As I mentioned in my post a few weeks a ago, the PEL Allocate statement is a powerful method of moving data between PerformancePoint models.
Although it's powerful, the end result of an Allocate statement is simply that the destination model will contain data that has been queried from the source model. What if the destination model already contains data for the target scope? If this is the case, then you will have double counting. When creating the Allocate rule, there is unfortunately no option that lets you decide what you want to do with existing data. Ideally you need to have the 'Existing Data' option that you get when running an association, which gives you the 'Append', 'Scoped Replacement' and 'Full Replacement' options.
To get around this, I tend to place two rules inside a rule set. The first rule is a SQL implementation assignment rule with the same scope as the allocate rule, and deletes existing data by using the PEL statement this = NULL inside the rule. Then the second rule is the allocate statement, which appends the data. This is shown below:
The idea is then to execute the rule set, which will execute each of its rules in order. The advantage to doing this is that your whole data movement process can be executed in one clean step (perhaps scheduled or via Excel), without the hassle of executing several individual rules.