Hi Csaba
Is this the requirement really from business or you wanted to test something? Because business always needs upto-date stock information. I mean stock data is something that moves very very frequently and you cant keep updating everything to them
All they want to see is how much stock is available as on today. 
Anyways, coming to the solution, my suggestion is to add a date field as a keyfield if you are using a DSO or add in time dimension if you are using a cube. (doesn't matter)
In the transformation, the date field should be mapped as system date and you could get what you need.
For example, if on 01.01.2015, the stock of the material is 10, then your system date will be 01.01.2015 and the stock will 10 for stock date 01.01.2015
On 02.01.2015, when the stock for 01.01.2015 is changed, then your system date will be 02.01.2015 and the stock value is 9 for stock date 01.01.2015
You could easily get this as per your expectation, but the data volume will be more based on the size of your business.
Also I wonder which source table you will get this information. The stock information is not bound to any dates (as far as I know). All the best
Hope this helps. Revert for any clarifications please
Regards
Karthik