DocDori

Fixing garbled Korean text in CSV files opened in Excel

Last updated 2026-11-10

Why Korean text shows up as ?? or strange characters when you open a CSV in Excel (it’s the encoding), and how to fix it step by step.

A CSV downloaded from another program often looks garbled when you double-click it to open in Excel, with Korean text showing up as something like ‘ë‚´ìš©’. The file isn’t broken. The problem is that the file’s character encoding differs from the encoding Excel guessed.

Why does this happen?

A CSV is plain text, so it may not contain any information about how its characters were stored. Most web services today save files as UTF-8, but Excel on Korean-language Windows tries to read unmarked CSVs using the older Korean legacy encoding (CP949/EUC-KR), which garbles the text. If the file starts with a marker called a BOM, Excel reads it correctly as UTF-8.

Fixing it directly in Excel

  1. Open Excel and choose Data → From Text/CSV.
  2. Pick the file, then change File Origin in the preview window to ‘65001: Unicode (UTF-8)’.
  3. Once the text looks right, click ‘Load’.

Fixing and saving it with DocDori

  1. Open the CSV in the Workspace. Commas inside quotes and line breaks within cells are read correctly.
  2. If a file can’t be read as UTF-8, DocDori asks whether to reread it as EUC-KR (the older Korean encoding). After you review or edit the content, it saves as UTF-8, and keeps the BOM if the original had one.
  3. If the saved file is UTF-8 without a BOM and looks garbled in Excel, open it with the ‘Data → From Text/CSV’ method above.

Things to keep in mind

When this usually happens

Frequently asked questions

It looks fine in Notepad but garbled in Excel. The file is fine; Excel guessed the wrong encoding. Opening it with the ‘Data → From Text/CSV’ method above solves it.

I already saved it while it was garbled. If the garbled characters were saved as they were, it’s hard to get the original text back. Downloading the original CSV again and starting over is the reliable option.

The leading zeros disappeared too. That’s a separate issue from encoding. See When leading zeros disappear.

Try it in the workspace