I have XML (~4MB) file with n-levels nested structure.
Opening it as LONG, 149 cols.
Every time I run EDT and open same .transform - it open exactly same, unmodified XML in different cols order (??).
My first step is “Remove Cols” (so only 5 cols is left) - and now it is removing different cols each time when I open this .transform file, so everything downstream (20+ steps) is wrong…
It is known issue?
I do not want to share whole file because its content.
Did some search - the “header-only.csv” trick will solve this? →
Yes, you should be able to use a Stack transform to handles this, as described in the link provided. However the same XML file input with the same settings should produce the same dataset each time. If you can provide us an example where that isn’t the case, we will investigate. If the data is sensitive, you can email it to support. Or perhaps you can create a non-sensitive example that displays the issue?
Do you mean that:
A. the exact same XML file can be input twice with the same setting and give different column ordering
B. Two similar XML files with the same basic schema can give different column ordering
B. is known and expected, as discussed here:
A. is not expected.
Thanks for the data. I believe this issues was fixed in v1.39.0, see number item 26 here:
I could reproduce the issue in v1.38.2, but not in v1.41.0. So I strongly recommend you upgrade to the latest version, v1.41.0:
There have also been a lot of other improvements.