Is it a bug? Is it a Feature Request? CSV failed rows re-importing with errors

[acornforth]

The feature to download a CSV of failed import rows is great, since it doesn't mess with the original format, and the column mappings will ignore the errors column on re-upload, it is trivial for a user to correct the CSV and re-upload it to the importer.

If there are any further errors and rows fail to import, another CSV with the failures is presented to the user, but in this case, unless the user first removes the appended errors column, we will now have a CSV with duplicate column headings, since it will add the new errors column (at the beginning, for some reason) and not take the original errors column into account at all.

Obviously this is very much an edge case. But a more frictionless experience would be to replace the errors column if it exists, instead of appending a duplicate.

This would allow users to rinse and repeat until all rows are imported.

I get that we can simply instruct our users to remove the errors column before attempting a re-upload. But better to make it as idiot-proof as possible.

It might be a good idea to make the name of the errors column customisable in any case, if it isn't already. aside from this case, for some apps, it's conceivable that an import might have a column with that name.

To reproduce the case above:

1. Import a CSV with deliberate validation errors.
2. Download the list of failures
3. Fix some rows, or not, just make sure there are still some errors before you re-upload and import the new file, leaving the errors column in the file.
4. download the new failures csv and note the duplicate errors column heading and that this file will now fail to import at all.