Do not input private or sensitive data. View Qlik Privacy & Cookie Policy.
Skip to main content

Announcements
Join us in Bucharest on Sept 18th for Qlik's AI Reality Tour! Register Now
cancel
Showing results for 
Search instead for 
Did you mean: 
ali_hijazi
Partner - Master II
Partner - Master II

Pivot table with custom column dimension

Hello
below is a screenshot of a pivot table I developed (not all of it appears in this screenshot)

ali_hijazi_0-1747294232421.png

I got on the left 20 expressions and each cell has its own number formatting, money, or percentage or integer
and as a column dimension I have field defined in an inline table as follows:
load * inline [
rpt_colulmn
1
2
3
4...

];

in the definition of the column dimension in the pivot table I defined the following:
pick(rpt_column,'Unit','Actual','Actual %', 'Budget',....)

now in each expression I do the following
pick(rpt_column,

expression1,
expression2,
...

)

the result is correct but the table takes one minute to display result
is there a faster way to accomplish this?
kindly advise

I can walk on water when it freezes
Labels (2)
3 Replies
Aditya_Chitale
Specialist
Specialist

If you are using multiple match parameters in pick, it is bound to take more time to display the table,

one thing that you can check if you are using nested if conditions in your expression part, try to replace them with set analysis if possible.

Regards,

Aditya

rafaelencinas
Partner - Creator II
Partner - Creator II

Sometimes, the better code to implement our customizations isn't the better for the charts 😞 

In this case, you can try to create variables for each column and line to improve speed.

 

Bye!

 

Senior Qlik Architect
Cobra, Stallone, "You're a problem and I'm the solution"
marcus_sommer

Your approach is a nested if-loop - using pick(match()) is only a different syntax - and Qlik calculates at first all branches and picked then the first TRUE result. That is a rather unusual behaviour but it has big benefits by only a few loops with rather simple matching and the complete opposite by returning n complex calculations.

In other words such approach couldn't be fast. Better would be to transfer as much as possible of the calculations within the dimensionalities.