I'm not sure if you're referring to a columnar data export or the DATAEXPORT calculation command.
But in either case, the answer is "no". With DATAEXPORT you can pick the dimension to be used for the values columns from the dense dimensions in your cube.
If you really must have a specific order I would post-process the file using a script (although there are other options - ETL tools etc).
As TimG said, with DATAEXPORT you cannot do it.
You have other possible options though
1. If you are extracting a subset of data and if the amount of data is less, you can go with report script to define the order you want to
2. As TimG said, you can post-process it using a script or any other tool. If you have ODI in your in-house process, then you can easily do it
3. Long back when DATAEXPORT was not there (prior to system9) there was a CDF JExport with which you can extract a subset of data. This is similar to what we are doing now using DATAEXPORT. Have a look at (Oracle Essbase Sample Code) and look for CDF_Exports. Haven't used it after 18.104.22.168. AFAIK, it should work and you can define the dimension order
Hope this helps
Why do you try this link .. Cameron's Blog For Essbase Hackers: The fastest way to export targeted data from BSO Essbase with NONEMPTYBLOCK
I suggest use every dimension @relative than fixing and try year by year .. hope it helps