A little challenge for hard core app developers ^^

[herard_bertrand]

Hello,

I'm working on an app, in which I need to show which boots can be set up on which Ski
(there is an example attached to play with). 
Here is the simplified model

Here is table of the app : 
(Boots, Ski, a simple calculation to explicit if there is a match or not) and Ski_season is a filter

When I click on a value on Ski_season here is the actual displays (I understand Why) : 

So my aim is to have the following behaviour : 
When I select the Tyre Season, I'd like to still sees every Boots_Name in the table (in this example 4 lines), and if the ski_name is excluded by a ski_season selection, to see a null value in the Ski_name column and the value Unmatch in the calculation ? Is it possible ?

Constraints : 
- We can't modify the link table to add lines to specify that there are no match between ski and boots. (That is to say, the link table cannot be a cartesian product between Boots and Ski with a flag indicating is there is a match or not) (because of the size of the true application=> performance issue). 
- We cannot put the ski_season field in the link_table.(because in reality there are much more fields an values in the true application => performance issue).

Infos:  
- In the link table : every Boots is in the link Table, not every Ski
- If a Boots have no corresponding Ski => a line with the Boots_ID and null value for Ski_ID exists in the Link_Table.

There is an example attached 

Best regards

[Oleg_Troyansky]

Hi there!

It looks like your biggest problem is how to show all Boots, no matter what was selected - right?

See if this blog post can give you some ideas:

https://www.naturalsynergies.com/q-tip-20-showing-all-values-in-qlik-sense/

[herard_bertrand · Accepted solution]

Thanks, 

"It looks like your biggest problem is how to show all Boots, no matter what was selected - right?" 

=> That's exactly the point.

I tried the solution on your bolg link and it works for my mini example,

Now I'm going to try it on the true huge complex model 🙂 

[marcus_sommer]

Especially if it's a huge data-set and there are already issues with the performance it might be worth to rethink the data-model approach and not using a link-table else a star-scheme data-model - even more if the link-table isn't capable to resolve the missing key-stuff.

Beside this the populating of missing data does not mandatory mean to keep a cartesian data-set from any to everything else quite often it could be reduced to the essential data and are not so heavy as all the workaround-stuff of link-tables and calculated dimensions ...