(From @porosnie:) The problem with splitting Donlon is that it takes time and it can enter in conflict with further updates online.
(sanity check) ensure that all features has a gml:identifier and that one is truly unique
look at how Andrei has split the file and create a simple XML file with the list of gml:identifiers and the destination file after the split
create an XSLT that can split Donlon based on that simple XML configuration file
The tasks can be allocated to a student who is currently in Eurocontrol and who has got a bit of XSLT experience.
(From @vog:) Donlon should be split along the following dimensions.
Per feature type, such as aixm:Airspace, aixm:AirportHeliport, etc.
Per time slice type (aixm:interpretation), i.e. BASELINE, TEMPDELTA, PERMDELTA, SNAPSHOT
This is important when creating a WFS-TE with Donlon dataset, where it is allowed to only insert BASELINE and TEMPDELTA time slices
(From @porosnie:) The problem with splitting Donlon is that it takes time and it can enter in conflict with further updates online.
(From @vog:) Donlon should be split along the following dimensions.