Problem with calculation of a new field/expression over 2 dimensions

[#1]

Hi,

I have a very simple test data set (see INPUT table in xls file for better formatting):

INPUT TABLE

apo rel_vrd gebj_bucket ddd_bucket

A 1 j i

A 2 j ii

A 2 j ii

A 4 m iii

B 5 m ii

B 3 m iiii

B 7 m i

C 8 m iii

C 9 o iii

C 7 o ii

Now I'm trying to calculate a pivot with the following values for selection apo = A (CORRECTION MATRIX UNNORMALIZED AND UNSUMMED in the xls):

j m o

i 3 1 -

ii 6 1 -

iii - 2 -

iiii - 1 -

The values should be calculated as "count( {<apo = >} rel_vrd)" * the vector (j=3, m=1, o=0 for all ddd_bucket), but I can't seem to get this to work in Qlikview.

Any idea on how to do this?

xls: http://www.box.net/shared/55mlyrnhs9

[#2]

Any help would be greatly appreciated!

[#3]

Hi Michel,

I looked at your example - xls the other day. Unfortunatly I have no Idea how to solve your problem, especially the product vector with matrix. So I am waiting (with you) for some other answers. Or do you have another algorithm ?

Regards Roland

[#4]

Roland,

I have gotten a little bit further, but I'm not there yet. The missing link now is that I need to create a pivot / matrix with the following values, where each row is equal to count( {<ddd_bucket= >} rel_vrd), but I have not been succesful in this last step yet:

j m o

i 3 1 -

ii 3 1 -

iii 3 1 -

iiii 3 1 -

I hope someone knows a "workaround" or trick. I think having matrices / tables in qlikview that can be dynamically updated (e.g. an extension of the variables concept) could make the tool extremely (more) powerful. This would allow many types of weighing and correction of data by demographic/population/etc.

[#5]

Anyone?