TM1 Forum

Hi All

Has anyone come across the order of the data that gets exported is not the same as the order of elements in the dimension ?
Static subset is created through TI prior to exporting the data out.

If memory serves this something that can be found in the known issues forum. I might be wrong however. Have a look in there,

Jim.

http://blog.tm1tutorials.com/tag/asciioutput/

Found this posted by Wim Gielis

"The aim is to export a list of projects that generated revenues in a chosen year, but sorted by project name in column A. The project name is an alias on the project number element names. It is possible to create a subset containing projects at n level, showing the alias, and sorted on that alias. It is possible to assign this custom subset to a view. But when TM1 processes the view, he pays no attention to ordered subsets whatsoever."

Wim did mention on the post that it ignores the order of the subsets when exporting the data. Abit surprise that such functionality is not found in the TM1 process. Do you know if such problem still exist in the latest Tm1 10.2 ?

I doubt that's something that's changed. Hopefully if Wim see's this he may have tested it within 10.2 and will have a better idea of where this is,

Jim.

winsonlee wrote:Hi All

Has anyone come across the order of the data that gets exported is not the same as the order of elements in the dimension ?
Static subset is created through TI prior to exporting the data out.

Don't you mean "the data that gets exported is not the same is the order of the subset"???? When you process a view as a TI data source, regardless of whatever subsets you may be using, the data is processed according to the index number of the actual elements in each dimension and from Dimension 1 to Dimension x order. If you want to control the order yourself you would have to skip using a view as a data source and write code in the Prolog to loop through the subsets yourself and use a combination of CellGets and ASCIIOutputs to achieve what you want.

Yes, you are right, i am actually reffering to the order of the subset.
That would mean the best option now is to export the data out via prolog.

The only disadvantage of looping through the subset is not being able to supress zero.

Tom is correct.
And in TM1 10.2, this still holds true. No possibility to export in a quick way, taking advantage of views skipping non-existant data points.
But honestly, I don't get asked for this type of solutions often. Once a year or so, and then I use a batch process to do some kind of sorting afterwards, after the initial text file was created by TM1.