Why visual similarity is not enough

"AB-104" and "AB-104 " contain different text if one has a trailing space. "00125" may be an identifier while "125" is a number. A blank cell may mean missing data rather than a duplicate. A cleanup rule should reflect the meaning of the dataset, not only what the grid appears to show.

A safe review workflow

1. Preserve the source

Save a copy of the XLSX or CSV before starting. Duplicate removal is a transformation, so you need a way to compare the output with the original.

2. Choose the identity columns

Use the smallest set of columns that identifies one record. A name alone may not identify a customer, while an order ID may be enough. Selecting every column can miss duplicates that differ only in a note or timestamp.

3. Inspect blanks and spaces

Check whether blank values should match one another. Look for leading or trailing spaces in copied data. Do not silently convert missing information into a real value just to make rows match.

4. Check case and number meaning

Decide whether uppercase and lowercase should be treated as the same key. For IDs, keep leading zeros when they carry meaning. For dates and numbers, verify that the displayed format is not hiding different underlying values.

5. Review groups before export

Inspect duplicate groups and confirm which row should remain. Then export, compare row counts, open the output in the target spreadsheet application, and check formulas, filters, and important formatting.

Example

A customer list has the values "North-01", "North-01 ", and "north-01". Before removing rows, decide whether the space is accidental and whether case matters to the business key. If they represent one customer, normalize deliberately and document the rule. If they represent separate codes, keep them.

Common mistakes

  • Deleting duplicates before saving a source copy.
  • Using a visible label when a hidden ID defines identity.
  • Treating every blank as the same record.
  • Dropping leading zeros from codes.
  • Assuming the browser output preserves every workbook feature without checking.

For a row-by-row workflow, read how to keep the first matching row. For previewing groups before deletion, see the duplicate review tutorial and open the tool guide. Microsoft's Remove duplicate values guidance explains the standard spreadsheet workflow.