Set analysis P({$}) vs $::

[#1]

Here is the case. Base on a existing report, I need to create 2 straight table which only filter by Sales Person and Payment mode.

1st table display list of products which the Customer Group has purchased. It should looks like this:

Group Name	Product
Asia Grp 1	prod1, prod2, prod6, prod11
Asia Grp 2	prod3, prod4, prod5, prod6
Eur Grp 1	prod11
US Grp 1	prod 10, prod11

2nd table display list of products which a customer has purchased. It should looks like this:

Customer	Product
AAA	prod1, prod11
BBB	prod1
CCC	prod3, prod 4
DDD	prod 5, prod6

This 2 table can filter each other, as selection of a Group Name in the 1st table will filter 2nd table to display Customer belongs to that group. when a selection of a Customer, 1st table should display the Group information which this customer belong to.

So I create Product field as expressions like this:

Concat(<1{ SalesPerson= P({$} SalesPerson), PaymentMode= P({$} PaymentMode), CustomerGroup= P({$} CustomerGroup) }> [product], ' ,') for table one.

Concat(<1{ SalesPerson= P({$} SalesPerson), PaymentMode= P({$} PaymentMode), Customer= P({$} Customer) }> [product], ' ,') for table two.

It looks OK, but once I click on a Customer, the Customer Group will display the product purchased by the selected customer instead of the product purchased by the Group.

So I change the Set analysis to this:

Concat(<1{ SalesPerson= $::SalesPerson, PaymentMode= $::PaymentMode, CustomerGroup= $::CustomerGroup }> [product], ' ,') for table one.

Concat(<1{ SalesPerson= $::SalesPerson, PaymentMode= $::PaymentMode, Customer= $::Customer }> [product], ' ,')

for table two.

This time whatever I selected in one table it will not filter the other table....

Can someone explain to me what is the different between SalesPerson= P({$} SalesPerson) and SalesPerson= $::SalesPerson. How should I use them???