CSV Parser to provide "error file" suitable for recycling, after fixing
Let us say, input CSV has 1 header row and 10 data rows. 3 out 10 data rows failed for good reason.
Error output will have just the header row and 3 failed data rows, with additional column for error.
With that facility, Data Loader can fix errors as required and recycle the second file as input.
And there may even be another output with “succeeded rows”, but showing extra information about generated IDs etc. That will help the Data Loader to quickly query freshly loaded data.