Fix CSV Dates Excel Flipped Between DD/MM and MM/DD

A UK file says 04/11/2023 meaning 4 November; a US machine reads it as 11 April, saves, and the damage is invisible. Drop the CSV below, state the order the file was written in, and convert every date column to the format you need — nothing is ever guessed.

✔ 100% free ✔ No signup ✔ Files never leave your browser — prove it ✔ Works offline after load

Quick Answer: CSVUndo standardizes CSV date formats flipped between DD/MM/YYYY, MM/DD/YYYY, and ISO YYYY-MM-DD. You specify the source date ordering so ambiguous dates are never guessed, and all conversions run 100% locally in your browser with zero data egress.

Drop your CSV file here or click to choose a file — it opens in your browser tab, never on a server
…or paste the data straight into the box below

No file to hand? Grab sample-dates.csv — ambiguous DD/MM dates — and drop it above. All samples.

Date order in and out
Output options for the converted file

Excel-safe export writes each converted date as ="2023-11-04", which Excel treats as text even on a double-click, so the next person to open the file cannot re-flip your work. Leave it off if the file goes straight to an importer, which may show the = literally.

🔒 Try it: switch off your Wi-Fi — the repair still runs, because your file never leaves this tab. We can't read, store, or leak what we never receive. Prove it in 30 seconds →

Why Excel flips the day and month in a CSV

A CSV is plain text with no date type, so Excel reads every date using your computer’s regional (locale) settings. A machine set to US English takes the first number of 04/11/2023 as the month and gets 11 April; a UK machine reads the same characters as 4 November. Open a file from the other side of the Atlantic, hit Save, and Excel writes its reading back into the CSV — and because only the days 12 and under can flip, the file looks mostly fine while a third of its dates now mean something else. The full story, including the tell-tale half-dates-half-text column, is in the companion guide: Excel changed my CSV date format. This exact mismatch is what makes QuickBooks reject a bank CSV outright — see our QuickBooks CSV import errors guide.

Why this tool never guesses — and what it does instead

Here is the honest core of it: 04/11/2023 is a valid date under both readings, so the source order is unknowable from the value alone. Any tool that claims to auto-detect it is guessing with your data. This one refuses to:

How to use it

  1. Drop the file or paste the rows, then click Find the date columns. Every date-like column is listed with its evidence: how many values prove which order, how many are ambiguous, and how many carry a risky two-digit year.
  2. Confirm the source order. If the data proved it, it is already selected; if not, choose the order the file was written in — you know where the export came from, the file does not say.
  3. Pick the output format. ISO YYYY-MM-DD is the safe default: the year leads, every part is fixed-width, and no locale on Earth misreads it.
  4. Click Convert & Download. The summary reports exactly how many values were converted, already correct, or left untouched — and the preview shows the result before you trust it.

The output formats

The four output formats and when each one is the right choice
FormatExampleUse it when
YYYY-MM-DD2023-11-04Always, if you can — ISO 8601 is unambiguous in every locale and sorts correctly as text.
DD/MM/YYYY04/11/2023A system that insists on UK/EU order.
MM/DD/YYYY11/04/2023A system that insists on US order (QuickBooks US, for example).
DD-MMM-YYYY04-Nov-2023Human readers — the spelled month cannot be misread in either direction.

Tick Excel-safe export and each converted date is written as ="2023-11-04", which Excel treats as text even on a double-click — so the next person to open the file cannot re-flip your work.

Two-digit years: the century is Excel’s rule, not ours

A year written with two digits, like 1/2/03, does not say which century it belongs to. Excel resolves it with a fixed rule — 00 to 29 becomes 2000 to 2029, and 30 to 99 becomes 1930 to 1999 — and this tool applies the same rule so the result matches what you already see on screen in Excel. The summary counts every value pivoted this way, because a date-of-birth or archive column from before 1930 will land in the wrong century. If the warning fires on data that old, fix the years at the source before converting.

What this tool can and cannot recover

If the file was only viewed in Excel, nothing is lost — the original characters are still on disk, and converting them to ISO here makes them permanently safe. If the file was misread and re-saved, only the days of 13 and above are provably intact; a half-flipped column shows up in the analysis as contradictory evidence, and for the ambiguous rows the honest answer is to re-export from the source. A different trap wearing the same costume: if Excel turned codes like MAR1 into 1-Mar, that damage cannot be reliably reversed by any tool — this page converts real dates between formats, it does not resurrect destroyed codes. And the way to stop all of it recurring is the safe import routine in how to open a CSV in Excel without breaking it.

Dates flipped right now?

Scroll back up, drop the file, state the order it was written in, and download the converted CSV in seconds — nothing leaves your machine.

Fix my dates now Read why Excel flips the day and month

Frequently asked questions

Why doesn't the tool detect the date order automatically?

Because 04/11/2023 is a valid date under both readings — 4 November and 11 April — the source order is unknowable from the value alone, so the tool never guesses. You state the order the file was written in. The analyzer only pre-selects an order the data proves, which takes a day of 13 or more: a value like 27/03/2023 can only be day-first, so it settles the whole column.

What happens to values that don't parse under the order I chose?

They are copied through untouched, byte for byte, and counted in the summary. A wrong choice therefore shows up as a large untouched count instead of silently mangling the file. Values that carry a time, like 04/11/2023 09:30, land in that count deliberately — converting them would mean dropping the time, and silent data loss is the thing this site exists to undo.

Can this tool un-flip a file Excel already re-saved?

Only partly, and we would rather say so than pretend. In a re-saved file, dates with a day of 13 or more were never touched by Excel, and a half-flipped column shows up here as contradictory evidence — some values proving day-first and some month-first — so the analyzer suggests nothing and tells you why. For the ambiguous rows there is no way to tell from the file which were flipped; the dependable fix is to re-export from the source system, ideally as ISO YYYY-MM-DD.

What does the two-digit year warning mean?

A year written with two digits, like 1/2/03, does not say which century it belongs to. Excel resolves it with a fixed rule — 00 to 29 becomes 2000 to 2029, and 30 to 99 becomes 1930 to 1999 — and this tool applies the same rule so the result matches what you already see on screen in Excel. The summary counts every value pivoted this way, because a date-of-birth or archive column from before 1930 will land in the wrong century.