Remove duplicate rows from a CSV

Pick either Whole row identical or a single key column, then click Deduplicate. The output reports how many duplicates were dropped and how many rows survived, so you can sanity-check the result before using it.

Whole row or one column?

These answer different questions and picking the wrong one is the usual mistake.

• Whole row removes only exact copies — every field identical. Use it for genuine double-imports, where the same record was appended twice.
• One key column keeps the first row for each distinct value and drops later ones, even when the other fields differ. Use it for "one row per customer" when the same email appears with different timestamps or order values.

The first occurrence is always the one kept, so if the file is sorted oldest-first you keep the oldest record. Sort it the other way before deduplicating if you want the newest.

Why near-duplicates survive

Matching is exact and case-sensitive. These are all distinct values:

• ana@example.com and Ana@Example.com
• Belgrade and Belgrade  (a trailing space)
• +381601234567 and 0601234567

Exact matching is the safe default — fuzzy matching merges records that only look alike and is very hard to undo. If your key needs normalising, lowercase and trim it before deduplicating.

Excel's Remove Duplicates

It does the same job well, up to the point where the file fits — and it cannot open more than 1,048,576 rows. Deduplicating the parts of a split file separately does not work either, because duplicates spanning two parts are never compared. Deduplicate the whole file first, then split.

Common questions

Which duplicate is kept? The first occurrence in file order. Sort the file before deduplicating if you need the newest record instead.

Is matching case-sensitive? Yes. Ana@example.com and ana@example.com are treated as different values. Normalise the column beforehand if that matters.

Can I deduplicate on two columns at once? Not currently — either the whole row or one column. Concatenating the two into a single column beforehand achieves the same result.