Qliksense - Replace blanks/empty values after outer join

[lost_rabbit]

Here's the sample Qlik script 

Table1:

load * inline [
id,colA
1,34
2,45
,87
5,89
,11
];

Table2:

load * inline [
id,colB
1,22
6,78
7,22
,38
];

How do I replace the blank/empty values shown below with the text 'NOVALUE'

I even tried using len(trim()) and isnull() function but it does not work

Is it possible to do it in the chart?

[Vicky_Z]

You can find the "Show null values" in Properties panel . try to untick the option. See the screenshot in the post https://community.qlik.com/t5/New-to-Qlik-Sense/null-value-handling-in-table-box-i/td-p/86847

[lost_rabbit]

That would remove the entire row. I want to retain the row but replace the "-" values.

[lost_rabbit]

@Kushal_Chawda Any thoughts on how to handle nulls that show up due to  Qliksense join between tables?

[marcus_sommer]

You need to do this kind of job within the data-model maybe with a resident-load on the joined table and applying there the len(trim()) approach or using a mapping-load instead of the join - applymap() has a third parameter to define the non-matching results. Another way could be to check both tables against each other and then populating the missing data.

[lost_rabbit]

Gotcha.. Is there absolutely no way to handle these on the front end ?

[marcus_sommer]

I don't want to say that there is no way but IMO usually there is no sensible way. The reason is that there is nothing what could be shown. The relevant dimension-values don't exists within the dimensional context which is the base on which all calculations happens.

In some scenarios there are workarounds possible with functions like: alt(), coalesce() or range-functions as rangesum(YourExpression, 0) but quite often an appropriate bypass would require to create synthetic dimensions. Mostly this is complex and not really performant and related to more or less disadvantages in the UI navigation. If you think it are too much efforts to do it the script - each UI solution will be more expensive.

Beside this - these NULL's in the chart are not wrong else they show what the dataset contains respectively what's missing. IMO there is no mandatory need to adjust it.