A repeated order number can be perfectly valid. An order for a notebook and a pen often occupies two rows in a sales export. Removing duplicates on the order number alone would erase one of the items.
Work out what one row represents
Consider this small order export.
| Order | Line | SKU | Quantity |
|---|---|---|---|
| 4108 | 1 | NOTE-A5 | 2 |
| 4108 | 2 | PEN-BL | 1 |
| 4108 | 2 | PEN-BL | 1 |
Each row represents an order line. The second and third records match in every field. If the source confirms that one line was exported twice, keeping a single copy leaves two lines and three units. Keeping a single row for the entire order would report either one or two units, depending on which row survived.
Read the differences before keeping a copy
Suppose the last record has quantity 3 instead. The two records now share the same order and line number but disagree about the quantity. An exact-row check will keep them both. Matching by Order + Line surfaces the disagreement.
Look for a transaction timestamp, amendment number, or another field that explains the change. Check whether the second entry replaced the first or added to it. If the system keeps revisions, sort by a reliable revision field before choosing which occurrence to retain. If the file lacks that information, refer the conflict back to its owner.
Verify the business total
Compare three quantities before and after the repair: record count, distinct order count, and units sold. In the first example the expected result is two records, one order, and three units. A row count alone would miss an incorrect choice of key.
Use the duplicate removal tool to preview the records your chosen columns bring together.
