CSV import task schema check fails without header line

The CSV import task's schema check should detect that the field names are defaulted when there is no header (--no-header) and check each field schema by position. Currently, the check will fail because the inferred names, like "field_5", don't match the dataset's schema field names.