to many datas, or I'll get a circular reference.

I hope, someone could help me.

Thanks in adv.

Greetings

Ramon

[#2]

I'm afraid to have to do that.

Because I have to know what we have promoted, and what was bought from who.

So I really have to get all possible combinations at all.

Unfortunately that would be 18'00000000000 records, impossible to handle 😞

[marcus_sommer]

Yes these amount of data is too big - it also won't work with other tools ...

The only possibility is to generate such data on demand for only ONE promotion.

This could be done in load per where-clause or in GUI per creation from synthetic (and static) dimensions over the core-data - it's definitely complicated, slow and the amount of data is also limited with which it will work (RAM) and the usability is rather bad ...

- Marcus

[#4]

ok, got it.

So the requirements must be changed. I will go back to the management and explain the situation.

Anyway ... thanks for your help.

Greetings

Ramon

[#5 · Accepted solution]

finally I could solve the problem with setanalysis and alternate states.

so i was able to get de min. common amount and max. common amount without calculation all combinations of it.

greetings

Ramon