Also use filename in the featuremerger if that makes it unique.

To expand a bit on what Niels said: Expose the fme_basename attribute, then create new xml_id's and xml_parent_id’s by prefixing the fme_basename. If the filenames are unique, then that should create completely unique ID’s.

Also: I’ve found myself moving away from the FeatureMerger in favor of the FeatureJoiner more and more, because sometimes I do have duplicates and still want them merged with other features, but not necessarily in lists. Might also help.

Hello @nielsgerrits, Hello @s.jager,

thank you for replies! Sorry, to ask, but don´t you know where to find this fme_basename? I was checking the “additional attributes to expose” but it´s empty.

@s.jager, for the calculation, do you use “Expression Evaluator” Or another? 

Thank you

It is deep in the featurereaders parameters.  Additional attributes to expose. On mobile so no screenshots or samples, sorry.

That’s strange. That ‘Additional attributes to expose’ dialog should be listing a bunch of things, in a FeatureReader and in a regular Reader:

How exactly are you reading your xml’s? Because I’m getting this….

And prefixing the fme_basename to create a new ID is easy, just use the AttributeCreator. Or, if you want to be specific, a StringConcatenator. 

Hello @s.jager, thank you! 

As a reader I´ve used the XSD-driven XML in The Reader, I think nothing special :/

Hmmm. I’ve never used the XSD-Driven XML Reader, but when I try that it does leave the ‘Additional attributes to expose’ dialog blank. 

Maybe try the regular XML Reader? That will also allow you to specify an XSD, if you need that.

Hello @s.jager,

with XML input it works fine - I had to make it from scratch but now it seems to work!

Thank you for help @s.jager, @nielsgerrits.