My tools

Excel dedupe

Drop an .xlsx or paste cells — duplicates show up at once.

Compare

How to use it

  1. Drag an .xlsx onto the card, pick one with “Choose .xlsx”, or copy cells in Excel and paste them here. The counts and the preview show up at once — there is no button to press, and nothing leaves your device.
  2. Leave “Compare” on “Whole row” to remove rows that repeat entirely. Switch to “Selected columns” and tick the business key (order ID, email, phone) when only part of a row has to match — the order you tick columns in does not matter.
  3. Check the “Removed” view: it lists each dropped row’s number and the row that was kept. If the newest record should survive, set “Keep” to “Last”, then download the result as .xlsx or .csv.

Example

Input

Order ID	Customer	Email	Country
1001	Ann Lee	ann@example.com	US
1002	Ben Carter	ben@example.com	UK
1003	Chen Wei	chen@example.com	CN
1002	Ben Carter	ben@example.com	UK
1004	Dana Ito	dana@example.com	JP
1002	Ben Carter	ben@example.com	UK

Result

6 rows → 4 kept · 2 removed

#  | Order ID | Customer   | Email           | Country
2  | 1001     | Ann Lee    | ann@example.com | US
3  | 1002     | Ben Carter | ben@example.com | UK
4  | 1003     | Chen Wei   | chen@example.com| CN
6  | 1004     | Dana Ito   | dana@example.com| JP

Removed: rows 5 and 7, both kept as row 3

Row 1002 appears three times (rows 3, 5 and 7), so rows 5 and 7 are dropped and row 3 is kept. Leading zeros such as 0001 survive, and comparison follows the text you see on screen — the number 1 and the text “1” count as the same value.

Rules, limits and known issues

  • Values only: the downloaded .xlsx is a brand-new workbook. Formatting, formulas, comments, conditional formatting, data validation, charts, images and merged cells are not carried over — dedupe moves rows, so relative formula references would break anyway. Your original file is never modified.
  • The limits are hard: up to 10 MiB per file, and up to 50,000 rows, 200 columns or 500,000 cells in the selected sheet. Above that the file is refused with an explicit message instead of freezing the tab. The preview draws the first 200 rows; the counts are always exact.
  • Dates and times are converted from Excel serial numbers using the workbook format (1900 or 1904 date system) and shown as ISO (2025-08-01 or 2025-08-01 12:00:00). Other number formats — thousands separators, decimal places, percentages, currency — are not applied; numbers are shown at full precision instead.
  • Comparison uses the text you see, so the number 1 and the text “1” are treated as the same value (Excel’s own Remove Duplicates keeps them apart). “Ignore outer spaces” also strips full-width spaces (U+3000), non-breaking spaces and zero-width characters; spaces inside a cell are never changed.
  • In “Selected columns” mode, a row whose selected cells are all empty is always kept: a blank key is not evidence of a duplicate. Row numbers are the physical row numbers of the sheet, so row 1 is the header when you use one.
  • One worksheet at a time — pick another in the sheet list. .xls, .xlsm, .ods, password-protected workbooks, cross-sheet dedupe, merging, approximate matching (Ann vs Anna) and Unicode normalisation are out of scope.
  • Exports: .xlsx keeps the values with a bold header, a frozen first row and real date cells; .csv is UTF-8 with a BOM so Excel opens Chinese and Arabic text without garbling. With “First row is header” off, no header row is written in either format.

What people usually need next, once the list is clean.

CSV viewer

Text cleaner