@an-fischer Currently, if I don't sort the annotations alphabetically for the CSV export, they are sorted by their occurrence count (most frequent first). The problem still is that the application has no concept of any "order" of annotations. During import, the annotations are applied to the tokens and then that's where they are. If someone adds an annotation later, they might add it to the end, at the beginning or in the middle of the annotations. So as soon as data is edited through the platform, the order of annotations might even be differ from token to token. So in the context of the platform there is no "order of annotations" :(
@an-fischer Currently, if I don't sort the annotations alphabetically for the CSV export, they are sorted by their occurrence count (most frequent first). The problem still is that the application has no concept of any "order" of annotations. During import, the annotations are applied to the tokens and then that's where they are. If someone adds an annotation later, they might add it to the end, at the beginning or in the middle of the annotations. So as soon as data is edited through the platform, the order of annotations might even be differ from token to token. So in the context of the platform there is no "order of annotations" :(