[andreaskagedal]

Great! Thanks. I will experiment and see what I can put into my PDFs using the pypdf library!

I think pypdf cannot write XMP Metadata.
I wrote a Script using the exiftool Library.
And currently I am working to replace it with xmp-toolkit

The  "PDF library method" worked fine. I could write all attributes with pypdf and they were picked up by MobileSheets when I imported the PDF. Great! (Actually, I did not try them all, but the ones I tried did work)

What exactly is the point of using the "xpath" attributes? It seems to be the same ones supported.

For me it would have been nice to also be able to import the "tempos" field. But it is not critical. If it is ever supported, I will probably use it. It is supported by the CVS import, so it does not seem completely unrealistic.

I see that forscore (the competing iPad app) also support importing some of the fields using the standard PDF field "Keywords". I guess the only point in suporting this would be to be able to import the same scores in both apps with the same result.