How to use the Convert CSV Encoding
- 1 Drop the CSV file on the studio, or try the example (a Windows-1252 file).
- 2 Under Source, open “Character encoding”. Try the encodings until names like Joël and Zoë read correctly.
- 3 If text is already garbled in the file itself, add the step “Repair garbled characters”.
- 4 Under Export options choose the output encoding, for example UTF-8, and download.
Characters are bytes plus a guess
A text file is a series of bytes. Which letter a byte stands for depends on the encoding, and the file does not say which one was used. Open a Windows-1252 file as UTF-8 and “é” turns into a replacement character; open UTF-8 as Windows-1252 and it becomes “é”. The studio detects UTF-8 (including the byte-order mark and UTF-16) reliably, because invalid UTF-8 is easy to spot, and otherwise falls back to Windows-1252, the usual Western European encoding. It tells you when that is a guess.
Choosing and converting
Under Source you can pick UTF-8, UTF-16, Windows-1252, ISO-8859-1, ISO-8859-15 (with the euro sign), Windows-1250, Windows-1251 or Mac Roman. The preview updates immediately. On export the same list, minus the rarest, is available. Characters that the target encoding cannot hold – a euro sign in ISO-8859-1, or any Chinese character in Windows-1252 – are written as “?” and counted in a warning, so you know before you download. UTF-8 holds everything and is what the WHATWG Encoding Standard recommends for new files.
Text that was already damaged
If a program already read the file with the wrong encoding and saved it again, the garbled characters are now real text. “Repair garbled characters” reverses the typical case, UTF-8 read as Windows-1252, including text damaged twice. It leaves text alone that has no sign of damage.
Limits
Automatic detection cannot tell the single-byte encodings apart (Windows-1250 from Windows-1252, for instance); choose the right one by eye. Chinese, Japanese and Korean legacy encodings are not supported.
Frequently asked questions
Which encoding should I export?
UTF-8 unless the receiving program insists otherwise. Excel on Windows reads UTF-8 correctly when the file starts with a byte-order mark; use “UTF-8 with byte-order mark” for that.
What does a “?” in the output mean?
The chosen encoding has no character for it. The warning shows how many. Export as UTF-8 to keep them.
The file looks right in the preview but wrong in my program – why?
Your program is guessing the encoding. Save as UTF-8 with a byte-order mark, which most Windows programs recognise.
Last updated: