Validity table defines the range of the characteristic values for which the non-cumulative KF is valid.
Let us say your cube has 0calday, and it has data from 01/01/1999 to 12/31/2000. If you run a query and check (say its stock cube) stock balance per day per plant, it will calculate balances for each plant and show starting from 01/01/1999 onwards.
However, if you had a plant that became operational on 06/01/2000, you wouldn't want to see a 0 stockbalance for it starting from 01/01/1999 (or say when calculating the average, take as if stock existed from 01/01/1999) but from the date the plant actually existed ie 06/01/2000.
In this case, if your validity table had only 0calday, your report would show 0 stock for each day from 01/01/1999 to 06/01/2000 for this new plant (not a good idea), while if you defined it at plant level, you will see the report starting at 06/01/2000 while it will start from 01/01/1999 for other plants. This is what you want.
Let us say your cube has 0calday, and it has data from 01/01/1999 to 12/31/2000. If you run a query and check (say its stock cube) stock balance per day per plant, it will calculate balances for each plant and show starting from 01/01/1999 onwards.
However, if you had a plant that became operational on 06/01/2000, you wouldn't want to see a 0 stockbalance for it starting from 01/01/1999 (or say when calculating the average, take as if stock existed from 01/01/1999) but from the date the plant actually existed ie 06/01/2000.
In this case, if your validity table had only 0calday, your report would show 0 stock for each day from 01/01/1999 to 06/01/2000 for this new plant (not a good idea), while if you defined it at plant level, you will see the report starting at 06/01/2000 while it will start from 01/01/1999 for other plants. This is what you want.
