topic Re: $() Expansion is ignoring the dimension filters in a chart table in App Development

$() Expansion is ignoring the dimension filters in a chart table

skafimn — Thu, 28 Mar 2024 10:59:20 GMT

Hello Qlik lovers, I have an interesting case that I came across and I hope you can help me solve it.

For this case, I am trying to filter against a numeric field [KPI_Code]

I created a Table and inserted [KPI_Code] as a Dimension.

I have 2 expressions, the results are in the table below

Expression A: Max(KPI_Code)

Expression B: Aggr(Only({<KPI_Code={"$(=Max(KPI_Code))"}>}KPI_Code),KPI_Code)

when I use a dollar sign expansion, it always gives me the maximum number in the dataset, disregarding the dimension value.

How can I use $() and still get values just like in Expression A ???

------------

Note: I need to use the $() because I need the output value inside a variable to call its expression

Example: vKPI_$()_Expression

if the $() resulted in 1001 then the variable would become vKPI_1001_Expression and I would use it's saved expression output in the cell beside its corresponding KPI_Code

KPI_Code	Expression A	Expression B
	1004	1004
1001	1001	-
1002	1002	-
1003	1003	-
1004	1004	1004

Re: $() Expansion is ignoring the dimension filters in a chart table

BrunPierre — Thu, 28 Mar 2024 11:11:48 GMT

=Aggr(Max({<KPI_Code={"$(=Max(KPI_Code))"}>} KPI_Code), KPI_Code)

Re: $() Expansion is ignoring the dimension filters in a chart table

marcus_sommer — Thu, 28 Mar 2024 14:44:59 GMT

Th $-expansion creates an adhoc-variable which could have only a single value at a time - and this is calculated before the chart and then applied to all rows.

Beside this you couldn't any condition within a set analysis which needs to respect the dimension-values because it worked like a selection on a column-level. To consider any row-level conditions you will need an if-loop.

In your case you may need an expression like:

if(Code = 1001, $(vKPI_1001), if(Code = 1002, $(vKPI_1002), ...))

which is probably not what you want to do.

I suggest that you consider to change the approach completely and moving the expression-logic into the dimensional layer within the data-model.