Filtering bar items without changing percentages.

[SMC1990]

Hi there,

I´m working with a bar graphic that looks like this 

This graphic reflects  a table in which we have some information about drugs that have been administered to a set of patients in a determinate clinical context. A very simple example is displayed below in order to illustrate the way data are represented.

Patient_ID	Item
11	A
11	C
13	B
14	C

We have also a table in which all Patient ID`s are collected.

Study ID

11

12

13

14

The "Item" attribute is placed at the "Axis" dimension in the graphic, and the "Patient ID" attribute of the second table is placed at the "Value" dimension, expresed as a percentage. So that, for each drug, the graphic shows the percentage of patientes that have been medicated with.

Well, here comes my doubt. I´m trying to create a filter object that allows to display only the items we are interested in, "hiding" the rest of them without -this is where the problem is- modifying the percentages. For example, if "A"=37,8% and "B"=20,8%, our desired filter would not end up resulting in a percentage change when both options were selected. If I simply create an "Item" filter, selecting a single option leads to a percentage change (100%). Selecting multiple options will lead to the desired result only if all patients have been medicated with at least one of the selected drugs. Is there some way to design a filter that can fit to these conditions?

Thanks in advice.

[lance_6]

VAR _num = COUNTROWS(<table>) // Will be taken into context with the visual
VAR _den = CALCULATE(COUNTROWS(<table>), ALL(<table>)) //all rows ignoring filters

RETURN
_num/_den