I am trying to export data in column format and import back to replace 0 with #Missing. I want to see if it will reduce any blocks. When I export I am getting data header members from a different dimension other that Period. It is getting tricky to map these many data rows in rule file as oppose to 12 from Period.
What determines this behavior. I know I have tried in the past (different app) and Period came through as data header record.
Any help is appreciated.
You can't directly control which dimension is selected (at least, not without changing your cube) in a native export.
However, for what you are doing, you do not need a load rule - you could just find-and-replace in the file.
You can also try converting zero to #Missing with a calc script (see Re: BSO Level 0 Block Analysis Advice) although it's going to be slow and you'd need to defrag with a dense restructure afterward.