Thank you very much for your answer, which has put me on the right track. Your idea with the “Extract Variables” as an entry point for the reader was spot on.
I have now implemented a slightly modified way with Try/Catch. As I need an entry point anyway, I have put an empty table in front of it, but it already has the same structure as the CSV file. The catch statement now either accepts the existing data from the CSV file or the previously generated empty table.
Problem solved. Thank you very much!