Overwriting Deselection within Pivot Table

[Anonymous]

I have a Pivot table with 3 dimensions (2 vertical (location, specialty) & 1 horizontal (date).

I have 1 expression (sum of occupancy). This is conditionally set not to display 2 of the dimensions have one selection

(count (DISTINCT PK_LOCATION)>1 or count(DISTINCT DATEVALUE) >1 ) ).

(When these conditions are met another table displays).

If you select a value within the expressions, the conditions will be met and therefore the second table will display instead.

However, if you select a location first, or a date, selecting the value in the expression will deselect the original selection.

I.e. if you select a location and then select a expression value for one date, it will switch to displaying details for all locations on that one date, and therefore conditions will never be met.

Is there a way of disabling certain deselect funtions, or a similar workaround to make this functionality possible?

[swuehl]

I can't really reproduce your selection issue in a pivot table as I am able to select first on the first dimension, then on the second and keeping all selections.

Note that selecting a chart cell will actuall select the dimensional values.

Could you upload a small sample QVW that demonstrates your issue?

[Anonymous]

how can I share a qvw file on here? My previous attempt was rejected

[oknotsen]

Start a reply, click on "Use Advanced editor" in the top right, find the attach option near the bottom.

[Anonymous]

Attached is re-created example

[swuehl]

Ok, when I select location 1, the chart is limited to location one, still showing two dates.

Then I select the first date, and the selections are kept and your logic kicks in.

Is this not what you want?

[Anonymous]

If you select location 1 :

Then if you select the '1' value for specialty 1 on the 2/04/2016

The location selection is removed. This is the case with the total numbers, and is more prevalent in the dashboard I am working with

[swuehl]

Still I can't really reproduce your issue. On which version are you?

[Anonymous]

v11.20

The logic of this is correct, but i'm trying to find a work around.