What is a CSV file, and why do special characters break in Excel?

You downloaded a list from your bank, your accounting package or a web dashboard, opened it in Excel, and the names are wrong: a customer called Zoë shows up as Zoë, José as José, and the price column has € in front of every figure. Sometimes the whole row is crammed into a single cell instead, with no columns at all. Both problems come from the same place, and neither means your file is damaged — Excel simply guessed wrong about what it was reading.

Below you will find what a CSV file actually is, why the letters fall apart, and the concrete steps that put them back. Start from the reassuring part: nothing has been lost. The information inside the file is intact; it is only being read through the wrong lens.

What a CSV file actually is

A CSV file is a plain text file with nothing in it but writing. Each line is one row of the table, and the cells within a line are separated by a single punctuation mark. In other words, a CSV file is not an Excel file. Excel is simply one of many programs that can open it.

You can see this for yourself. Right-click the file, choose "Open with", and pick Notepad (or TextEdit if you are on a Mac). What appears is not a grid but a stack of lines. The first line is usually the column headings; every line under it is one record. Everything on that screen is the whole file — there is no hidden formatting and no saved settings behind it.

That is exactly where the trouble starts. An XLSX workbook records inside itself which cells are text, which are numbers, which are dates, and which alphabet table the file was written with. A CSV file records none of that. The program opening it has to guess all of it, and the glitches you are seeing are the moments when the guess was wrong.

Where é and € come from

Computers do not store letters; they store numbers. The list that says which number means which letter is called an encoding. Almost everything written today uses an encoding called UTF-8, in which the plain letters of the English alphabet take one number each, while accented letters such as é, ñ and ü — and symbols such as £ and € — take two or three.

If the program opening the file does not know which list the file was written with, it falls back on an older Windows list in which every number is exactly one letter. So the two numbers stored for é get read as two separate letters and land on screen as é. The letter was not corrupted, it was split in half, and the numbers inside the file never changed at all.

There is a practical consequence: if you tidy up the garbled version in Excel and save it, you make the damage permanent. Try to open the file with the right encoding first, and leave manual correction as a last resort.

The garbled pairs you are most likely to see

If what is on your screen matches the table below, the diagnosis is settled: the file was written in UTF-8 and opened with an older Windows list. You can also read the table backwards — find the sequence in the right-hand column and the left-hand column tells you which character it should have been.

Three ways to open the file correctly

All three work; which one you want depends on what you plan to do with the file. For a one-off look, the first is quickest. For a report that lands in your inbox every week, the third saves the most time.

• The import dialogue: open a blank workbook in Excel, go to the Data tab on the ribbon, click "From Text/CSV" and pick your file. The window that opens has a box labelled "File Origin"; choose the line reading 65001: Unicode (UTF-8) and the preview corrects itself immediately. The same window is where you set the delimiter.
• Re-saving through Notepad: open the file in Notepad, choose "Save As" from the File menu, and change the Encoding box at the bottom of the window to "UTF-8 with BOM". A BOM is an invisible mark placed at the start of the file; once Excel sees it, it stops guessing. Double-clicking the re-saved file now shows the accents correctly.
• Converting to XLSX: passing the file through the CSV to XLSX tool removes the guesswork entirely, because an XLSX workbook stores the encoding and the cell types inside itself. This is also the safer option if you are forwarding the file to colleagues, since nobody downstream has to repeat the fix.

When the whole row lands in one cell: the delimiter

The mark that separates cells is not the same everywhere. English-language installations of Excel expect a comma, which is why the format is called comma-separated in the first place. Much of continental Europe writes decimals with a comma — 12,5 rather than 12.5 — so exports produced there use a semicolon instead. Feed such a file to an English Excel and it sees no delimiter at all, so it puts the entire line into one cell.

The reverse happens just as often: hand a comma-separated file to a system expecting semicolons and you get the same wall of text. To find out which one your file uses, open it in Notepad and look at the first line.

The fix is the import dialogue again: pick the right mark in the Delimiter box and the preview splits into columns. When you are producing a file for another program or a web service rather than for a person, prefer the comma; the widely followed convention for CSV is built around it.

It is not only letters: the silent data loss

Double-clicking a CSV file has a consequence that is easier to miss than a broken accent. Excel treats anything that looks like a number as a number, and a leading zero in a number means nothing, so it gets dropped. Phone numbers written as 0207 lose their zero, product codes beginning with 00 shrink, and very long reference numbers have their last digits replaced with zeros.

None of this appears as an error on screen, and saving the file makes it permanent. So for any file containing phone numbers, reference codes or account identifiers, use the import dialogue rather than a double-click, and mark those columns as "Text" while importing.

If the data is headed for a program rather than a spreadsheet, converting straight to JSON is usually the shorter route; the CSV to JSON tool turns your heading row into field names and Excel never enters the picture. And when what you have is a workbook and what you need is a CSV, the XLSX to CSV tool covers the opposite direction.

Frequently asked questions

How do I open a CSV file in Excel without breaking the accents?

Instead of double-clicking, open Excel with a blank workbook, use "From Text/CSV" on the Data tab, and set File Origin to 65001: Unicode (UTF-8) in the window that appears. If you open the same file regularly, converting it once to XLSX is the permanent fix.

Has my data been lost in a garbled file?

No. The numbers inside the file never changed; they were simply read with the wrong alphabet table. Reopen the file with the correct encoding and the text comes back intact. The only way to lose it for real is to save over the garbled version.

Why does my file open in a single column?

The separator in the file does not match the one Excel expects. English installations expect a comma, so a file exported with semicolons ends up in one cell per row. Choose the right mark in the Delimiter box of the import dialogue.

Should I use CSV or XLSX?

For moving data into another program, CSV is enough and it is readable everywhere. If people will open it, if it holds phone numbers or codes, or if you need formulas and formatting, choose XLSX: the encoding and the cell types are stored inside the file, so there are no surprises.

Are my files uploaded to a server?

No. The conversions on this site run inside your browser; the file you choose never leaves your device and is not sent anywhere.

Last updated: August 27, 2026