There is a difference between placing a Column in a filter bar vs. placing a parameter.

• A column is a defined data element that it is clear how it will be used in a visualization which is based on the same dataset (or model), as such, when placing it on a filter bar it behaves as a filter.
• A parameter is basically a list of values that might not have any relationship to a data element. You can select a value from the list, but you should also explicitly tell the system how and where to use this value. As such, when placed on a filter bar it is basically a value selector, not a filter. The value can be used to filter (currently using an expression filter, later on this year, by using filter binding), it can be used in calculations, or in strings. But the act of selecting a value cannot trigger and implicit action without the author defining this action.

Having said that, as I hinted above, we do see the use of parameter to filter as a common use case. In an upcoming update we will introduce the concept of binding parameters to column filters which will allow the parameter to drive the filter selection without the need to create expression filters. It will also help is reusing filter selections between canvases (which will allow dashboard filters to maintain their selection as the user move from one canvas to the next).

For the original use case, you will always need to have a filter, a parameter selector is not a filter. We will make it much easier to combine column filters and parameters, but it will remain the author responsibility to define how a parameter is used, filter or something else.