Skip to main content
Announcements
Have questions about Qlik Connect? Join us live on April 10th, at 11 AM ET: SIGN UP NOW
cancel
Showing results for 
Search instead for 
Did you mean: 
marcus_steggall
Creator
Creator

Rolling 12 months calc, not showing correct figures

Hi all, hope you're having a good day.

I am hoping, you can solve my head scratching query, around a rolling 12 month calculation not working.

As per my attached qvd, there are two tables.

First one showing, 12 month totals per month.

The second, showing the individual months total

So if you added up 2018 Oct to 2017 Nov = 387,496,198, which you'll see from the first table is 100% correct.

If you do the next one being, 2018 Sep to 2017 Oct = 413,311,256 which then doesnt equal the first table of 413,307,617.

So i'm a little confused, why this is not showing the correct total for the 12 months total?!

Any help, would be greatly appreciated.

1 Solution

Accepted Solutions
sunny_talwar

The problem was that your inner set analysis check against Client_Name also included set analysis to only look at >Max_Period - 11 <Max_Period. So, the limits were only checked for those revenues which fell within that date range. So, essentially, any customer which was in the limit but didn't have any sales within the date range was excluded. I removed that bit of the set analysis and boom you got what you needed.

Does that make sense?

View solution in original post

19 Replies
Gysbert_Wassenaar

You need to sort the virtual table you create in the aggr function by the period/month. Try

sum(aggr(rangesum(above(sum({<Period_Sort>} Revenue),0,12)),(Period_Sort,(NUMERIC, ASCENDING))))

* Avg({<[Period_Sort]={">=$(=max([Period_Sort])-11)<=$(=max([Period_Sort]))"},Month= ,Year_YYYY= ,Quarter_QQ= , Period_YYYY_Month=>} 1)


talk is cheap, supply exceeds demand
marcus_steggall
Creator
Creator
Author

Hi Gysbert, many thanks for your reply, it's very much appreciated

using just the calc you stated above does work on it's own!

However, did you see within my expression i've got two variables working at client level for vMinLimit and vMaxLimit?

i thought putting the aggr at the beginning of the calc and the numeric ascending after the Revenue),0,12 would work but this doesn't seem to work? have i missed something obvious?

=RangeSum(Above(Sum({<[Period_Sort]=,Month= ,Year_YYYY= ,Quarter_QQ= , Period_YYYY_Month=,

Client_Name={"=round(Sum({<[Period_Sort]={"">=$(=max([Period_Sort])-11)<=$(=max([Period_Sort]))""},Month= ,Year_YYYY= ,Quarter_QQ= ,Period_YYYY_Month =>}Revenue))<=$(vMinLimit)
or
round(Sum({<[Period_Sort]={"">=$(=max([Period_Sort])-11)<=$(=max([Period_Sort]))""},Month= ,Year_YYYY= ,Quarter_QQ= ,Period_YYYY_Month= >}Revenue))>=$(vMaxLimit)"
}>} Revenue), 0, 12))

*
Avg({<[Period_Sort]={">=$(=max([Period_Sort])-11)<=$(=max([Period_Sort]))"},Month= ,Year_YYYY= ,Quarter_QQ= , Period_YYYY_Month=>} 1)

Gysbert_Wassenaar

Yes, you're missing a sum(aggr(.... around your rangesum.


talk is cheap, supply exceeds demand
jensmunnichs
Creator III
Creator III

I think he's also missing the red bit in ,(Period_Sort,(NUMERIC, ASCENDING)))) as he said he's only adding the numeric ascending.

By the way, is this only available in Qlikview 12? I've been following this thread because it seems interesting but when I enter your expression I just get 'Error in expression: ')' expected'. Am using Qlikview 11.

sunny_talwar

Yes, this is only available in QV12 and above versions of QlikView. Available in most versions of Qlik Sense.

The sortable Aggr function is finally here!

jensmunnichs
Creator III
Creator III

Thanks Sunny!

marcus_steggall
Creator
Creator
Author

Hi guys, many thanks for the replies, an interesting subject this,

I've attached the updated expression, which I think it should be based on your replies and the answer is still not what i expected.

2018 Sep is Showing 413,307,617 when it should be 413,311,256.

Can you see any glaring mistakes, in my code?

Many thanks for taking the time

sunny_talwar

There is a difference between the two tables that you have... Individual month totals is for all clients (because it uses Sum(Revenue) as expression) whereas Rolling 12 month totals only include Client_Name which meet this condition

Client_Name = {"=Round(Sum({<[Period_Sort]={"">=$(=max([Period_Sort])-11)<=$(=max([Period_Sort]))""}, Month, Year_YYYY, Quarter_QQ, Period_YYYY_Month>}Revenue)) <= $(vMinLimit)

or

Round(Sum({<[Period_Sort]={"">=$(=max([Period_Sort])-11)<=$(=max([Period_Sort]))""}, Month, Year_YYYY, Quarter_QQ, Period_YYYY_Month>}Revenue)) >= $(vMaxLimit)"}

If you add this condition to the individual month totals, the monthly numbers will change and you give you 413,307,617. If you think, that 413,311,256 is the right number, then remove the set analysis from your Rolling 12 month totals table


Capture.PNG

marcus_steggall
Creator
Creator
Author

Hi Sunny,

Appreciate your views as always

The only reason I put the individual months table in there, was to show the real time numbers each month, so you could add the 12 months together.

As you can see below from the individual table, this is where i've got the 413,311,256 from!

So surely by using the set analysis, why doesn't this just include the Client_Name which meets this condition highlighted?

Thanks