Need a table on a sheet to ignore other sheets' filter selections

[waterscg]

I've been given the following requirement by a customer. I figured this would be doable with Set Analysis but I'm not sure where to begin. Appreciate any ideas.

The data model contains work-related incidents, falling into three categories: facility, equipment, and product. Every incident is associated with one, two, or all three categories. The model includes an incident date dimension; in addition, there are ancillary date dimensions: facility affected date, equipment affected date, and product affected date.

The app is to be used for data analysis and will have three pertinent sheets: Facility Incidents, Equipment Incidents, and Product Incidents. Each sheet displays, in a table, only the incidents associated with that category. The Facility Incidents sheet has a Facility Affected Date filter; the Equipment Incidents sheet has an Equipment Affected Date filter; and the Product Incidents sheet has a Product Affected Date filter. There are other filters common to all three sheets.

The customer wants the incidents that are displayed on each sheet to be driven by the 'affected date' filter selection(s) on that sheet; i.e., they don't want the results to be affected by the 'affected date' filter selection(s) on the other two sheets. In addition, they want each sheet's 'affected date' filter selection(s) (i.e., the date(s)) to be unaffected by the 'affected date' filters on the other two sheets.

I hope I've explained this well-enough. Appreciate any ideas.

[Mauritz_SA]

Hi there

From what I understand from your query I can think of two possible solutions. First, using set analysis and second using alternate states.

Set analysis:

You can tell set analysis to disregard certain selections by adding an open = at the end of the argument. For example:

Something like Count({<[facility affected date] =, [equipment affected date] = >} IncidentId) can be used to count the number of product related incidents while disregarding the facility affected date and equipment affected date selections.

Alternate states:

Basically, you can use the same data set and apply different selections on it. Each sheet object can be linked to a set of selections (alternate states). Have a look at Mike Tarallo's video to get some guidance: https://www.youtube.com/watch?v=tsbnG3tOdR4

Sorry that it's so brief, but I thought I would try and get you going since no one else replied.

Regards,

Mauritz

[waterscg]

Thanks for the suggestions -- and for the link to the Alternate States video!

It looks like Alternate States will meet my customer's requirements.

There are a few filters on each sheet that are meant to be shared across all sheets; e.g., Country. I'm not sure if the customer has thought through the implications of shared filters; I'll discuss it when I meet with them later today.

[waterscg]

The customer liked the results but wants selections from filters common to all three sheets (e.g., "Country", "Manager", etc.) to impact the tables on all three sheets. So if the user selects "Canada" on the facility sheet, they want the table results to be filtered on the facility sheet and on the equipment and product sheets. I don't see a way to accomplish that with alternate states. Any suggestions? The Qlik Help mentions inheritance in the context of altered alternate states but I don't think it applies in this case.

[Mauritz_SA]

Hi there

I think set analysis will be the easiest solution. Exclude the filtered date fields on the other sheets like I showed in my first reply. If you still struggle then you can post your app and I will see if I can help.

Good luck.

[waterscg]

Are you suggesting I use set analysis instead of alternate states? I can try that ... but I'm unsure how to translate your pseudocode...

"Something like Count({<[facility affected date] =, [equipment affected date] = >} IncidentId) can be used to count the number of product related incidents while disregarding the facility affected date and equipment affected date selections."

...into something I can use on (say) the product incident table. I have only one measure column; it subtracts the 'shutdown' dimension from the 'restart' dimension so as to display the length of each incident; i.e., the incident length expression is: Only(([Product Restart Timestamp] - [Product Shutdown Timestamp])).

Are you suggesting I place the set analysis within the Only() expression, prior to the actual arithmetic? I tried that, but the product incident table is still being affected by the facility incident date and equipment incident date filters selected by the user on the respective sheets; i.e., if a facility incident date of 2019-04-24 is selected, the product incident table is being filtered to show just those incidents.

Update: I think I've determined at least one reason why the above isn't working. When (say) the user selects a facility incident date, the rows in the product incident date filter are being filtered ... which filters the results of the product incident table. So I would need to somehow apply set analysis to the three incident date filters -- to make them be unaffected by each other. Is that possible?

[Mauritz_SA]

Hi there

It is very difficult to give an answer without seeing the data model. If alternate states got you close then look at this help post: https://help.qlik.com/en-US/qlikview/November2017/Subsystems/Client/Content/Examples%20of%20Alternat...

You should be able to use the common filters with the default state and set up the other date filters and assign them to their own states.

Otherwise you can post your app and I'll try to help.