island table - impacting data model

[Mansi_Gupta1]

Problem Statement :

In UI we have 3 charts
chart1 should not be impacted by chart 2 and chart3
so we initially gave alternate state to chart2 and chart3.(these charts are working with global filters-which are shown in top through scope operator)

However this was impacting UI performance badly. So we are trying to achieve this through data model.

Data Model Table:

KeyField (having values a b c d..)
field1
field2
field3
field4
.
.Chart 1 uses keyfield and field 3, field 4 here. (field 1 and field 2 are for remaining charts)

Island table 1 : (residented data model table)
KeyField as KeyField1 (having values a b c d..)
field1 as field11

Island table 2 : (residented data model table)
KeyField as KeyField2 (having values a b c d..)
field2 as field22

in UI :
in filter2 we are giving field11 when keyfield1 matches a value (amongst the list)
in chart 2 we are giving expression :

count({<keyfield={variable(definition : trigger in page selects a)}>}field11)

what is happening ..
in selection area it shows variable a selected..

when we are navigating to next page where variable value b should be selected.. it does not overwrite that.

[marcus_sommer]

The used approach seems to be more complicated as necessary. Just to ignore some selections it's not necessary to use alternate states because these selections could be ignored within a set analysis.

If I understand the applied data-model with the island tables right - just in duplicating two fields within separate tables - and querying later those selections within set analysis it should work very well and could be done quite simply without any variables and/or buttons and selections, for example with something like:

count({< Filter2, Filter1 = p(IslandFilter1) >} Field)

and similar by the other charts. To show further which selections of the filter-fields are applied you could just query the island-table, like:

concat(IslandFilter1, ' + ')