Are you getting the expected outcome of the loop? KNIME in general really likes to let the user know anything that might be happening out of expected, even if this does not affect the result. It is, in the end, good as you will always be warned and always be aware if there is something that could potentially be a problem.
I think your approach seems to be a solid approach for me. That said, some of your warnings seem to be related with the variable Path not existing. Can you share a minimal reproducible example of your workflow?
In fact, the CSV reader node does not function properly. The CSV reader node somehow remembers the spec/structure of the CSV file it previously read-in and affect the CSV file imported in the current iteration.
Hello christine_ywl,
just enable the Support changing file schemas option on the Advanced tab of the reader in which case the node does not output a table specification during configure but can handle changing file structres. For more details see the File Handling documentation.
Bye
Tobias