topic Very basic Set Analysis Question in QlikView

Very basic Set Analysis Question

Fri, 31 Dec 2010 15:44:02 GMT

I am trying to do a set analysis that involves EXCLUDING all of the 'Data related error' counts from dimension CategoryDescription while at the same time only counting only those CategoryResult that = 'Failure'. Is this level of set analysis even possible in QV 9?

To think of this visually, i have a chart with 5 different CategoryDescription dimensions (A,B,C,D, and Data Related error) The sum of Attempts for ABCD are about 100 each but the sum of data related error is like 10000 because its not valid data for what im trying to do and thus. I want to exclude that particular type of Description entirely from my chart. But at the same time, I want to make sure that everything that IS counted is only from the population of data where CategoryResult = 'Failure' (Success and a few other results should be excluded)

Is it possible to set inclusions and exclusions on 2 different dimensions in the same set analysis formula? Obviously my formula below is wrong... any help?

sum( {$-1<CategoryDescription = {"*Data related error*"}>} (sum({$<CategoryResult = {'Failure'}>} Attempt )) )

Very basic Set Analysis Question

mike_garcia — Fri, 31 Dec 2010 17:53:21 GMT

Hello,

If I understood correctly, this is what you want:

=Sum({$<CategoryDescription -= {"*Data related error*"}, CategoryResult = {'Failure'}>} Attempt)

If you are new to Set Analysis, you should check some of my articles at http://iQlik.wordpress.com

Best Regards,

Mike

Very basic Set Analysis Question

Fri, 31 Dec 2010 19:45:38 GMT

Mike,

Thanks for quick reply. The updated formula works great... the only followup question I might have is.. Is there a way to have the chart still include the NULL values? Basically I am excluding "data related errors' but want to include the counts of blanks (nulls)

This may be something completely unrelated to the set analysis piece, but i am not selecting any option within that chart that suppresses or excludes nulls.

Thanks again

James