Description Missing name Mising name 2009-10-17 10:16:59 EDT

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; de; rv:1.9.0.11) Gecko/2009061212 Iceweasel/3.0.14 (Debian-3.0.14-1)
Build Identifier: Eclipse BIRT Designer Version 2.5.0.v20090603 

Data might have different path length to root, Birt should ignore empty/NULL dimensions and aggregate to next non null level.

Sample Dimension
US / NY / NY
US / CA / LA
US / Headquarters

if data is stored as parent/child this would typically result in this data sets:
US,NY,NY
US,CA,LA
NULL,US,HQ
in this case NULL should be ignored and US be promoted to top level.

if data is stored in flat table, data set would be:
US, NY, NY
US, CA, LA
US, NULL, HQ
NULL should be ignored and HQ attached to US

Depending on how dimension data is stored (parent/child or flat) NULL value might appear on different places. The basic problem remains the same Headquarters has to be on bottom level in order to be able to join with measure.
But in the crosstab HQ should be below country on the same level as states

This will also have effects on the sub totals. Sub totals on state should not result in sub total for HQ, since HQ is not a state, its data directly attached to country.

Know workarounds

for parent/child, copy child level to parent level if parent level is null
this results in dimension /US/US/HQ and in a sub total called 'US TOTAL'

for flat hierarchy, leave NULL as null
result /US/ /HQ ; sub total called 'null TOTAL'

Reproducible: Always