Sounds like you should model your data like this.
Create the following dimension tables
Sales - > Unique sales fields
Claims - > Unique claim fields
Products -> Product name, Product Type
Customers -> Customer Name, Customer Type
You will need to create unique keys perhaps concatenating 2 or 3 fields to make the unique value if you do not have a unique sales or claim number.
Create a Fact table and load the common data into it.
ProductsKey -> Material = PARTS NUMBER
CustomerKey -> S-t-P = IMPORTER NUMBER
Fact Type = 'SALES VALUE (Euro), CLAIMS NET VALUE, CLAIMS NET VALUE (Euro) , Quanties Etc..'
Date -> Created On = DOCUMENT DATE
Do this and you should have a model you can use to answer the questions you have asked and more.
Hope this helps
You can create the Key Column like this.
[Created On]&'|'&[S-t-P]&'|'&[Material] as [KeyColumn]
[DOCUMENT DATE]&'|'&[IMPORTER NUMBER]&'|'&[PARTS NUMBER] as [KeyColumn]
The above keys are for the respective tables and the association will be done based on that. Join the main Columns and make the key column so that the data wont be cross join.
Tell me whether it works or not.
You Can do the association using composite key.
For Sale2016 Table:
num([Created On])&'-'&[S-t-P]&'-'&[Material] as [Key]
convert [Created on] into num format else it will become trouble.
For Raw Claims Data Table:
num([DOCUMENT DATE])&'-'&[IMPORTER NUMBER]&'-'&[PARTS NUMBER] as [Key]
The two keys automatically get associated.