Unlock a world of possibilities! Login now and discover the exclusive benefits awaiting you.
I have this Variable vUB (works great) which concists of the below expression:
vUB =
({<[FLOW TYPE]= {'Actual EOD (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'TCM (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'FW (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'O (Active) (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'T (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'S (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'W → (OLD DATA)'}>}AMOUNT)
-
SUM({<[FLOW TYPE]= {'1W (OLD DATA)'}>}AMOUNT)
However when i try to calculate the StDev results as per below:
stdev(vUB) = null
stdev($(vUB)) = error
any ideas?
Rgds,
Olle
Try a variable vUB definition like this:
SUM({<[TYPE]= {'A'}, Date *= {"<=01/01/2011"} >}Amount) +
SUM({<[TYPE]= {'B'}, Date *= {">01/01/2011"} >}Amount)
I think that the expression doesn't work because Sum and StDev are both aggregation functions, so you have to use stdev as you use sum, not togheter
I see, so I have to use aggr some way in my expressions?
Could you pls give me an example?
Use
stdev(ValueList(1,2,3,4))
but instead of 1,2,3,4
can be your Variable but you would need iit to be SUM(...),SUM(...) etc
Assuming vUB works on its own (ie as expression in the chart =$(vUB) ), then you should be able use an Aggr() like this:
=stdev(Aggr(($(vUB)), dim1, ... )
where dim1, ... is a comma separated list of all the chart/table dimensions
Or if you just wanting STdev of all [FLOW TYPE]
stdev(aggr(sum(AMOUNT), [FLOW TYPE]))
Closing in i think!
However getting different values when calculating in Excel and QV.
Any idea why? Pls see below, and pls see attached example doc.
I take it that qlikview is doing
STdev(Aggr(SUM({<TYPE={'A'}>}Amount)+SUM({<TYPE={'B'}>}Amount),Date))
?
I beleive you are ríght, question is why it differs from the Excel result?
Mystery of the day
Because you are using STDEV.P in Excel.
I believe the QV stdev() function works like STDEV.S in Excel.
For large sample sizes, both Excel functions (STDEV.S and STDEV.P) converge.
You can transform Stdev.S into Stdev.P by calculating something like
Stdev.S * sqrt(N-1) / sqrt(N) = Stdev.P
or
Stdev.S = sqrt(N) /sqrt(N-1) * Stdev.P
N being your sample size (count of values).