Read the sample results
The original sample contains a header followed by four customer records. The starting rules use customer_id as the exact unique key, require customer_id and credit_limit, and check joined_on as a YYYY-MM-DD date. There are five issues involving three data records. CSV record numbers include the header: the first data record is record 2. A quoted multiline field is still one record, so record numbers need not equal physical text-line numbers.
| CSV record | Field | Observation | Follow-up |
|---|---|---|---|
| 3 | customer_id | C-0007 was already present in record 2 | Check identity and correction rules; do not delete a row automatically |
| 3 | credit_limit | Empty under the selected required rule | Ask the field owner; a blank does not establish zero |
| 3 | joined_on | 03/04/2026 lacks a confirmed convention | Confirm 3 April versus 4 March before changing the value |
| 4 | joined_on | 2026-02-30 is not a real calendar date | Check the original record instead of rolling into March |
| 5 | Whole record | Four fields against three header columns | Check the separator and original export before discarding the extra field |
Set rules that match the target fields
After reading your own text or file, the tool leaves the unique-key, required and date rules unselected. Choose them from the actual import contract. Exact key comparison keeps case, spaces and leading zeros: A, a and space-A are different values. Empty keys are excluded from duplicate checking; select Required for that column if emptiness is prohibited. Every repeated occurrence after the first is listed, while the first is retained as the comparison reference.
A file contains 00042 and 42. Neither is flagged as a duplicate. Does this prove they are different customers?
Show the worked answer
No. It establishes only that the two strings differ. Confirm the source identity rule and target field type before deciding whether they refer to one customer. Do not strip leading zeros or convert identifiers to numbers merely to make the comparison pass.
Required means missing or whitespace-only content is an issue. A date check accepts real calendar dates written exactly as YYYY-MM-DD, including valid leap days, without interpreting time zones. A blank optional date is left alone. Blank and exact repeated header names are reported separately. The issue export contains record numbers, field names and issue labels; it does not include your cell values or change the original file.
Use the checker within its limits
| Input | Handling |
|---|---|
| UTF-8 CSV, at most 2 MiB | Choose comma, semicolon or tab explicitly; the first record is treated as headers |
| Quoted commas, line breaks and doubled quotes | Keep them within the same field; malformed quotes are rejected |
| Up to 20,000 data records and 150 columns | Display at most 200 issues; export includes every detected issue |
| XLSX, formulas, numeric ranges or target lookups | Not checked; export UTF-8 CSV and validate other rules separately |
This is a bounded structural and selected-field check, not an import approval. It does not detect every encoding, decimal, business-key, permission or relationship problem. Re-reading columns resets selected rules, because the file structure may have changed. Switching page language carries the current text and rules once through browser session storage, then consumes that transfer; Clear removes the pending transfer. Test a representative import and reconcile the target records before expanding the migration.