Hi,
I'm out of ideas on the set analysis / aggregation side but I think the problem for Qlikview may be that there is not a complete matrix to work with.
When you add 'ddd_bucket' and 'gebj_bucket' to a pivot chart there are some combinations not present in the 'in memory' table you've created from your input file.
Perhaps if you were to add in rows for the missing combinations - with no value in rel_vrd - you'd get the result requested ,
e.g. if you've already loaded your input file called, say, 'INPUT'
then repeat a number of join loads to combine each element used with every other element to create a complete matrix.
INPUT_PLUS:
LOAD DISTINCT ddd_bucket as Xddd_bucket
RESIDENT INPUT;
OUTER JOIN
LOAD DISTINCT gebj_bucket as Xgebj_bucket
RESIDENT INPUT;
OUTER JOIN
LOAD DISTINCT apo as Xapo
RESIDENT INPUT;
OUTER JOIN
LOAD atc5 as Xatc5
RESIDENT INPUT;
Then use the following to fill in the rel_vrd values -
LEFT JOIN
LOAD apo AS Xapo,
ddd_bucket AS Xddd_bucket,
gebj_bucket AS Xgebj_bucket,
rel_vrd AS Xrel_vrd,
atc5 AS Xatc5
RESIDENT INPUT;
Now try working with INPUT_PLUS as the in memory table for your pivot, using the code from the previous posts to ignore ddd_bucket and 'TOTAL' by gebj_bucket, you should at least now be working with a complete matrix of dimension combinations (only some of which have rel_vrd values)
Regards,
HD
