topic Is it a glitch or bad set analysis in QlikView

Is it a glitch or bad set analysis

jduenyas — Wed, 25 Nov 2020 16:16:04 GMT

In the attached QVW (some of the data encrypted) the chart shows Division "Bitzu" (in red) calculated value 1.00 for the month of Dec 2018. When selecting this Division from the List Box, the value for Dec 2018 changes to 3.67 (which is the correct value.)

Before selecting Bitzu

and after selecting the division

This is obviously wrong. The calculated value should not change when selecting the Division.

QVW Attached

Does anyone understand what is going on?

Thanks

Re: Is it a glitch or bad set analysis

sunny_talwar — Mon, 07 Jan 2019 17:44:53 GMT

Try this expression

=Sum(Aggr(Accuracy, ID, MonthYear, Div))/Count(DISTINCT ID)

Re: Is it a glitch or bad set analysis

jduenyas — Mon, 07 Jan 2019 18:01:30 GMT

Yes it does take care of the miscalculation but it should not be calculated this way. None of the other divisions present this error and - selecting both Division AND Dec 2018 from the MonthYear shows in the table that the data does not have to be further aggregated. (Or does it?)

Thanks

Re: Is it a glitch or bad set analysis

jduenyas — Mon, 07 Jan 2019 18:10:40 GMT

In the meantime I will use your solution as it does take care of the issue (and reports are due out today... 🙂 ) but I am not completely satisfied with that.

Thanks though.

Re: Is it a glitch or bad set analysis

sunny_talwar — Mon, 07 Jan 2019 18:18:55 GMT

I am unable to find the exact issue, but to me this looks like a grain mis-match problem... you can read about this here

Pitfalls of the Aggr function

Re: Is it a glitch or bad set analysis

jduenyas — Mon, 07 Jan 2019 18:23:03 GMT

OK. I found where the error is.

The ID which is supposed to be unique actually is not! and that creates the problem.

Once again, thanks for looking into it.

Josh