I have a wrangler transform whose first directive is
However, by mistake, my CSV file has records separated by blanks. As a consequence the Wrangler emits 0 outputs, which is fine. But it also emit 0 errors, and that is wrong. It should send all the bad records to the error output.