Forum Discussion
Helper IEquivalent to QV's Indirect Set Analysis
All,
I am an advanced QV developer and looking for a PBI equivalent to QV's indirect set analysis; the P and E functions.
For example, the following set analysis will calculate the average value for all excluded companies within a sector. Say I have 10 companies, selected one and this expression will perform average across remaining 9.
avg({<Company_Name= e(Company_Name), Sector= {'$(vSector)'}>} Value)
Hope someone can share some ideas.
Thanks!
Hi mapko
You can use the EXCEPT function to return the complement of the selected values of a particular column, similar to your example with the E function.
Here is an example pattern using arbitrary table/column/measure names:
= CALCULATE( [Value],
EXCEPT(
ALL( Company[Company_Name] ),
VALUES( Company[Company_Name] )
)
)I'm not really familiar with QlikView or the P & E functions (apart from Bing-ing them last night), but as for the P (possible values) function, I'm sure there are ways to replicate its behaviour in Power BI as well.
4 Replies
- OwenAuger
Super User
Hi mapko
You can use the EXCEPT function to return the complement of the selected values of a particular column, similar to your example with the E function.
Here is an example pattern using arbitrary table/column/measure names:
= CALCULATE( [Value],
EXCEPT(
ALL( Company[Company_Name] ),
VALUES( Company[Company_Name] )
)
)I'm not really familiar with QlikView or the P & E functions (apart from Bing-ing them last night), but as for the P (possible values) function, I'm sure there are ways to replicate its behaviour in Power BI as well.
- ankitpatira
Community Champion
mapko With slight variation you can do similar in power bi desktop. Just create a measure as below which will give you average of values of all the selection you will make. So instead of getting avergae for ones that you don't select, instead select all that you want the average for.
Measure = CALCULATE( AVERAGE(YOURTABLE[valueColumn]), FILTERS(YOURTABLE[companyColumn]))
mapkoankitpatira, thank you for your reply. I don't think this is a good workaround for me. Simply because there will be cases when I have to compare a single company's values against a big number of other companies and I can't expect from my users to hand pick those. Hope that makes sense.
