Set analysis is a way to define an aggregation scope different from current selection. Think of it as a way to define a conditional aggregation. The condition – or filter – is written inside the aggregation function. For example, the following will sum the amounts pertaining to 2021:
Sum({<Year={2021}>} Amount)
This syntax however has a couple of drawbacks: First, it is not easy to combine a master measure with different set expressions, since the set expression is hard-coded inside the master measure. Secondly, if you have an expression with multiple aggregations, you need to write the same set expression in every aggregation function.
Therefore, we introduce an additional position for set expressions: They can now be written outside the aggregation function and will then affect all subsequent aggregations. This means that the below expression is allowed:
{<Year={2021}>} Sum(Amount) / Count(distinct Customer)
For master measures, this change will allow a very powerful re-usability: You can now add set expressions to tweak existing master measures:
{<Year={2021}>} [Master Measure]
Lexical scoping
The outer set expression will affect the entire expression, unless it is enclosed in round brackets. If so, the brackets define the lexical scope. For example, in the following expression, the set expression will only affect the aggregations inside the brackets - the Avg() call will not be affected.
( {<Year={2021}>} Sum(Amount) / Count(distinct Customer) ) – Avg(CustomerSales)
Position
The set expression must be placed in the beginning of the lexical scope.
Context and inheritance
Aggregation functions that lack set expression, will inherit the context from the outside: In earlier versions the context was always defined by the current selection. Now we have added the possibility of having the context defined by a set expression. So, now “context” means current selection or an outer set expression.
Inner set expression
If an aggregation function already contains a set expression, this will be merged with the context. The same merging rules as today will apply:
- An inner set expression with a set identifier will NOT inherit from the context. It will inherit the selection from the set identifier instead.
- An inner set expression that lacks set identifier – it has only a set modifier – will inherit from the context.
- How the merge is made depends on the set assignment for the field; whether it is made with an equals sign “=” or with an implicit set operator, e.g. “+=”. The logic is identical to how current selection is merged with a set expression.
Examples:
{<OuterSet>} Sum( {<InnerSet>} Field )
The OuterSet will be inherited into the InnerSet, since the inner set lacks set identifier.
{<OuterSet>} Sum( {$<InnerSet>} Field )
The OuterSet will not be inherited into the InnerSet, since the inner set expression contains a set identifier.
Aggr()
The set expression of the outer aggregation will never be inherited into the inner aggregation. But a set expression outside the outer aggregation will be inherited into both.
Examples:
Sum({<Set1>} Aggr(Count({<Set2>} Field )))
The Set1 will not be inherited into Set2.
{<OuterSet>} Sum({<Set1>} Aggr(Count({<Set2>} Field )))
The OuterSet will be inherited into both Set1 and Set2.
Summary
Nothing changes for existing set expressions – they will continue to work. But with this additional syntax we hope to simplify your work and your expressions and allow you to re-use your master measures more effectively.
This change affects all Qlik Sense editions from the August 2022 release. It will also be included in the next major QlikView release, planned for late spring 2023.
See more on
https://community.qlik.com/t5/Qlik-Design-Blog/A-Primer-on-Set-Analysis/ba-p/1468344
HIC
.png "Former Employee"){kind=icon}